id
stringlengths 4
10
| text
stringlengths 4
2.14M
| source
stringclasses 2
values | created
timestamp[s]date 2001-05-16 21:05:09
2025-01-01 03:38:30
| added
stringdate 2025-04-01 04:05:38
2025-04-01 07:14:06
| metadata
dict |
|---|---|---|---|---|---|
1023067228
|
Update ignoring-files.md
Clarified that a successful "touch .gitignore" gives no output.
Why:
Closes #11077
What's being changed:
Check off the following:
[ ] I have reviewed my changes in staging (look for the latest deployment event in your pull request's timeline, then click View deployment).
[x] For content changes, I have completed the self-review checklist.
Writer impact (This section is for GitHub staff members only):
[ ] This pull request impacts the contribution experience
[ ] I have added the 'writer impact' label
[ ] I have added a description and/or a video demo of the changes below (e.g. a "before and after video")
@bphermansson
Thanks so much for opening a PR! I'll get this triaged for review :zap:
|
gharchive/pull-request
| 2021-10-11T20:01:14 |
2025-04-01T04:34:21.934023
|
{
"authors": [
"bphermansson",
"ramyaparimi"
],
"repo": "github/docs",
"url": "https://github.com/github/docs/pull/11090",
"license": "CC-BY-4.0",
"license_type": "permissive",
"license_source": "github-api"
}
|
1113427409
|
Update index.md
Why:
Closes [issue link]
What's being changed:
Check off the following:
[ ] I have reviewed my changes in staging (look for "Automatically generated comment" and click Modified to view your latest changes).
[ ] For content changes, I have completed the self-review checklist.
Writer impact (This section is for GitHub staff members only):
[ ] This pull request impacts the contribution experience
[ ] I have added the 'writer impact' label
[ ] I have added a description and/or a video demo of the changes below (e.g. a "before and after video")
This seems to be opened by accident, so I am going to close it now :yellow_heart: Take a look at the contributing.md to learn how to contribute in this repo. You can open a discussion if you have any questions :sparkles:
|
gharchive/pull-request
| 2022-01-25T04:59:47 |
2025-04-01T04:34:21.938447
|
{
"authors": [
"RaghuRV1709",
"ramyaparimi"
],
"repo": "github/docs",
"url": "https://github.com/github/docs/pull/14630",
"license": "CC-BY-4.0",
"license_type": "permissive",
"license_source": "github-api"
}
|
1932242778
|
Update create_properties.md to reference repository and pusher_type.
It looks like if a CreateEvent is generated specifically when a repository is created, separate from the main branch being created. In that event, the ref_type is repository and ref is null.
Additionally, out of the 22,000 events I looked at for a particular hour a day ago, all of them had pusher_type=user. Not sure how valuable it is to include in the table if it's always the same value, but it's there.
Here's an example CreateEvent payload:
"payload":{"ref":null,"ref_type":"repository","master_branch":"main","description":null,"pusher_type":"user"}
Why:
Closes: https://github.com/github/docs/issues/28994
What's being changed (if available, include any code snippets, screenshots, or gifs):
Just a bit of the docs at this page: https://docs.github.com/en/rest/overview/github-event-types?apiVersion=2022-11-28#event-payload-object-for-createevent
Check off the following:
[x] I have reviewed my changes in staging, available via the View deployment link in this PR's timeline.
For content changes, you will also see an automatically generated comment with links directly to pages you've modified. The comment won't appear if your PR only edits files in the data directory.
[x] For content changes, I have completed the self-review checklist.
dependabot merge
|
gharchive/pull-request
| 2023-10-09T03:32:36 |
2025-04-01T04:34:21.943286
|
{
"authors": [
"Diego200208",
"nanodeath"
],
"repo": "github/docs",
"url": "https://github.com/github/docs/pull/28996",
"license": "CC-BY-4.0",
"license_type": "permissive",
"license_source": "github-api"
}
|
2140056902
|
docs(notifications): add missing security_advisory_credit reason
Why:
Closes: #31696
security_advisory_credit reason is missing in the Notifications REST API documentation
Closes:
What's being changed (if available, include any code snippets, screenshots, or gifs):
Check off the following:
[x] I have reviewed my changes in staging, available via the View deployment link in this PR's timeline (this link will be available after opening the PR).
For content changes, you will also see an automatically generated comment with links directly to pages you've modified. The comment won't appear if your PR only edits files in the data directory.
[x] For content changes, I have completed the self-review checklist.
@setchy Thanks so much for opening a PR! I'll get this triaged for review ✨
|
gharchive/pull-request
| 2024-02-17T12:51:39 |
2025-04-01T04:34:21.947177
|
{
"authors": [
"nguyenalex836",
"setchy"
],
"repo": "github/docs",
"url": "https://github.com/github/docs/pull/31695",
"license": "CC-BY-4.0",
"license_type": "permissive",
"license_source": "github-api"
}
|
2674823315
|
Fix bullet list formatting
Why:
This PR unifies the formatting of bullet lists (see issue).
Closes: #35344
What's being changed (if available, include any code snippets, screenshots, or gifs):
The format **Keyword**—Text has been changed to the **Keyword**: Text format.
Check off the following:
[x] I have reviewed my changes in staging, available via the View deployment link in this PR's timeline (this link will be available after opening the PR).
For content changes, you will also see an automatically generated comment with links directly to pages you've modified. The comment won't appear if your PR only edits files in the data directory.
[x] For content changes, I have completed the self-review checklist.
@akordowski Your changes look great! 💛 I'm working on getting an approval from our Site Policy team since their files are being updated, and we will get this merged right after ✨
|
gharchive/pull-request
| 2024-11-20T07:42:51 |
2025-04-01T04:34:21.951064
|
{
"authors": [
"akordowski",
"nguyenalex836"
],
"repo": "github/docs",
"url": "https://github.com/github/docs/pull/35359",
"license": "CC-BY-4.0",
"license_type": "permissive",
"license_source": "github-api"
}
|
633253624
|
4.6 python for everyone (Coursera)
4.6 Write a program to prompt the user for hours and rate per hour using input to compute gross pay. Pay should be the normal rate for hours up to 40 and time-and-a-half for the hourly rate for all hours worked above 40 hours. Put the logic to do the computation of time-and-a-half in a function called computepay() and use the function to do the computation. The function should return a value. Use 45 hours and a rate of 10.50 per hour to test the program (the pay should be 498.75). You should use input to read a string and float() to convert the string to a number. Do not worry about error checking the user input unless you want to - you can assume the user types numbers properly. Do not name your variable sum or use the sum() function.
code:
def computepay():
hrs=input("Enter hours:")
h=float(hrs)
rate=input("Enter Rate:")
r=float(rate)
pay=(hr)
if h<=40:
return(pay)
else:
ovrpay = ((r40) +(h-40)(r1.5))
return(ovrpay)
result=computepay()
print("Pay",result)
https://github.com/github/explore
|
gharchive/issue
| 2020-06-07T09:52:41 |
2025-04-01T04:34:21.954198
|
{
"authors": [
"Abrar-aslam",
"Jackson0852"
],
"repo": "github/explore",
"url": "https://github.com/github/explore/issues/1521",
"license": "CC-BY-4.0",
"license_type": "permissive",
"license_source": "github-api"
}
|
731436375
|
Update index.md
The is an editor basically a text editor.
Thank you for contributing! Please confirm this pull request meets the following requirements:
[x] I followed the contributing guidelines: https://github.com/github/explore/blob/master/CONTRIBUTING.md
[ ] I have no affiliation with the project I am suggesting (as a maintainer, creator, contractor, or employee).
Which change are you proposing?
[x] Suggesting edits to an existing topic or collection
[ ] Curating a new topic or collection
[ ] Something that does not neatly fit into the binary options above
Editing an existing topic or collection
I'm suggesting these edits to an existing topic or collection:
[ ] Image (and my file is *.png, square, dimensions 288x288)
[ ] Content (and my changes are in index.md)
Please replace this line with an explanation of why you think these changes should be made.
Curating a new topic or collection
[ ] I've formatted my changes as a new folder directory, named for the topic or collection as it appears in the URL on GitHub (e.g. https://github.com/topics/[NAME] or https://github.com/collections/[NAME])
[ ] My folder contains a *.png image (if applicable) and index.md
[ ] All required fields in my index.md conform to the Style Guide and API docs: https://github.com/github/explore/tree/master/docs
Please replace this line with an explanation of why you think this topic or collection should be curated.
Something that does not neatly fit into the binary options above
[ ] My suggested edits are not about an existing topic or collection, or at least not a single one
[ ] My suggested edits are not about curating a new topic or collection, or at least not a single one
[ ] My suggested edits conform to the Style Guide and API docs: https://github.com/github/explore/tree/master/docs
Please replace this line with an explanation of your proposed changes.
Please note: we will close your PR without comment if you do not check the boxes above and provide ALL requested information.
It looks like this is an empty commit
|
gharchive/pull-request
| 2020-10-28T13:21:30 |
2025-04-01T04:34:21.962055
|
{
"authors": [
"Sahil-ojla",
"iancanderson"
],
"repo": "github/explore",
"url": "https://github.com/github/explore/pull/1892",
"license": "CC-BY-4.0",
"license_type": "permissive",
"license_source": "github-api"
}
|
1104299107
|
Add a collection with cheatsheets
Please confirm this pull request meets the following requirements:
[x] I followed the contributing guidelines: https://github.com/github/explore/blob/master/CONTRIBUTING.md.
[x] I have no affiliation with the project I am suggesting (as a maintainer, creator, contractor, or employee).
Which change are you proposing?
[ ] Suggesting edits to an existing topic or collection
[x] Curating a new topic or collection
[ ] Something that does not neatly fit into the binary options above
Curating a new topic or collection
[x] I've formatted my changes as a new folder directory, named for the topic or collection as it appears in the URL on GitHub (e.g. https://github.com/topics/[NAME] or https://github.com/collections/[NAME])
[ ] My folder contains a *.png image (if applicable) and index.md
[x] All required fields in my index.md conform to the Style Guide and API docs: https://github.com/github/explore/tree/master/docs
I think it's useful to add cheatsheets as the total number of different cheatsheets can be overwhelming for many programmers and it's great to have a collection of well-made ones.
Thanks again @Luois45!
|
gharchive/pull-request
| 2022-01-15T02:43:36 |
2025-04-01T04:34:21.966480
|
{
"authors": [
"Luois45",
"MikeMcQuaid"
],
"repo": "github/explore",
"url": "https://github.com/github/explore/pull/2851",
"license": "CC-BY-4.0",
"license_type": "permissive",
"license_source": "github-api"
}
|
3259640
|
catchAll doesn't receive unmatched messages when hubot addressed directly
I'm not sure if I'm using catchAll correctly, but it doesn't seem to receive messages which are addressed directly to hubot.
module.exports = (robot) ->
robot.catchAll (msg) ->
msg.send "I heard you say #{msg.message.text}"
$ ./bin/hubot
Hubot> qwerty
Hubot> I heard you say qwerty
Hubot> hubot qwerty
Hubot>
Am I using this correctly, or is there another way to listen for those messages? I only want to respond to messages that have no matches and are prefixed with /^hubot\s+/.
The easiest way would be to have a generic catch all named something that'd be loaded last in your scripts/ folder alphabetically.
Hmmm, I'm not sure what you mean---I figured this was as generic an example as I could make it, and it's not getting anything starting with "hubot". I could be wrong, but I assume I should still be able to get those messages...?
As for loading order, doesn't it still give pass a message to all scripts asynchronously? And if I do get the message, I see "done" and a "finished" to detect if something's already been processed, but I don't see any scripts that call it, so I don't see how I can know for sure whether a message I receive has been processed or not.
They're almost certainly getting matched by the respond action which technically is a match. :) Catch all wasn't designed to work with replies but probably should. I'll see if I have some time in the next few days to tackle it or maybe @tombell has some ideas.
Ah, I see, thanks. Everybody here seems to expect some pretty high-level answers from hubot so I hooked him into a pandorabot. The only part I can't seem to figure out is how to limit the response to a hubot-script XOR the pandorabot. I'll put the script on github when I can get this part working.
bah, sorry slipped. :/
I couldn't see an easy fix for this in robots.coffee---maybe when I'm a little more comfortable with coffeescript & nodejs I'll look at it again. I aliased him to hubie for the time being:
https://gist.github.com/1851316
Hubot> hubie what's your name?
Hubot> My name is Hubot.
Hubot> hubie where do you come from?
Hubot> I am originally from Pennsylvania. Now I live in here. Where are you?
Hubot> hubie I'm from Canada.
Hubot> What is it like growing up there? Tell me a little about your life in Canada.
Hubot> hubie normally it's cold as hell.
Hubot> Are they exactly the same?
Hubot> hubie, hell and cold? I suppose it's not quite the same.
Hubot> I call it "Hades". Ambiguous: cold weather or cold sickness? That seems a bit negative.
Hubot> hubie where are you right now?
Hubot> At this moment I am in A.L.I.C.E. AI Foundation's computer in here. Where are you?
|
gharchive/issue
| 2012-02-16T22:23:02 |
2025-04-01T04:34:21.983073
|
{
"authors": [
"atmos",
"mikebridge"
],
"repo": "github/hubot",
"url": "https://github.com/github/hubot/issues/245",
"license": "mit",
"license_type": "permissive",
"license_source": "bigquery"
}
|
394529710
|
GDScript highlighting breaks on signal token
GDScript view is breaking with this script:
Preliminary Steps
Please confirm you have...
[ ] reviewed How Linguist Works,
[ ] reviewed the Troubleshooting docs,
[ ] considered implementing an override,
[ ] verified an issue has not already been logged for your issue (linguist issues).
Problem Description
I do not know the location that describes the syntax for gdscript, since the atom has no native support for this language.
since the atom has no native support for this language.
Which atom are you talking about?
I think you should read the troubleshooting docs.
The CSharp coloration comes from an atom library, however there is no native coloring library for GDScript syntax, only on github. I can not find the repository to which I can report this bug.
From the Troubleshooting docs:
Linguist detects the language of a file but the actual syntax-highlighting is powered by a set of language grammars which are included in this project as a set of submodules as listed here.
Is the wording unclear? Should we improve it?
|
gharchive/issue
| 2018-12-28T02:24:07 |
2025-04-01T04:34:21.988724
|
{
"authors": [
"pchaigno",
"schweigert"
],
"repo": "github/linguist",
"url": "https://github.com/github/linguist/issues/4364",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
}
|
536470346
|
FR: Script that generates "open" colors
Preliminary Steps
Please confirm you have...
[x] reviewed How Linguist Works,
[x] reviewed the Troubleshooting docs,
[x] considered implementing an override,
[x] verified an issue has not already been logged for your issue (linguist issues).
Problem Description
When trying to add a language, it is rather difficult to find a color that isn't in close proximity to other colors. It would be nice if there was a script you could run that would generate a bunch of valid colors that aren't close to other ones, based on the color algorithm used and the current colors used in language.yml.
Aren't there still a very large number of possible available colors?
Perhaps users of the script should be prompted for a seed color, and the script will try to get as close as possible.
Alternatively, without a seed color, perhaps the script should find the color at the farthest distance from any other color. If there are multiple good choices, use a pseudo-random number generator to choose one arbitrarily.
Another possibility -- have a script that would generate (on build, both locally and on GitHub) a color map image showing where all the currently used language colors are located, and also the areas which are too close to existing colors. It would then be trivial for contributors to open this image, find an open area by hand, and select a color from there..
I like the color map idea - that would be a fun project.
re: possible colors, I reckon in the "color space" there are, but it's not obvious how to find them...I tried several greens, blues, yellows, and red, and kept on getting rejected - which is what inspired this FR...
Ideally, there shouldn't be any need for this. See #4506; I've been pressing @lildude to press the other GitHub staff for a consensus on whether the colour proximity can be removed (as it was obviated long ago when dividers were added to the language bar).
Regarding the script, both @lildude and I had the same idea in the past. However, it's actually not that easy, because the colour proximity calculation is complex and non-linear (as in, it's not a simple matter of comparing RGB values).
Yes, this is incredibly hard and unfeasible to script for several reasons:
The CIEDE2000 color-difference formula is pretty complex
Assuming we scan through the entire valid HTML colour list, the resulting list of colours will vary based on where you start. If you start at #000000 and iterate up to #FFFFFF you'll get a different set of results from starting at #FFFFFF and iterating down to #000000, or starting in the middle and iteration outwards.
My initial attempt using an iterative method took several hours to do on my MacBook Pro in just one direction.
If someone finds a colour that doesn't clash and adds it to Linguist without consulting the previously determined list, that list would become invalid and would need regenerating.
A predetermined list isn't really much better than the current colour proximity enforcement.
So ultimately, the solution is to remove this, but I'm still waiting on the 👍 from our design team. As soon as I get it, it'll be removed within minutes.
2\. Assuming we scan through the entire valid HTML colour list, the resulting list of colours will vary based on where you start. If you start at `#000000` and iterate up to `#FFFFFF` you'll get a different set of results from starting at `#FFFFFF` and iterating down to `#000000`, or starting in the middle and iteration outwards.
I'm not sure I understand that point. Is CIEDE2000(color1, color2) different from CIEDE2000(color2, color1)?
I would expect that we "just" need to compute the distance for all colours and all languages with a color assigned, so 16M times a few hundreds computations.
I was referring to starting with an array of colours we know are in use and then use the color-proximity gem to iterate through all the free colours. If there's no clash, add it to the array and try the next considering that colour is now in use. This is probably far from the best way of doing it, but it was the way I initially tested this idea a long time ago.
Another approach might be a randomized algorithm; start with just choosing 1000 random colors in the RGB space and testing those all against proximity, as soon as you get ~20 colors or so, generate an HTML and open it in the users browser so they can pick one.
you could get more clever, and divide the RGB space into say 512 cubes (8x8x8), calculate the number of colors extant in each mini-cube, then start your search in the cubes with the fewest number of extant colors. I know it's not the same as the color difference algorithm but it might be close enough to work.
for someone submitting a new language, they just need to pick a color, so if we could generate even like 20 colors that will work that will save time.
|
gharchive/issue
| 2019-12-11T16:10:22 |
2025-04-01T04:34:21.999425
|
{
"authors": [
"Alhadis",
"karlwbrown",
"lildude",
"pchaigno",
"zspitz"
],
"repo": "github/linguist",
"url": "https://github.com/github/linguist/issues/4743",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
}
|
362914970
|
Fix to analysis GROUP BY
Context: https://github.com/github/orchestrator/issues/631
This PR fixes the analysis query, that has an invalid GROUP BY term. While the fix was easy, I'm puzzled at how this went under the radar with CI, and in particular with SQLite. I would have assumed it would have been caught earlier. I'll look into that.
cc @sharad-jha
Turns out SQLite, like MySQL without ONLY_FULL_GROUP_BY, allows invalid GROUP BY statements.
Work in progress to run ONLY_FULL_GROUP_BY sql_mode on CI MySQL.
|
gharchive/pull-request
| 2018-09-23T05:22:21 |
2025-04-01T04:34:22.001967
|
{
"authors": [
"shlomi-noach"
],
"repo": "github/orchestrator",
"url": "https://github.com/github/orchestrator/pull/632",
"license": "apache-2.0",
"license_type": "permissive",
"license_source": "bigquery"
}
|
827934566
|
Dark mode theme GA
Summary
GitHub's dark mode is generally available to all customers with a seamless dark mode experience on github.com and GitHub Enterprise Cloud.
Intended Outcome
Whether as a personal preference or as a way to reduce eye strain in low light conditions, dark mode gives developers more choice and flexibility in how and when they use GitHub.
How will it work?
From the appearance settings page, a user is able to choose between default to system (matching OS preferences), light mode or dark mode for their preferred GitHub theme setting.
https://github.blog/changelog/2021-04-14-dark-and-dimmed-themes-are-now-generally-available/
|
gharchive/issue
| 2021-03-10T15:42:44 |
2025-04-01T04:34:22.005042
|
{
"authors": [
"Sid-ah",
"github-product-roadmap"
],
"repo": "github/roadmap",
"url": "https://github.com/github/roadmap/issues/177",
"license": "CC-BY-4.0",
"license_type": "permissive",
"license_source": "github-api"
}
|
243866332
|
github-terms-of-service D: Locally condition grants on existing license
With 7a72ae0a (#1), the ToS grew new language about conditionally granting licenses in section D.3. However, the text in D.4 and D.5 still read as if the license grant was unconditional. This commit adjusts sections D.4 and D.5 to clearly separate the permissions needed (which are always true) from the ToS grant (which is only needed if the existing license doesn't already grant those permissions). Also
Define a new term, “Other Users” so we can compactly refer to other users without ambiguity.
Adjust the Your Content criteria. For example, if you fork another user's repository, the forked repository is Your Content, just like it would be if you cloned the other user's repository and then created a new GitHub repository seeded from that local clone. This clarifies the case where Alice creates the content and Bob uploads it to GitHub (it would be Your Content for Bob, but not for Alice unless she took other action).
Trim D.3 to cover termination and payment of any permissions granted in the following D.*. The old “we need you to grant us” wasn't conditional, so now that's gone. The old “unless other Users have forked it” is no longer covered in the ToS (more on this below). The remaining lines removed from D.3 are now covered in more detail in their specific sections.
In D.4, remove “to do things like host Your Content, publish it, and share it”, which are all covered in more detail in the subsequent section.
In D.4, add a new paragraph explicitly granting GitHub the required permissions if and only if your existing license doesn't already. This section doesn't address things like fair use, where GitHub might have permission to parse and share search results even in the absence of an explicit license, but I couldn't think of a concise way to cover that.
In D.5, I've removed the focus on public content, since any other user who is authorized to view the content requires this permission, regardless of whether the content is public.
In D.5, I've removed the need for fork permission. Users can view your content and decide for themselves whether they're allowed to fork it. If they fork content that you do not consider forkable, you can file a DMCA takedown request. The old “unless other Users have forked it” from D.3 is no longer needed because the forked content is no longer Your Content (see the earlier points about the Your Content criteria).
Related discussion in #53, #52, #38, #37, and #7.
Adjust the Your Content criteria. For example, if you fork another user's repository, the forked repository is Your Content
This would reopen #8 which was fixed only because §D.7 only applies to “Your Content”. I was being told that it’s deliberate that both “Your Content” and §D.7 only apply to code which uploader can grant the waiver on.
I've just addressed this in https://github.com/github/site-policy/issues/7#issuecomment-319542309. I believe that will resolve the issues here. Closing this for now, but will refer back to this when drafting the new guidelines.
On Wed, Aug 02, 2017 at 01:32:26AM +0000, Hannah Poteat wrote:
I've just addressed this in https://github.com/github/site-policy/issues/7#issuecomment-319542309.
The help article you propose there sounds nice, but note that 1 has:
These Terms of Service, together with the GitHub Privacy Statement,
represent the complete and exclusive statement of the agreement
between you and us. This Agreement supersedes any proposal or prior
agreement oral or written, and any other communications between you
and GitHub relating to the subject matter of these terms.
So if your help article says something like “projects released under
the GPL-3.0 need no additional rights granted to GitHub or Other
Users” (or whatever you end up saying), I still think you'd want to
improve the ToS wording to avoid the currently unqualified 2:
You grant us and our legal successors the right to store and display
your Content and make incidental copies as necessary to render the
Website and provide the Service.
and similar.
Also note that the current ToS currently require you to grant other
users the right to fork 3. And while that may be something that
GitHub intentionally requires you to grant other users, it limits what
people can host on GitHub (e.g. you can't host a public repository
that is “all rights reserved” while only granting other users view
permission 4).
And I think the current “Your Content” still needs clarification,
although I've also documented that in 5.
|
gharchive/pull-request
| 2017-07-18T22:08:46 |
2025-04-01T04:34:22.042529
|
{
"authors": [
"mirabilos",
"nsqe",
"wking"
],
"repo": "github/site-policy",
"url": "https://github.com/github/site-policy/pull/54",
"license": "CC0-1.0",
"license_type": "permissive",
"license_source": "github-api"
}
|
110065413
|
add-collaborators tweaks
A couple of tweaks to add-collaborators after using it in a particularly busy class last week:
Grabs a hash of current collaborators before working through issue comments: skips anyone who is already a collaborator
Old version would still post to the collaborators API for every comment in the issue
If there are 100s of comments, this should be considerably faster
Posts a comment as teacher at the end of the script with foo, bar, baz are now repository collaborators
should cut down on student confusion - they should now get a notification & see a comment when they've been added
One or two other minor things while I was in there. (mostly abort instead of just puts on error...)
@loranallensmith @brntbeer @patrickmckenna any feedback here? I've done some testing and confirmed it "works on my machine" -- got a minute to try this out?
@jaw6 :+1: to all of this.
Also, works on my machine, too.
@jaw6 This is so :cool:. Thanks for making these updates. I just ran this and it appears to be working on my machine.
:tada::balloon:
Just wanted to give another :+1: to this type of refinement. I'm not sure if any other @github/training-teachers have experienced the same thing I have, but students regularly ask me during class if I would mind sharing how we do "tricks" like this. Utilitarian as its roots may be, it's a really slick tool to have. Students love knowing that they have such an accessible platform for automation at their disposal.
:muscle: @jaw6 and @patrickmckenna
@jaw6 this is amazing. Looks great. Particularly like the use of Aborts instead of puts! :wink:
@jaw6 I'm late to the party but this is so awesome! Thank you for making these updates.
|
gharchive/pull-request
| 2015-10-06T18:00:46 |
2025-04-01T04:34:22.048681
|
{
"authors": [
"brntbeer",
"crichID",
"jaw6",
"loranallensmith",
"patrickmckenna"
],
"repo": "github/training-utils",
"url": "https://github.com/github/training-utils/pull/6",
"license": "mit",
"license_type": "permissive",
"license_source": "bigquery"
}
|
228771979
|
MukthaNarayan Recommendations
Adding my recommendations for things to do at hometown:
- [ ] Branch 1
- [ ] Adding a file to test Issue #1
- [ ] Commit
- [ ] Initiate a Pull Request #1
- [ ] Request review
- [ ] Make additional changes to PR #1
- [ ] Get approval
- [ ] Merge PR #1
This exercise is been completed.
|
gharchive/issue
| 2017-05-15T16:31:17 |
2025-04-01T04:34:22.055199
|
{
"authors": [
"MukthaNarayan"
],
"repo": "githubschool/automatic-waffle",
"url": "https://github.com/githubschool/automatic-waffle/issues/10",
"license": "CC-BY-4.0",
"license_type": "permissive",
"license_source": "github-api"
}
|
866393181
|
In step 3, Add headers, states click on the pencil icon to edit the file when no pencil icon is shown (only ... )
Bug Report
Current behavior
Step 3 instruction asks us to click on the pencil icon which does not exist as seen in these screenshots:
Reproduction
Steps to reproduce the behavior in the course:
Go to Step 3 Add headers
Follow the Activity instructions near the bottom to edit your file with headers
After clicking on the 'Files changed' tab...
There will be no pencil icon to click.
Expected behavior
I expect that the lesson will match the instructions and I become disorientated and discouraged when not matched.
Possible solution
Re-word the "in the upper right corner..." to say "click either the pencil or ... (hamburger) for the..."
Additional context
When a final change decision is made, I would like to make the change and commit to my repository then create a pull request to yourselves. If not, then go-ahead and make the changes yourself.
Well, I keep busy and just could not wait for a discussion to occur or asking permission to make a change. So, I boldly went ahead and made the changes and submitted a pull request. This is after all one of my learning experiences. And, now I wait to see what response occurs. Thanks for this opportunity.
The changes I made were approved and I closed the pull request and closed this issue also. If there is anything I missed let me know as I am still learning the cycles.
|
gharchive/issue
| 2021-04-23T19:59:17 |
2025-04-01T04:34:22.063542
|
{
"authors": [
"UNIVAC-Coder"
],
"repo": "githubtraining/communicating-using-markdown",
"url": "https://github.com/githubtraining/communicating-using-markdown/issues/33",
"license": "CC-BY-4.0",
"license_type": "permissive",
"license_source": "github-api"
}
|
137307195
|
Verificar versão do gitpie
Olá, senti falta de uma opção para verificar qual a versão que instalei do gitpie. não encontrei esse recurso no software.
abs.
Obrigado, melhoria será publicada no próximo hotfix.
|
gharchive/issue
| 2016-02-29T16:50:30 |
2025-04-01T04:34:22.097804
|
{
"authors": [
"lvlofenix",
"mapaiva"
],
"repo": "gitpie/GitPie",
"url": "https://github.com/gitpie/GitPie/issues/38",
"license": "mit",
"license_type": "permissive",
"license_source": "bigquery"
}
|
811883849
|
GitPitch Is Shutting Down on March 1, 2021
Hello,
The documentation says:
GitPitch is shutting down on March 1, 2021. The trial software is no longer available for download.
What does it mean? Is it about the desktop or the online service?
Hi Royi,
What does it mean?
I'm shutting down the business on March 1, 2021. That means the end. No more service or software.
So both the online and the desktop application won't be available anymore, right?
Too bad, it is a great service!
So both the online and the desktop application won't be available anymore, right?
Exactly right.
Cheers, David.
|
gharchive/issue
| 2021-02-19T10:05:08 |
2025-04-01T04:34:22.100505
|
{
"authors": [
"RoyiAvital",
"gitpitch"
],
"repo": "gitpitch/gitpitch",
"url": "https://github.com/gitpitch/gitpitch/issues/328",
"license": "mit",
"license_type": "permissive",
"license_source": "bigquery"
}
|
1505588184
|
ui and nuxt export different version of NavDrawer
I recently have the opportunity to use the ui package and I found that some props doesn't work as expected while using NavDrawer from @gits-id/ui package. Meanwhile, in another project that uses the same component from @gits-id/nuxt with the same props applied, works as expected (per documentation and storybook example).
Diving into the code, it seems that these package export different version of NavDrawer as VNavDrawer. Below are the export code in ui and nuxt:
https://github.com/gitsindonesia/ui-component/blob/cda5923c77eea564f1df3696753e7b21a5da97c5/packages/nuxt/src/module.ts#L266-L270
https://github.com/gitsindonesia/ui-component/blob/cda5923c77eea564f1df3696753e7b21a5da97c5/packages/ui/src/index.ts#L31
https://github.com/gitsindonesia/ui-component/blob/cda5923c77eea564f1df3696753e7b21a5da97c5/packages/ui/src/index.ts#L96
And this is how NavDrawer actually export its components
https://github.com/gitsindonesia/ui-component/blob/cda5923c77eea564f1df3696753e7b21a5da97c5/packages/nav-drawer/src/index.ts#L1-L3
Based on the above findings, I assume that the reason why NavDrawer behave differently between ui and nuxt is because the one exported by ui is the deprecated one which by storybook example does not implement some props that currently exist in not-deprecated one.
Hi. Yes it is intended for preventing breaking change. For older user will still use the deprecated version, while the new user like in the nuxt starter will use the new component.
Okay, so it's advisable for newer user to export NavDrawer from gits-id/nav-drawer then?
If so, maybe there should be a note in ui package, because the docs doesn't exactly have a page dedicated for the old version of NavDrawer nor point out which props are available from specific ver. onwards. Or, maybe in upgrade path or something. Just a section to tell how user can use the new version of certain components while waiting for new major ver. release to have this as default would be helpful.
Yes. It's recommended to use the new NavDrawer component. The old one could be removed on the new major version.
import {NavDrawer} from '@gits-id/nav-drawer`;
I already document and add deprecation info on the old nav-drawer page: https://gitsindonesia.github.io/ui-component/components/nav-drawer.html
The new NavDrawer doc is here: https://gitsindonesia.github.io/ui-component/components/navigation-drawer.html
My bad, somehow my eyes skipped that entry. Thanks for the link, I've read the deprecated one now. I know of the new one, but I guess my eyes glazed over when I see the deprecated ones.
I'll close this issue and remove the bug tag as it wasn't a bug.
|
gharchive/issue
| 2022-12-21T01:57:11 |
2025-04-01T04:34:22.224621
|
{
"authors": [
"gits-gretchelin",
"gravitano"
],
"repo": "gitsindonesia/ui-component",
"url": "https://github.com/gitsindonesia/ui-component/issues/72",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
}
|
2151166179
|
🛑 IceGiant-Node is down
In ed05b6e, IceGiant-Node (https://tendermint.icegiant.xyz) was down:
HTTP code: 502
Response time: 110 ms
Resolved: IceGiant-Node is back up in 8b02b14 after 14 minutes.
|
gharchive/issue
| 2024-02-23T14:06:07 |
2025-04-01T04:34:22.228437
|
{
"authors": [
"gitsrc"
],
"repo": "gitsrc/upptime",
"url": "https://github.com/gitsrc/upptime/issues/1700",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
}
|
1353879475
|
🛑 MSP-103 is down
In 1ea1ed1, MSP-103 (https://msp-portal.gw103.oneitfarm.com/) was down:
HTTP code: 0
Response time: 0 ms
Resolved: MSP-103 is back up in 4abaa17.
|
gharchive/issue
| 2022-08-29T08:08:47 |
2025-04-01T04:34:22.231146
|
{
"authors": [
"gitsrc"
],
"repo": "gitsrc/upptime",
"url": "https://github.com/gitsrc/upptime/issues/815",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
}
|
252016535
|
Add support for custom string tables
It is currently possible to customize the bundle used to localize string values, but not the string table itself. In addition to calling bundleForStrings, LMViewBuilder should also look for a tableForStrings method on the document owner. This method would return the name of the string table to use when localizing resources, and could be used with or without bundleForStrings.
This is done.
|
gharchive/issue
| 2017-08-22T16:29:30 |
2025-04-01T04:34:22.293997
|
{
"authors": [
"gk-brown"
],
"repo": "gk-brown/MarkupKit",
"url": "https://github.com/gk-brown/MarkupKit/issues/102",
"license": "apache-2.0",
"license_type": "permissive",
"license_source": "bigquery"
}
|
57970943
|
Allows use of tspans with transitions along a path
@gka
This should allow for the tspan values to update when for example transitioning along a path a la Mike Bostock's example http://bl.ocks.org/mbostock/3902569.
I don't think the tspans code needs to be modified for this to work - try saving a reference to the text element and translating it.
|
gharchive/pull-request
| 2015-02-17T19:38:09 |
2025-04-01T04:34:22.295418
|
{
"authors": [
"1wheel",
"lfarrell"
],
"repo": "gka/d3-jetpack",
"url": "https://github.com/gka/d3-jetpack/pull/3",
"license": "bsd-3-clause",
"license_type": "permissive",
"license_source": "bigquery"
}
|
133975082
|
windows build for 0.6.1
I created 4 builds with the latest master....so other people could try it too
i changed the version to 0.6.1
Ok, thanks very much. But I'd rather include the builds with the release, not in the repo itself. That makes it much easier to sync with current development. I'll tag the 0.6.1 release today and include the prebuilt extensions for Windows.
I've created the release for 0.6.1 and included your builds (see Release)
|
gharchive/pull-request
| 2016-02-16T12:57:29 |
2025-04-01T04:34:22.303744
|
{
"authors": [
"ThaDafinser",
"gkralik"
],
"repo": "gkralik/php7-sapnwrfc",
"url": "https://github.com/gkralik/php7-sapnwrfc/pull/7",
"license": "mit",
"license_type": "permissive",
"license_source": "bigquery"
}
|
1064506528
|
feat - fix navbar + add languages
What
Slightly improve navbar + add language dropdown.
Not perfect but I had too many constrains from the original theme. I tried not to edit original files (we might have conflicts in the future) - this is the reason I created extra files.
I deployed theme to prod: https://insights.glassnode.com/ (almost no visible diff)
Other improvements:
fix dark mode toggle (localstorage stores values as string)
fix ::selection in light/dark mode
other fixes / overrides to make it work a bit better
Looks good 👍
|
gharchive/pull-request
| 2021-11-26T13:59:46 |
2025-04-01T04:34:22.319919
|
{
"authors": [
"Whyseverythingtaken",
"rafalbromirski"
],
"repo": "glassnode/Casper",
"url": "https://github.com/glassnode/Casper/pull/2",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
}
|
1925457232
|
🛑 Schoelcher is down
In 5db75a5, Schoelcher (https://www.mairie-schoelcher.fr) was down:
HTTP code: 0
Response time: 0 ms
Resolved: Schoelcher is back up in 78c6796 after 13 minutes.
|
gharchive/issue
| 2023-10-04T05:54:58 |
2025-04-01T04:34:22.331243
|
{
"authors": [
"glefait"
],
"repo": "glefait/martinique-public-websites",
"url": "https://github.com/glefait/martinique-public-websites/issues/13551",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
}
|
1971864644
|
🛑 Schoelcher is down
In e5e2c04, Schoelcher (https://www.mairie-schoelcher.fr) was down:
HTTP code: 0
Response time: 0 ms
Resolved: Schoelcher is back up in 4ef5bd7 after 10 minutes.
|
gharchive/issue
| 2023-11-01T07:39:08 |
2025-04-01T04:34:22.333618
|
{
"authors": [
"glefait"
],
"repo": "glefait/martinique-public-websites",
"url": "https://github.com/glefait/martinique-public-websites/issues/14616",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
}
|
2145031758
|
🛑 Saint-Pierre is down
In 4e8dea8, Saint-Pierre (https://saintpierre-mq.fr) was down:
HTTP code: 0
Response time: 0 ms
Resolved: Saint-Pierre is back up in b455d95 after 5 minutes.
|
gharchive/issue
| 2024-02-20T18:33:26 |
2025-04-01T04:34:22.336134
|
{
"authors": [
"glefait"
],
"repo": "glefait/martinique-public-websites",
"url": "https://github.com/glefait/martinique-public-websites/issues/16285",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
}
|
2151520995
|
🛑 Saint-Pierre is down
In 92bfa31, Saint-Pierre (https://saintpierre-mq.fr) was down:
HTTP code: 0
Response time: 0 ms
Resolved: Saint-Pierre is back up in 70c097e after 10 minutes.
|
gharchive/issue
| 2024-02-23T17:35:50 |
2025-04-01T04:34:22.338436
|
{
"authors": [
"glefait"
],
"repo": "glefait/martinique-public-websites",
"url": "https://github.com/glefait/martinique-public-websites/issues/16443",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
}
|
1425265195
|
🛑 Schoelcher is down
In da790bd, Schoelcher (https://www.mairie-schoelcher.fr) was down:
HTTP code: 0
Response time: 0 ms
Resolved: Schoelcher is back up in 92d0132.
|
gharchive/issue
| 2022-10-27T08:48:59 |
2025-04-01T04:34:22.340668
|
{
"authors": [
"glefait"
],
"repo": "glefait/martinique-public-websites",
"url": "https://github.com/glefait/martinique-public-websites/issues/1732",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
}
|
2181680127
|
🛑 Saint-Pierre is down
In 7cb71d0, Saint-Pierre (https://saintpierre-mq.fr) was down:
HTTP code: 0
Response time: 0 ms
Resolved: Saint-Pierre is back up in 1d95061 after 11 minutes.
|
gharchive/issue
| 2024-03-12T13:52:07 |
2025-04-01T04:34:22.342986
|
{
"authors": [
"glefait"
],
"repo": "glefait/martinique-public-websites",
"url": "https://github.com/glefait/martinique-public-websites/issues/17348",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
}
|
1287439249
|
🛑 Schoelcher is down
In fe20bbb, Schoelcher (https://www.mairie-schoelcher.fr) was down:
HTTP code: 0
Response time: 0 ms
Resolved: Schoelcher is back up in 9d331b2.
|
gharchive/issue
| 2022-06-28T14:29:08 |
2025-04-01T04:34:22.345265
|
{
"authors": [
"glefait"
],
"repo": "glefait/martinique-public-websites",
"url": "https://github.com/glefait/martinique-public-websites/issues/452",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
}
|
1565814990
|
🛑 Schoelcher is down
In cb4462c, Schoelcher (https://www.mairie-schoelcher.fr) was down:
HTTP code: 0
Response time: 0 ms
Resolved: Schoelcher is back up in 72f18ae.
|
gharchive/issue
| 2023-02-01T10:05:15 |
2025-04-01T04:34:22.347736
|
{
"authors": [
"glefait"
],
"repo": "glefait/martinique-public-websites",
"url": "https://github.com/glefait/martinique-public-websites/issues/4670",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
}
|
1633922903
|
🛑 Schoelcher is down
In d1aea67, Schoelcher (https://www.mairie-schoelcher.fr) was down:
HTTP code: 0
Response time: 0 ms
Resolved: Schoelcher is back up in 5315d9a.
|
gharchive/issue
| 2023-03-21T13:25:03 |
2025-04-01T04:34:22.350026
|
{
"authors": [
"glefait"
],
"repo": "glefait/martinique-public-websites",
"url": "https://github.com/glefait/martinique-public-websites/issues/6458",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
}
|
1322181860
|
🛑 Le Lorrain is down
In 6cc3181, Le Lorrain (https://villedulorrain.com) was down:
HTTP code: 0
Response time: 0 ms
Resolved: Le Lorrain is back up in e47d199.
|
gharchive/issue
| 2022-07-29T12:21:46 |
2025-04-01T04:34:22.352300
|
{
"authors": [
"glefait"
],
"repo": "glefait/martinique-public-websites",
"url": "https://github.com/glefait/martinique-public-websites/issues/731",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
}
|
1691459660
|
🛑 Biodiversite Martinique is down
In cba0714, Biodiversite Martinique (http://www.biodiversite-martinique.fr) was down:
HTTP code: 503
Response time: 576 ms
Resolved: Biodiversite Martinique is back up in 972a724.
|
gharchive/issue
| 2023-05-01T21:55:13 |
2025-04-01T04:34:22.354617
|
{
"authors": [
"glefait"
],
"repo": "glefait/martinique-public-websites",
"url": "https://github.com/glefait/martinique-public-websites/issues/8244",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
}
|
89528554
|
OpenCL object definitions should be fixed at compile time
Right now, the kernel/objects.cl file contains definitions for objects that have to be equal to the corresponding definitions in the Lensed sources. It would thus make more sense to hard-code the OpenCL definitions at compile time.
Ideally, the same bit of code could be used in the C and OpenCL codes. Maybe this can be done by including the same file, if either all compilers understand the same syntax for struct packing, or suitable macros can be defined.
PR #182 fixed this whole problematic part of the code in a much better way. Closing this.
|
gharchive/issue
| 2015-06-19T10:30:35 |
2025-04-01T04:34:22.356286
|
{
"authors": [
"ntessore"
],
"repo": "glenco/lensed",
"url": "https://github.com/glenco/lensed/issues/179",
"license": "mit",
"license_type": "permissive",
"license_source": "bigquery"
}
|
416640403
|
Flatlist inside Gesture recognizer not working on iOS
When I set the flat list inside the gesture recognizer, It works fine on android but flatlist is not able to scroll on iOS..Any help please!
I'm having the same issue!
Same here.
When making ScrollView a child of a GestureRecognizer this causes the same problem. The answer for that was to switch the order, ie. make the GestureRecognizer a child of the ScrollView. How can we do the same thing for FlatList?
Same issue here. I need GestureRecognizer to be the parent. There is a ScrollView as a children somewhere down the tree, but it doesn't work. My scenario is not hard, I need GesturaRecognizer to listen to right and left swipes but ignore up and down and leave it to the ScrollView to handle those naturally. Any ideas?
There is a solution on another Issue: https://github.com/glepur/react-native-swipe-gestures/issues/16#issuecomment-344181328
Same issue coming at my side. Any idea?
@thegamenicorus has a solution here: #16 (comment)
Doesn't help any more(
For anyone looking out for a solution, you can try this library: https://www.npmjs.com/package/rn-swipe-gestures
It's basically a clone of react-native-swipe-gestures. The only difference is you can add props to the config such as detectSwipeUp, detectSwipeDown, etc. This will solve your problem.
Guys,
I have inserted GestureRecognizer INSIDE renderItem of Flatlist, and it worked both in Android and iOS. Try it~
renderItem = (data) => {
const gestureConfig = {
velocityThreshold: 0.3,
directionalOffsetThreshold: 80
};
return (
<GestureRecognizer
onSwipeLeft={(state) => this.onSwipeLeft(state)}
onSwipeRight={(state) => this.onSwipeRight(state)}
config={gestureConfig}
style={{flex: 1}}
>
test
)
}
For anyone looking out for a solution, you can try this library: https://www.npmjs.com/package/rn-swipe-gestures It's basically a clone of react-native-swipe-gestures. The only difference is you can add props to the config such as detectSwipeUp, detectSwipeDown, etc. This will solve your problem.
I installed this library and the interface for the GestureRecognizerConfig didn't contain those properties to detect swiping up or down...
|
gharchive/issue
| 2019-03-04T05:43:23 |
2025-04-01T04:34:22.368936
|
{
"authors": [
"HisonKahng",
"Manish123Pandia",
"RazaShehryar",
"Vladislava9009",
"bobber205",
"cs-manughian",
"hamzaehsan95",
"matheusbaumgart",
"olinations"
],
"repo": "glepur/react-native-swipe-gestures",
"url": "https://github.com/glepur/react-native-swipe-gestures/issues/31",
"license": "mit",
"license_type": "permissive",
"license_source": "bigquery"
}
|
138753709
|
Add support for wgl window transparency (Vista and later)
This method uses desktop composition for efficient window transparency.
Since windows 8, composition is always on. If composition is disabled on vista or 7, transparency is simply ignored. Also, since it requires a pixel format supporting compositing, if we can't find it (e.g. windows xp) we ignore transparency similarly to how @datenwolf is doing it for glx's frame buffer configs.
This shouldn't be tested in a virtual machine.
On a side note, desktop composition can be enabled in the "Performance Options" dialog (win+r systempropertiesperformance). It is on by default but won't work unless the Aero score index has been refreshed after installing gpu drivers (Control Panel\All Control Panel Items\Performance Information and Tools).
I tested it on Windows Vista, 7, 8, 8.1 and 10.
It ins't so straight forward to achieve transparency with opengl on Windows but I hope that you will find that this is heading to the right direction. This method has quirks which can all be resolved with hacks but by comparison to other methods, this method is the fastest. The glReadPixels or BitBlt methods to achieve transparency use a lot of cpu when fullscreen vs almost idle with Dwm (bellow 1%).
It is worth also noting that there is a color key method on layered windows. This would require a different type of glfw hint because all pixels painted by the window in this color will be transparent and forces the user to not use that color in his rendering.
Thank you!
@uglycoder good catch! I did purposely use SetWindowLongPtr but I overlooked the Get version.
A quick search shows that we previously used GetWindowLongPtrW in the windowProc callback to retrieve the _GLFWwindow pointer. For the sake of consistency, maybe we could use GetWindowLongPtrW and SetWindowLongPtrW everywhere.
I checked out 43f116f and used cmake -G "Visual Studio 12 2013 Win64" .. to create the solution and then tried to build with VS2013 - got these errors:
Anyone else sees this? Or have I borked my build?
I removed the inline keyword of isWindowsVersionOrGreater and then it built fine. I removed it based on a hint from ReSharper++.
Once built it's working fine on my Win10 machine: (nVidia GeForce GTX 750)
I'll try on Win7 later.
|
gharchive/pull-request
| 2016-03-06T04:26:29 |
2025-04-01T04:34:22.375382
|
{
"authors": [
"christopherp",
"elmindreda",
"thomthom"
],
"repo": "glfw/glfw",
"url": "https://github.com/glfw/glfw/pull/723",
"license": "Zlib",
"license_type": "permissive",
"license_source": "github-api"
}
|
216780884
|
Add .babelrc support
Uses the .babelrc in the consuming app if available.
LGTM - thanks @robbiepitts!
|
gharchive/pull-request
| 2017-03-24T13:10:33 |
2025-04-01T04:34:22.379928
|
{
"authors": [
"dgeb",
"robbiepitts"
],
"repo": "glimmerjs/glimmer-application-pipeline",
"url": "https://github.com/glimmerjs/glimmer-application-pipeline/pull/13",
"license": "mit",
"license_type": "permissive",
"license_source": "bigquery"
}
|
2156668441
|
🛑 INTG-APP is down
In 1e6b1d4, INTG-APP ($SITE_6) was down:
HTTP code: 403
Response time: 244 ms
Resolved: INTG-APP is back up in d3c06bb after 19 minutes.
|
gharchive/issue
| 2024-02-27T14:00:25 |
2025-04-01T04:34:22.399037
|
{
"authors": [
"glo-markoaular"
],
"repo": "glo-markoaular/status-navigate",
"url": "https://github.com/glo-markoaular/status-navigate/issues/1179",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
}
|
2214933947
|
🛑 INTG-APP is down
In 352a56d, INTG-APP ($SITE_6) was down:
HTTP code: 403
Response time: 108 ms
Resolved: INTG-APP is back up in 3e5be36 after 31 minutes.
|
gharchive/issue
| 2024-03-29T08:40:54 |
2025-04-01T04:34:22.401400
|
{
"authors": [
"glo-markoaular"
],
"repo": "glo-markoaular/status-navigate",
"url": "https://github.com/glo-markoaular/status-navigate/issues/1475",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
}
|
1378740018
|
🛑 CMS is down
In 30a55ca, CMS ($SITE_4) was down:
HTTP code: 0
Response time: 0 ms
Resolved: CMS is back up in 2409886.
|
gharchive/issue
| 2022-09-20T02:44:27 |
2025-04-01T04:34:22.403453
|
{
"authors": [
"glo-markoaular"
],
"repo": "glo-markoaular/status-navigate",
"url": "https://github.com/glo-markoaular/status-navigate/issues/346",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
}
|
1504094881
|
Update osmnx save_graph_shapefile() calls to use save_graph_geopackage()
On the new enhancements branch we are using newer osmnx, which has some changed methods and some flagged as being deprecated --- this is one of those, so will be good to update while we're modifying things. A nice to have.
/env/lib/python3.9/site-packages/osmnx/io.py:87: UserWarning: The `save_graph_shapefile` function is deprecated and will be removed in a future release. Instead, use the `save_graph_geopackage` function to save graphs as GeoPackage files for subsequent GIS analysis.
Yes OSMnx is going this route: http://switchfromshapefile.org/
I was just looking at implementing this, and reading the OSMnx feature request thread at https://github.com/gboeing/osmnx/issues/411#issuecomment-608125305 it occurred to me that we could bypass the gpkg (and the shapefiles; happy to be done with them!) and write directly to postgis, adapting Geoff's example.
if not (db_contents.has_table('nodes') and db_contents.has_table('edges')):
print("\nPrepare and copy nodes and edges to postgis... ")
nodes, edges = ox.graph_to_gdfs(G)
with engine.begin() as connection:
nodes.to_postgis('nodes',connection,index=True)
edges.to_postgis('edges',connection,index=True)
This simplifies the code considerably, makes the processing flow more logical, and should save time with reads and writes to files. I've tested on one study region, and will try on others shortly.
One difference i noticed in the postgis version loaded in this way was that it was missing the 'from' and 'to' columns. These were present in the shape files, and the geopackage version.
Looking at previous versions of edges and nodes that had these fields, as well as the u,v,key fields they are apparently derived from it seems that
"from" is v and "to is u if key == 0 or key ==1
"from" is u and "to" is v if key == 2
So, I'll recreate these fields as part of the network processing based on the above pattern.
|
gharchive/issue
| 2022-12-20T07:15:40 |
2025-04-01T04:34:22.409409
|
{
"authors": [
"carlhiggs",
"gboeing"
],
"repo": "global-healthy-liveable-cities/global-indicators",
"url": "https://github.com/global-healthy-liveable-cities/global-indicators/issues/171",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
}
|
2562648901
|
Fix: README
New url
@dmvieira
I was a maintainer, but I left globo.com...
|
gharchive/pull-request
| 2024-10-02T21:13:08 |
2025-04-01T04:34:22.427996
|
{
"authors": [
"dmvieira",
"fabiohcnobre"
],
"repo": "globocom/ghdfs",
"url": "https://github.com/globocom/ghdfs/pull/17",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
}
|
1726973205
|
is this lossy conversion?
anyone know how Glomatico's ✨ Apple Music ✨ Downloader works? so far I have learned that it download the fragment and decrypt it, but does decrypting the file but does it convert the file to m4a afterwards? since that would be a lossy to lossy converting which will lower the music quality, I have tried downloading the same link of a song from there and compared it with a itune matched song, the audio stream are different. So it would be nice if someone could explain to me how it works and why the differences in audio stream.
flac? 48/96/192 Hi-Res???? 🤔
Flac uses a different DRM! There is no public method for ripping flac.
|
gharchive/issue
| 2023-05-26T06:15:10 |
2025-04-01T04:34:22.429444
|
{
"authors": [
"Rafawell",
"Rainrainraind",
"futaride"
],
"repo": "glomatico/gamdl",
"url": "https://github.com/glomatico/gamdl/issues/33",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
}
|
1770682316
|
🛑 Rezepte is down
In 3e3a064, Rezepte ($REZEPTE_URL) was down:
HTTP code: 0
Response time: 0 ms
Resolved: Rezepte is back up in 2a9070a.
|
gharchive/issue
| 2023-06-23T02:34:54 |
2025-04-01T04:34:22.437409
|
{
"authors": [
"gloriousDan"
],
"repo": "gloriousDan/Homelab-uptime",
"url": "https://github.com/gloriousDan/Homelab-uptime/issues/47",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
}
|
1837813828
|
🛑 Nextcloud is down
In 090ee6c, Nextcloud ($NEXTCLOUD_URL) was down:
HTTP code: 0
Response time: 0 ms
Resolved: Nextcloud is back up in cfbb4c1.
|
gharchive/issue
| 2023-08-05T15:38:26 |
2025-04-01T04:34:22.439454
|
{
"authors": [
"gloriousDan"
],
"repo": "gloriousDan/Homelab-uptime",
"url": "https://github.com/gloriousDan/Homelab-uptime/issues/59",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
}
|
2309487261
|
Nativewind TextInput and TextArea does not apply custom colors defined from config
Description
TextInput and Texarea does not apply custom colors from nativewind defined in config
CodeSandbox/Snack link
https://github.com/gluestack/ui-example-nativewind
Steps to reproduce
No response
gluestack-ui Version
nativewind
Platform
[X] Expo
[ ] React Native CLI
[ ] Next
[ ] Web
[ ] Android
[ ] iOS
Other Platform
No response
Additional Information
Issue is re producable even in example nativewind demo (look at search text input in dark mode it's ignoring the text-typography-900)
Hi @Nathan-Charles,
Thanks for reporting this issue. We are working on it and will inform you once it is resolved.
Thank you.
This issue is now resolved, so I am closing it.
|
gharchive/issue
| 2024-05-22T03:19:49 |
2025-04-01T04:34:22.661909
|
{
"authors": [
"nathan-charles",
"rajat693"
],
"repo": "gluestack/gluestack-ui",
"url": "https://github.com/gluestack/gluestack-ui/issues/2176",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
}
|
1876373744
|
cargo install gluon_repl fails to compile
When I try to run cargo install gluon_repl it fails to install due to:
error[E0308]: mismatched types
--> ~/.cargo/registry/src/index.crates.io-6f17d22bba15001f/gluon_base-0.18.0/src/types/mod.rs:1042:24
|
1042 | const _: [(); 8 * 6] = [(); std::mem::size_of::<Type<Symbol, ArcType>>()];
| ----- ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ expected an array with a fixed size of 48 elements, found one with 40 elements
| |
| help: consider specifying the actual array length: `40`
Compiling rand_core v0.6.4
Compiling num_cpus v1.16.0
For more information about this error, try `rustc --explain E0308`.
error: could not compile `gluon_base` (lib) due to previous error
warning: build failed, waiting for other jobs to finish...
error: failed to compile `gluon_repl v0.18.0`, intermediate artifacts can be found at `/tmp/cargo-installSWuQke`
I've removed the other package compilation messages to focus on the error message
Closed by #959 . Thanks!
|
gharchive/issue
| 2023-08-31T22:04:59 |
2025-04-01T04:34:22.663680
|
{
"authors": [
"Marwes",
"rowanfr"
],
"repo": "gluon-lang/gluon",
"url": "https://github.com/gluon-lang/gluon/issues/957",
"license": "mit",
"license_type": "permissive",
"license_source": "bigquery"
}
|
443059122
|
关于 Class 'think\facade\Env' not found
你好，我使用
createTable
初始化数据库表，提示报错：
Class 'think\facade\Env' not found
我使用的是 thinkphp5.024
这个是类库是不是仅支持 5.1呀？我是需要升级5.1才能解决这个问题嘛？
是的 5.0.*没有Evn可以找到包的CreateTable.php文件找到env的位置改为自己runtime的路径
|
gharchive/issue
| 2019-05-12T03:12:47 |
2025-04-01T04:34:22.760855
|
{
"authors": [
"gmars",
"nongzhenli"
],
"repo": "gmars/tp5-rbac",
"url": "https://github.com/gmars/tp5-rbac/issues/11",
"license": "mit",
"license_type": "permissive",
"license_source": "bigquery"
}
|
2610116355
|
Nombres de microservicios irregulares
En la rama de cada modulo tienen que ponerse de acuerdo como van a nombrar los nombre de las carpetas, porque por ejemplo en el modulo de gestion se divide en servicios, en el de votYVal se dividen en submodulos, y el de visualizacion ni siquiera esta en su propia carpeta.
Asi que cada uno tendria que corregir en su rama eso y despues hacer un PR a development
@Billones142 porfa revisá el merge #13 y cerrá este issue si corresponde
Me parece correcto el cambio, cierro el issue.
|
gharchive/issue
| 2024-10-24T00:42:20 |
2025-04-01T04:34:22.762443
|
{
"authors": [
"Billones142",
"gmartineza"
],
"repo": "gmartineza/bienal-backend",
"url": "https://github.com/gmartineza/bienal-backend/issues/10",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
}
|
1405243511
|
Added initial documentation for command specific details.
This document will also be used for describing outputs for other commands, including install, cache, etc. For now it only includes list.
Everything looks good.
|
gharchive/pull-request
| 2022-10-11T21:22:25 |
2025-04-01T04:34:22.766321
|
{
"authors": [
"gmisail",
"meowox"
],
"repo": "gmisail/fontman-client",
"url": "https://github.com/gmisail/fontman-client/pull/19",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
}
|
677686762
|
Cancelled order "untilBatchId" is off by 1
Cancelled orders untilBatchId is off by one. Consider the order ID 0x7d82569f9871726c3259aac2e6057a745c70dc8c-0. Doing a graph query to get the untilBatchId returns 5324121.
graph query
query {
order(id: "0x7d82569f9871726c3259aac2e6057a745c70dc8c-0") {
untilBatchId
}
}
{
"data": {
"order": {
"untilBatchId": "5324121"
}
}
}
However, when reading the smart contract, the value is 5324120
truffle console
$ npx truffle console --network mainnet
truffle(mainnet)> const exchange = await BatchExchange.deployed()
truffle(mainnet)> const encodedOrders = await exchange.getEncodedUserOrders("0x7d82569f9871726c3259aac2e6057a745c70dc8c")
truffle(mainnet)> encodedOrders
'0x7d82569f9871726c3259aac2e6057a745c70dc8c00000000000000000000000000000000000000000000003635c9adc5dea000000017000000513d5700513d580000000000000000133f7b17b563fbd8000000000000003635c9adc5dea00000000000000000003635c9adc5dea000007d82569f9871726c3259aac2e6057a745c70dc8c00000000000000000000000000000000000000000000003635c9adc5dea000000007000000513d5a00513d66000000000000003c3bc3a4a2f75c7158000000000000003635c9adc5dea00000000000000000003635c9adc5dea00000'
truffle(mainnet)> parseInt(encodedOrders.substr(2 + (20 /*address*/ + 32 /*balance*/ + 2 /*buytoken*/ + 2 /*selltoken*/ + 4 /*validfrom*/) * 2, 8), 16)
5324120
FIxed, the query now returns:
{
"data": {
"order": {
"untilBatchId": "5324120"
}
}
}
|
gharchive/issue
| 2020-08-12T13:17:52 |
2025-04-01T04:34:22.834279
|
{
"authors": [
"nlordell"
],
"repo": "gnosis/dex-subgraph",
"url": "https://github.com/gnosis/dex-subgraph/issues/70",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
}
|
378646935
|
Hotfix/v0.7.11
Check markets before selling after receiving EVENT_AUCTION_CLEARED
Pull Request Test Coverage Report for Build 1050
5 of 5 (100.0%) changed or added relevant lines in 1 file are covered.
No unchanged relevant lines lost coverage.
Overall coverage increased (+0.08%) to 48.076%
Totals
Change from base Build 1042:
0.08%
Covered Lines:
1742
Relevant Lines:
3389
💛 - Coveralls
@anxolin Yes, I assert that no calls are done to ensureLiquidity functions (I'm checking bot function and service funcion calls)
|
gharchive/pull-request
| 2018-11-08T09:32:30 |
2025-04-01T04:34:22.839367
|
{
"authors": [
"coveralls",
"dasanra"
],
"repo": "gnosis/dx-services",
"url": "https://github.com/gnosis/dx-services/pull/29",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
}
|
835869917
|
Must set objective=surplus when calling http solver.
Since the default is to optimize for volume, and we're already ranking solutions based on surplus.
Does this task still require a change?
Currently we're optimizing for volume since optimizing for surplus seems less robust. Once we're confident in the solver results when optimizing for surplus, we should reopen this issue.
|
gharchive/issue
| 2021-03-19T11:21:20 |
2025-04-01T04:34:22.840702
|
{
"authors": [
"fleupold",
"marcovc"
],
"repo": "gnosis/gp-v2-services",
"url": "https://github.com/gnosis/gp-v2-services/issues/400",
"license": "Apache-2.0",
"license_type": "permissive",
"license_source": "github-api"
}
|
997140278
|
[L2-UX] Fix multiple clicks on switching network
What it solves
Resolves #2724. PR #2734 has to be merged first.
How this PR fixes it
Fixes onboard initialization. It also updates onboard to the latest version so it fixes the repeated polling when switching networks.
Add some improvements to recover connection used last time on each network
How to test it
Try switching networks.
Check that if Metamask was used last time connections is automatically recovered
Check that connection with Ledger/Trezor is possible in those compatible networks
Clicking multiple times in switch network doesn't trigger a continuous error, just once and show onboard modal second time (this is good to allow the user switch wallet in case wallet is not able to change network)
Screenshots
I guess we can merge it straight to l2-ux and test it there later.
I guess we can merge it straight to l2-ux and test it there later.
@katspaugh Yes, I think the same. It will avoid overhead to QA as this bug is detected in an integration branch, so will be tested during regression
|
gharchive/pull-request
| 2021-09-15T14:16:25 |
2025-04-01T04:34:22.866708
|
{
"authors": [
"dasanra",
"katspaugh"
],
"repo": "gnosis/safe-react",
"url": "https://github.com/gnosis/safe-react/pull/2735",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
}
|
2319693958
|
🛑 Beacon RPC is down
In 575a33b, Beacon RPC (https://rpc-gbc.gnosischain.com/eth/v1/beacon/headers) was down:
HTTP code: 0
Response time: 0 ms
Resolved: Beacon RPC is back up in cda0e96 after 5 minutes.
|
gharchive/issue
| 2024-05-27T19:39:31 |
2025-04-01T04:34:22.869269
|
{
"authors": [
"gnosis-deployer"
],
"repo": "gnosischain/gnosischain-uptime",
"url": "https://github.com/gnosischain/gnosischain-uptime/issues/982",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
}
|
789012134
|
[doc] unclear guidance on the https server setup
In the documentation (https://docs.gitea.io/en-us/https-setup/), it is said that the configuration should look like this to get https working:
[server]
PROTOCOL = https
ROOT_URL = https://git.example.com:3000/
HTTP_PORT = 3000
CERT_FILE = cert.pem
KEY_FILE = key.pem
Then it took me quite a while to get it working until I realized that the values for the CERT_FILE and KEY_FILE should be absolute paths instead of the literal relative path.
whould this had helped you ?
-CERT_FILE = cert.pem
-KEY_FILE = key.pem
+CERT_FILE = /path/cert.pem
+KEY_FILE = /path/key.pem
without absolute it's relative to $CUSTOM_ROOT ...
whould this had helped you ?
-CERT_FILE = cert.pem
-KEY_FILE = key.pem
+CERT_FILE = /path/cert.pem
+KEY_FILE = /path/key.pem
without absolute it's relative to $CUSTOM_ROOT ...
@6543 Yes. I think that /path/ hint is enough to help people understand it correctly. And probably CUSTOM_ROOT should be documented somewhere as well.
@6543 Yes. I think that /path/ hint is enough to help people understand it correctly. And probably CUSTOM_ROOT should be documented somewhere as well.
I have to correct myselve ... it's called GITEA_CUSTOM ... @cdluminate would you mind to create/propose a pull for the docs?
(It's generated by hugo, you find the files at /docs/
I have to correct myselve ... it's called GITEA_CUSTOM ... @cdluminate would you mind to create/propose a pull for the docs?
(It's generated by hugo, you find the files at /docs/
|
gharchive/issue
| 2021-01-19T13:13:38 |
2025-04-01T04:34:22.898037
|
{
"authors": [
"6543",
"cdluminate"
],
"repo": "go-gitea/gitea",
"url": "https://github.com/go-gitea/gitea/issues/14401",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
}
|
1082828142
|
Rename a repository twice, then the first rename action target became the latest name.
At first, rename repository A -> B, the action in dashboard displayed "xxx renamed A -> B".
Then rename B -> C, now there are two actions, the first one is "xxx renamed A -> C", the latest one is "xxx renamed B -> C".
I think because when rename, the target repository stored the repository id in the table but not the target name.
Hello, I have managed to fix this issue, but it manifests itself not only in renaming repo, but on all actions. For example, when we create issue (issue created in A) and rename repo (A->B), action with issu contains newest repo name (issue created in B), so as action with creation repo.
So do we need to replace ShortRepoPath call for new function in any place for each type of action in feed?
We can also rewrite ShortRepoPath but is it proper to all over cases?
I think yes, we need store the current name of that time in to database, maybe a new column?(need to check). So if add a new column, we need a migration. But the migration cannot do the right thing, because we have lost them. We can make future things right only.
I fix this issue without changing database. We can restore actual name of repo by following actions. So the question is should I change ShortRepoPath and change behavior for all dependencies or fix it only for feed by adding new function special for it?
|
gharchive/issue
| 2021-12-17T02:46:56 |
2025-04-01T04:34:22.901170
|
{
"authors": [
"Viktor-Yakovchuk",
"lunny"
],
"repo": "go-gitea/gitea",
"url": "https://github.com/go-gitea/gitea/issues/18007",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
}
|
2136552961
|
Upgrade from 1.18.5 to 1.21.5 in Docker
Description
DSM Synology, Docker. Previous container 1.18.5 switched-off. New container 1.21.5 switch-on, start Migrate(), recreate all my data base, so I can't use 1.18.5 container any more. Look like no errors was in migration process. Trying login in - got back to start page.
Gitea Version
1.21.5
Can you reproduce the bug on the Gitea demo site?
Yes
Log Gist
No response
Screenshots
No response
Git Version
No response
Operating System
No response
How are you running Gitea?
in Docker container in DSM Synology
Database
None
by the way is it doing backup of old data-base somewhere ? is it possible roll back some how? like partly desigion of this problem
The same with 1.21.4
Login and pass is correct - mirror project is logined and fo POST GET correctly according logs.
by the way is it doing backup of old data-base somewhere ?
No auto-backup from Gitea side. If your DSM has some backups, you could try.
Trying login in - got back to start page.
The information is not enough for debugging a problem. Maybe you need to check your logs first. And figure out why "got back to start page": eg: cookie doesn't work? login failed? etc
We close issues that need feedback from the author if there were no new comments for a month. :tea:
|
gharchive/issue
| 2024-02-15T13:39:16 |
2025-04-01T04:34:22.906484
|
{
"authors": [
"GiteaBot",
"VitaliyAT",
"wxiaoguang"
],
"repo": "go-gitea/gitea",
"url": "https://github.com/go-gitea/gitea/issues/29180",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
}
|
454451566
|
500 error when PR comment
Gitea version (or commit ref):1.8.2
Git version:2.17.1
Operating system:Ubuntu 18.04.2 LTS
Database (use [x]):
[ ] PostgreSQL
[x] MySQL
[ ] MSSQL
[ ] SQLite
Log gist:
[...routers/repo/pull.go:962 CompareAndPullRequestPost()] [E] NewPullRequest: newIssue: Error 1366: Incorrect string value: '\xF0\x9F\x98\xAC' for column 'content' at row 1
This has been fixed on v1.9 to set database as CHARSET = utf8mb4.
|
gharchive/issue
| 2019-06-11T02:28:53 |
2025-04-01T04:34:22.909266
|
{
"authors": [
"FerminPan",
"lunny"
],
"repo": "go-gitea/gitea",
"url": "https://github.com/go-gitea/gitea/issues/7178",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
}
|
858736706
|
migration: github: if rate limit is not enabled, ignore it
as title
Is there any reference document?
sadly no: github api v3 documentation is removed more and more in favour of GraphQL.
this issue was discovered & testend as you can read in Discord(#general)
refs:
it was https://web.archive.org/web/20200929150611/https://developer.github.com/v3/#rate-limiting
now you have https://docs.github.com/en/rest/reference/rate-limit
|
gharchive/pull-request
| 2021-04-15T10:25:07 |
2025-04-01T04:34:22.911638
|
{
"authors": [
"6543",
"lunny"
],
"repo": "go-gitea/gitea",
"url": "https://github.com/go-gitea/gitea/pull/15490",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
}
|
1075839122
|
Prevent services/mailer/mailer_test.go tests from deleteing data directory
Running make test-backend will delete data/ due to reloading the configuration and resetting the appdatapath.
This PR removes this unnecessary config reload but also adds extra code in to the unittest main to prevent its cleanup from deleting the wrong directory.
Signed-off-by: Andrew Thornton art27@cantab.net
Codecov Report
:exclamation: No coverage uploaded for pull request base (main@c7e2340). Click here to learn what that means.
The diff coverage is n/a.
@@ Coverage Diff @@
## main #17941 +/- ##
=======================================
Coverage ? 45.37%
=======================================
Files ? 815
Lines ? 90779
Branches ? 0
=======================================
Hits ? 41187
Misses ? 43025
Partials ? 6567
Continue to review full report at Codecov.
Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update c7e2340...eff1779. Read the comment docs.
How can we run only one unit test correctly? eg: I want to run TestGenerateMessageID only when developing the code.
setting.LoadForTest(`
[mailer]
ENABLED = true
FROM = test@domain.com
`)
setting.NewServices()
without the initialization code, TestGenerateMessageID couldn't run individually on my side.
How can we run only one unit test correctly? eg: I want to run TestGenerateMessageID only when developing the code.
setting.LoadForTest(`
[mailer]
ENABLED = true
FROM = test@domain.com
`)
setting.NewServices()
without the initialization code, TestGenerateMessageID couldn't run individually on my side.
Done
|
gharchive/pull-request
| 2021-12-09T17:45:38 |
2025-04-01T04:34:22.919633
|
{
"authors": [
"codecov-commenter",
"wxiaoguang",
"zeripath"
],
"repo": "go-gitea/gitea",
"url": "https://github.com/go-gitea/gitea/pull/17941",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
}
|
248006134
|
Make short link pattern greedy
Multiple links on the same line are currently not recognized correctly.
To reproduce, simply write something like this: [[link1]] [[link2]]
unit tests for this would be nice
LGTM
LGTM
@lunny Should we move this PR to milestone v1.2?
If unit test is added than yes I think
@lafriks: I have added a unit test. Is this okay or do you expect something else/more?
@michaelkuhn yes excactly I was thinking about. Thank you
|
gharchive/pull-request
| 2017-08-04T13:28:24 |
2025-04-01T04:34:22.922144
|
{
"authors": [
"appleboy",
"ethantkoenig",
"lafriks",
"michaelkuhn"
],
"repo": "go-gitea/gitea",
"url": "https://github.com/go-gitea/gitea/pull/2259",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
}
|
521480899
|
Githook highlighter
Shell code highlighter for GitHook editor
Codecov Report
Merging #8932 into master will decrease coverage by <.01%.
The diff coverage is 0%.
@@ Coverage Diff @@
## master #8932 +/- ##
==========================================
- Coverage 41.24% 41.24% -0.01%
==========================================
Files 547 547
Lines 70764 70765 +1
==========================================
Hits 29184 29184
- Misses 37860 37862 +2
+ Partials 3720 3719 -1
Impacted Files
Coverage Δ
routers/repo/setting.go
13.28% <0%> (-0.02%)
:arrow_down:
modules/task/migrate.go
25% <0%> (-3.95%)
:arrow_down:
models/error.go
32.67% <0%> (-1.19%)
:arrow_down:
modules/log/event.go
64.61% <0%> (-1.03%)
:arrow_down:
models/repo.go
46.51% <0%> (+0.04%)
:arrow_up:
modules/migrations/gitea.go
8.68% <0%> (+0.63%)
:arrow_up:
modules/migrations/migrate.go
22.9% <0%> (+1.67%)
:arrow_up:
modules/indexer/indexer.go
55.26% <0%> (+10.52%)
:arrow_up:
Continue to review full report at Codecov.
Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update e4e6d48...12f0ea6. Read the comment docs.
Sigh... TestGitHubDownloadRepo is failing because someone touched one of the issues it looks at.
@Konctantin - I think we're gonna merge #8973 before this one which will almost certainly conflict with this. Sorry about this.
Ok, then after merge #8973 i will fix the conflict
Heya! Yeah it did conflict but if you can fix we should be able to get this merged this evening
Thanks
@Konctantin Thank you for your first contribution.
|
gharchive/pull-request
| 2019-11-12T11:03:16 |
2025-04-01T04:34:22.935470
|
{
"authors": [
"Konctantin",
"codecov-io",
"sapk",
"zeripath"
],
"repo": "go-gitea/gitea",
"url": "https://github.com/go-gitea/gitea/pull/8932",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
}
|
1417789274
|
feat: support diy TableName
[] Do only one thing
[] Non breaking API changes
[] Tested
What did this pull request do?
User Case Description
@qqxhb 可以提供一些关于如何diy tableName的示例代码吗
|
gharchive/pull-request
| 2022-10-21T05:56:23 |
2025-04-01T04:34:22.939420
|
{
"authors": [
"qqxhb",
"wulorn"
],
"repo": "go-gorm/gen",
"url": "https://github.com/go-gorm/gen/pull/687",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
}
|
654532931
|
hope to recognize the default values
Describe the feature
go语言里面的很多默认值十分频繁的涉及到常用的业务值，比如int类型的0，bool类型的false。如果gorm完全不识别这些默认值会导致很多地方必须要做出特殊的处理从而破坏代码可读性。采取指针传入的方式也会有此类问题，这会导致很多业务场景下为了配合gorm强行做的指针转换。
建议gorm识别基础类型的默认值，把对默认值的判断交给用户决定
https://v2.gorm.io/docs/update.html
|
gharchive/issue
| 2020-07-10T05:59:07 |
2025-04-01T04:34:22.940799
|
{
"authors": [
"jinzhu",
"synuwxy"
],
"repo": "go-gorm/gorm",
"url": "https://github.com/go-gorm/gorm/issues/3132",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
}
|
860881485
|
Why logger ExplainSQL parameter case fmt.Stringer,this makes the log different from the actual
Your Question
Why logger ExplainSQL parameter case fmt.Stringer,this makes the log different from the actual
The document you expected this should be explained
type Letter int
func (l Letter) String() string {
switch l{
case A:
return "A"
case B:
return "B"
default:
return "C"
}
}
const(
A Letter= 1+iota
B
C
)
type Foo struct {
Id int
Letter Letter
}
func TestLogger(t *testing.T) {
db:=gorm.DB{}
var foos []Foo
db.Where("letter=?",A).Find(&foos)
}
# log
SELECT * FROM foo WHERE letter = 'A';
# actual
SELECT * FROM foo WHERE letter = 1;
Expected answer
I don't understand why case fmt.Stringer here ，I want to make a PR to delete case fmt.Stringer,Is it possible？
@jinzhu @liov
I'm facing the problem that logging SQL doesn't match actual SQL.
It made my team confused.
May I make this issue again and fix it ?
|
gharchive/issue
| 2021-04-19T04:25:10 |
2025-04-01T04:34:22.943226
|
{
"authors": [
"liov",
"sawadyecma"
],
"repo": "go-gorm/gorm",
"url": "https://github.com/go-gorm/gorm/issues/4303",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
}
|
2299792516
|
Executing count will clear Statement.Joins
https://github.com/go-gorm/playground/pull/736
type AB struct {
Id int `json:"id" gorm:"primaryKey"`
BId int `json:"aId" gorm:"index"`
AId int `json:"bId"`
B B `json:"b"`
A A `json:"a"`
}
type A struct {
Id int `json:"id" gorm:"primaryKey"`
Name string `json:"name" gorm:"uniqueIndex"`
}
type B struct {
Id int `json:"id" gorm:"primaryKey"`
Name string `json:"name" gorm:"uniqueIndex"`
}
func main() {
db, _ := gorm.Open(tests.DummyDialector{}, &gorm.Config{
NamingStrategy: schema.NamingStrategy{
SingularTable: true,
},
})
var (
count int64
appNodes = []*AB{}
)
db:= db.Model(&AB{}).Joins("A").Joins("B").Where(`id = ?`, 1)
db.Count(&count)
db.Limit(1).Offset(0).Order("id DESC").Find(&appNodes)
}
db := db.Model(&AB{}).Preload("A").Preload("B").Where(`id = ?`, 1)
The two pieces of code have different results, preload will get the correct result, because Count will call AfterQuery,it will "clear the joins after query because preload need it" ,I don't know why did it, this means need to write Joins twice.
@liov What version are you using?
var (
count int64
appNodes = []*AB{}
)
db1 := db.Model(&AB{}).Joins("A").Joins("B").Where("abs.`id` = ?", 1)
// [0.572ms] [rows:1] SELECT count(*) FROM `abs` LEFT JOIN `as` `A` ON `abs`.`a_id` = `A`.`id` LEFT JOIN `bs` `B` ON `abs`.`b_id` = `B`.`id` WHERE abs.`id` = 1
db1.Debug().Count(&count)
// [0.705ms] [rows:0] SELECT `abs`.`id`,`abs`.`b_id`,`abs`.`a_id`,`A`.`id` AS `A__id`,`A`.`name` AS `A__name`,`B`.`id` AS `B__id`,`B`.`name` AS `B__name` FROM `abs` LEFT JOIN `as` `A` ON `abs`.`a_id` = `A`.`id` LEFT JOIN `bs` `B` ON `abs`.`b_id` = `B`.`id` WHERE abs.`id` = 1 ORDER BY abs.`id` DESC LIMIT 1
db1.Limit(1).Offset(0).Order("abs.`id` DESC").Debug().Find(&appNodes)
var ab AB
// [0.630ms] [rows:0] SELECT `abs`.`id`,`abs`.`b_id`,`abs`.`a_id`,`A`.`id` AS `A__id`,`A`.`name` AS `A__name`,`B`.`id` AS `B__id`,`B`.`name` AS `B__name` FROM `abs` LEFT JOIN `as` `A` ON `abs`.`a_id` = `A`.`id` LEFT JOIN `bs` `B` ON `abs`.`b_id` = `B`.`id` WHERE abs.`id` = 1 ORDER BY abs.`id` DESC LIMIT 1
db1.Model(&AB{}).Preload("A").Preload("B").Debug().Find(&ab)
I run it with version v1.25.9, and the two pieces of code have same results.
And testing with your playground (after fix your syntax error with the: Error 1052 (23000): Column 'id' in where clause is ambiguous), the two pieces of code have same results too (with both version v1.25.9 and your playground version v1.25.4)
// [rows:0] SELECT `abs`.`id`,`abs`.`b_id`,`abs`.`a_id`,`B`.`id` AS `B__id`,`B`.`name` AS `B__name` FROM `abs` LEFT JOIN `bs` `B` ON `abs`.`b_id` = `B`.`id` WHERE abs.id = 11
db = DB.Model(&AB{}).Preload("A").Joins("B").Debug().Where(`abs.id = ?`, 11).Find(&ab)
// [2.033ms] [rows:1] SELECT count(*) FROM `abs` LEFT JOIN `bs` `B` ON `abs`.`b_id` = `B`.`id` WHERE abs.id = 11
db.Debug().Count(&count)
// [1.860ms] [rows:0] SELECT `abs`.`id`,`abs`.`b_id`,`abs`.`a_id`,`B`.`id` AS `B__id`,`B`.`name` AS `B__name` FROM `abs` LEFT JOIN `bs` `B` ON `abs`.`b_id` = `B`.`id` WHERE abs.id = 11 ORDER BY abs.id DESC LIMIT 1
db.Limit(1).Offset(0).Order("abs.id DESC").Debug().Find(&appNodes)
@ivila I'm mistaken, actually it's right to clear joins here, but count affects the select, please try the latest https://github.com/go-gorm/playground/pull/736
I guess that's where the problem lies
for _, join := range db.Statement.Joins {
// xxx
if isRelations {
for _, s := range relation.FieldSchema.DBNames {
if v, ok := selectColumns[s]; (ok && v) || (!ok && !restricted) {
clauseSelect.Columns = append(clauseSelect.Columns, clause.Column{
Table: tableAliasName,
Name: s,
Alias: utils.NestedRelationName(tableAliasName, s),
})
}
}
}
}
Is anyone looking into this? This bug was introduced by the fix for this issue https://github.com/go-gorm/gorm/issues/6715. Now, I just need to call count in front, and it can be fully reproduced. https://github.com/go-gorm/playground/pull/736
2024/05/21 03:15:28 /home/runner/work/playground/playground/main_test.go:35
[0.313ms] [rows:1] SELECT count(*) FROM `users` LEFT JOIN `accounts` `Account` ON `users`.`id` = `Account`.`user_id` AND `Account`.`deleted_at` IS NULL LEFT JOIN `pets` `Account__Pet` ON `Account`.`id` = `Account__Pet`.`account_id` AND `Account__Pet`.`deleted_at` IS NULL WHERE `users`.`deleted_at` IS NULL
2024/05/21 03:15:28 /home/runner/work/playground/playground/main_test.go:39
[0.144ms] [rows:2] SELECT * FROM `companies` WHERE `companies`.`account_id` = 1
2024/05/21 03:15:28 /home/runner/work/playground/playground/main_test.go:39
[0.445ms] [rows:1] SELECT * FROM `accounts` WHERE `accounts`.`user_id` = 1 AND `accounts`.`deleted_at` IS NULL
2024/05/21 03:15:28 /home/runner/work/playground/playground/main_test.go:39
[0.859ms] [rows:1] SELECT `users`.`id`,`users`.`created_at`,`users`.`updated_at`,`users`.`deleted_at` FROM `users` LEFT JOIN `accounts` `Account` ON `users`.`id` = `Account`.`user_id` AND `Account`.`deleted_at` IS NULL LEFT JOIN `pets` `Account__Pet` ON `Account`.`id` = `Account__Pet`.`account_id` AND `Account__Pet`.`deleted_at` IS NULL WHERE `users`.`deleted_at` IS NULL AND `users`.`id` = 1 ORDER BY `users`.`id` LIMIT 1
main_test.go:48: Failed, got ''
Also, I have fixed it now. https://github.com/go-gorm/gorm/pull/7027
|
gharchive/issue
| 2024-05-16T09:02:10 |
2025-04-01T04:34:22.950250
|
{
"authors": [
"ivila",
"liov"
],
"repo": "go-gorm/gorm",
"url": "https://github.com/go-gorm/gorm/issues/7025",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
}
|
1792143249
|
Some staticcheck fixes
I ran staticcheck and fixed the following truncated list of warnings.
$ staticcheck
db.go:823:65: should use time.Since instead of time.Now().Sub (S1012)
gorp.go:40:5: var zeroVal is unused (U1000)
table.go:79:9: unnecessary use of fmt.Sprintf (S1039)
table.go:84:2: should replace loop with columns = append(columns, fieldNames...) (S1011)
Oh my god it keeps happening. I opened this in the wrong repo again.
|
gharchive/pull-request
| 2023-07-06T19:55:15 |
2025-04-01T04:34:22.951881
|
{
"authors": [
"pgporada"
],
"repo": "go-gorp/gorp",
"url": "https://github.com/go-gorp/gorp/pull/447",
"license": "mit",
"license_type": "permissive",
"license_source": "bigquery"
}
|
607094317
|
"generate operation" returns "no operations were selected"
Problem statement
$ ./swagger generate operation --name=
2020/04/26 20:25:14 validating spec /home/borisov/tmp/swagger/swagger.yml
2020/04/26 20:25:15 preprocessing spec with option: minimal flattening
no operations were selected
If path has operationId defined then nothing is generated. If that property is removed then code is generated correctly:
$ ./swagger generate operation --name=
2020/04/26 20:31:21 validating spec /home/borisov/tmp/swagger/swagger.yml
2020/04/26 20:31:21 preprocessing spec with option: minimal flattening
2020/04/26 20:31:21 rendering 4 templates for operation
2020/04/26 20:31:21 name field GetFoo
2020/04/26 20:31:21 package field operations
2020/04/26 20:31:21 creating generated file "get_foo_parameters.go" in "restapi/operations" as parameters
2020/04/26 20:31:21 executed template asset:serverParameter
2020/04/26 20:31:21 name field GetFoo
2020/04/26 20:31:21 package field operations
2020/04/26 20:31:21 creating generated file "get_foo_urlbuilder.go" in "restapi/operations" as urlbuilder
2020/04/26 20:31:21 executed template asset:serverUrlbuilder
2020/04/26 20:31:21 name field GetFoo
2020/04/26 20:31:21 package field operations
2020/04/26 20:31:21 creating generated file "get_foo_responses.go" in "restapi/operations" as responses
2020/04/26 20:31:21 executed template asset:serverResponses
2020/04/26 20:31:21 name field GetFoo
2020/04/26 20:31:21 package field operations
2020/04/26 20:31:21 creating generated file "get_foo.go" in "restapi/operations" as handler
2020/04/26 20:31:21 executed template asset:serverOperation
2020/04/26 20:31:21 Generation completed!
For this generation to compile you need to have some packages in your GOPATH:
* github.com/go-openapi/runtime
You can get these now with: go get -u -f ./...
Swagger specification
$ cat swagger.yml
consumes:
- application/json
info:
title: Swagger
version: 0.1.0
paths:
/foo:
get:
operationId: list
responses:
200:
description: Successful response
schema:
type: object
required:
- data
properties:
data:
type: string
produces:
- application/json
schemes:
- http
swagger: "2.0"
$ ./swagger validate swagger.yml
2020/04/26 20:22:27
The swagger spec at "swagger.yml" is valid against swagger specification 2.0
Environment
$ ./swagger version
version: v0.23.0
commit: 3596b40bedd4f06cc918767098568456d2ec2370
$ go version
go version go1.14.2 linux/amd64
Yes, empty string is not pruned from the list... Fixing that and a similar issue with models. Thanks for reporting.
|
gharchive/issue
| 2020-04-26T19:32:09 |
2025-04-01T04:34:23.018577
|
{
"authors": [
"borisovg",
"fredbi"
],
"repo": "go-swagger/go-swagger",
"url": "https://github.com/go-swagger/go-swagger/issues/2280",
"license": "Apache-2.0",
"license_type": "permissive",
"license_source": "github-api"
}
|
114575152
|
add FindMap
func (session *Session) FindMap(rowsSlicePtr interface{}, condiBean ...interface{}) error {
defer session.resetStatement()
if session.IsAutoClose {
defer session.Close()
}
sliceValue := reflect.Indirect(reflect.ValueOf(rowsSlicePtr))
if sliceValue.Kind() != reflect.Slice && sliceValue.Kind() != reflect.Map {
return errors.New("needs a pointer to a slice or a map")
}
sliceElementType := sliceValue.Type().Elem()
fmt.Println("sliceValue.Kind()", sliceValue.Kind(), sliceElementType)
fmt.Println("sliceValue.Kind()")
/*
if len(condiBean) > 0 {
colNames, args := buildConditions(session.Engine, table, condiBean[0], true, true,
false, true, session.Statement.allUseBool, session.Statement.useAllCols,
session.Statement.unscoped, session.Statement.mustColumnMap)
session.Statement.ConditionStr = strings.Join(colNames, " AND ")
session.Statement.BeanArgs = args
} else {
// !oinume! Add "<col> IS NULL" to WHERE whatever condiBean is given.
// See https://github.com/go-xorm/xorm/issues/179
if col := table.DeletedColumn(); col != nil && !session.Statement.unscoped { // tag "deleted" is enabled
session.Statement.ConditionStr = fmt.Sprintf("(%v IS NULL or %v = '0001-01-01 00:00:00') ",
session.Engine.Quote(col.Name), session.Engine.Quote(col.Name))
}
}*/
fmt.Println("sliceValue.Kind()")
var sqlStr string
var args []interface{}
if session.Statement.RawSQL == "" {
var columnStr string = session.Statement.ColumnStr
if session.Statement.JoinStr == "" {
if columnStr == "" {
if session.Statement.GroupByStr != "" {
columnStr = session.Statement.Engine.Quote(strings.Replace(session.Statement.GroupByStr, ",", session.Engine.Quote(","), -1))
} else {
columnStr = session.Statement.genColumnStr()
}
}
} else {
if columnStr == "" {
if session.Statement.GroupByStr != "" {
columnStr = session.Statement.Engine.Quote(strings.Replace(session.Statement.GroupByStr, ",", session.Engine.Quote(","), -1))
} else {
columnStr = "*"
}
}
}
session.Statement.attachInSql()
sqlStr = session.Statement.genSelectSql(columnStr)
args = append(session.Statement.Params, session.Statement.BeanArgs...)
// for mssql and use limit
qs := strings.Count(sqlStr, "?")
if len(args)*2 == qs {
args = append(args, args...)
}
} else {
sqlStr = session.Statement.RawSQL
args = session.Statement.RawParams
}
fmt.Println("sliceValue.Kind()")
// var err error
/*
if session.Statement.JoinStr == "" {
if cacher := session.Engine.getCacher2(table); cacher != nil &&
session.Statement.UseCache &&
!session.Statement.IsDistinct &&
!session.Statement.unscoped {
err = session.cacheFind(sliceElementType, sqlStr, rowsSlicePtr, args...)
if err != ErrCacheFailed {
return err
}
err = nil // !nashtsai! reset err to nil for ErrCacheFailed
session.Engine.LogWarn("Cache Find Failed")
}
}
*/
fmt.Println("sliceValue.Kind()", sliceValue.Kind())
if sliceValue.Kind() != reflect.Map {
fmt.Println("sliceValue.Type()", sliceValue.Index(0).Type(), reflect.TypeOf(make(map[string]interface{})).Name())
if sliceValue.Index(0).Type() == reflect.TypeOf(make(map[string][]byte)) {
fmt.Println("sliceValue.Type()OK")
resultsSlice, err := session.query(sqlStr, args...)
if err != nil {
fmt.Println("sliceValue.Type()err", err.Error())
return err
}
for _, results := range resultsSlice {
fmt.Println("sliceValue.Type()OK", results)
sliceValue.Set(reflect.Append(sliceValue, reflect.Indirect(reflect.ValueOf(results))))
}
} else {
fmt.Println("sliceValue.Index(0).Type() == reflect.TypeOf(make(map[string]interface{}))")
}
} else {
fmt.Println("sliceValue.Kind() != reflect.Map")
}
return nil
}
见 #339
看代码没看明白。
就是Find 支持传递Map[string]interface{} 进去取值
最新版本已经支持https://github.com/go-xorm/tests/blob/master/testFind.go#L278
|
gharchive/issue
| 2015-11-02T11:19:06 |
2025-04-01T04:34:23.029787
|
{
"authors": [
"hzmnet",
"lunny"
],
"repo": "go-xorm/xorm",
"url": "https://github.com/go-xorm/xorm/issues/313",
"license": "bsd-3-clause",
"license_type": "permissive",
"license_source": "bigquery"
}
|
241757368
|
More clear log
I don't see nil and empty strings in log. I suggest the next changes
Before
INSERT INTO `dish` (`dish_id`,`restaurant_id`,`name`,`source`,`entry_id`,`created_by`,`menu_id`,`category`,`description`,`price`,`score`,`flog_count`,`approve`,`activated`,`created_at`,`updatated_at`) VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?) [59639fb56d9d9437b707a434 58a313719fee7b336da70c45 Chocolate Mousse Torte fsqi 51339970 Dessert Seasonal berry sauce 0 0 0 1 false 2017-07-10 18:39:33 2017-07-10 18:39:33]
After
INSERT INTO `dish` (`dish_id`,`restaurant_id`,`name`,`source`,`entry_id`,`created_by`,`menu_id`,`category`,`description`,`price`,`score`,`flog_count`,`approve`,`activated`,`created_at`,`updatated_at`) VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?) []interface {}{"59639ec16d9d943782ce2f6c", "58a313719fee7b336da70c45", "Chocolate Mousse Torte", "fsqi", "51339970", "", "", "Dessert", "Seasonal berry sauce", 0, 0, 0, 1, false, "2017-07-10 18:35:29", "2017-07-10 18:35:29"}
Codecov Report
Merging #638 into master will not change coverage.
The diff coverage is 100%.
@@ Coverage Diff @@
## master #638 +/- ##
=======================================
Coverage 49.53% 49.53%
=======================================
Files 33 33
Lines 7039 7039
=======================================
Hits 3487 3487
Misses 3077 3077
Partials 475 475
Impacted Files
Coverage Δ
engine.go
56.74% <100%> (ø)
:arrow_up:
xorm.go
69.49% <0%> (ø)
:arrow_up:
Continue to review full report at Codecov.
Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 39a812d...abe8474. Read the comment docs.
Thanks!
|
gharchive/pull-request
| 2017-07-10T15:41:07 |
2025-04-01T04:34:23.038506
|
{
"authors": [
"codecov-io",
"lunny",
"oherych"
],
"repo": "go-xorm/xorm",
"url": "https://github.com/go-xorm/xorm/pull/638",
"license": "bsd-3-clause",
"license_type": "permissive",
"license_source": "bigquery"
}
|
269885267
|
Add PingContext support for go1.8+
As title.
Codecov Report
Merging #766 into master will decrease coverage by 0.06%.
The diff coverage is 0%.
@@ Coverage Diff @@
## master #766 +/- ##
==========================================
- Coverage 51.38% 51.32% -0.07%
==========================================
Files 38 39 +1
Lines 8547 8557 +10
==========================================
Hits 4392 4392
- Misses 3670 3680 +10
Partials 485 485
Impacted Files
Coverage Δ
context.go
0% <0%> (ø)
xorm.go
66.19% <0%> (ø)
:arrow_up:
Continue to review full report at Codecov.
Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 865979f...9ec1968. Read the comment docs.
|
gharchive/pull-request
| 2017-10-31T09:13:08 |
2025-04-01T04:34:23.045569
|
{
"authors": [
"codecov-io",
"lunny"
],
"repo": "go-xorm/xorm",
"url": "https://github.com/go-xorm/xorm/pull/766",
"license": "bsd-3-clause",
"license_type": "permissive",
"license_source": "bigquery"
}
|
785048075
|
Update dependency zf-commons/zfc-user
Please update dependency zf-commons/zfc-user to fresh versions (3.0.0).
maybe other also.
thanks
???
|
gharchive/issue
| 2021-01-13T12:05:33 |
2025-04-01T04:34:23.051630
|
{
"authors": [
"lon9man"
],
"repo": "goalio/GoalioForgotPassword",
"url": "https://github.com/goalio/GoalioForgotPassword/issues/50",
"license": "bsd-3-clause",
"license_type": "permissive",
"license_source": "bigquery"
}
|
829414863
|
Async Mode: context deadline exceeded (Client.Timeout exceeded while awaiting headers)
When using Colly with Async turned off, the crawler works as expected - though slow as expected, about one item per second. However, enabling Async mode quickly turns into a barrage of
context deadline exceeded (Client.Timeout exceeded while awaiting headers)
this is fairly instant, within the first millisecond's Colly parses a few tens of instances, after which the timeout barrage is initiated.
I'm not sure if this generates the problem, but I am appending the parsed Struct from the main Collector's response to a child Collector via context:
items = append(items, item)
r.Request.Ctx.Put("item", &items[len(items)-1])
r.Request.Visit(nextLink)
Could this be a throttling problem, with the server timing out requests? How can I test/prevent it?
Current settings:
c.SetRequestTimeout(25 * time.Second)
c.Limit(&colly.LimitRule{
Parallelism: 2,
Delay: 1 * time.Second,
RandomDelay: 5 * time.Second,
})
What is strange is that even with the 25-second timeout rule, the error starts 3 or 4 seconds into the crawl start.
Hi @pdavis156879 ,
First you should Request to the secondary collector passing your context, as recommended in Colly's doc:
r.Ctx.Put("item", &listings[len(items)-1])
r2.Request("GET", nextLink, nil, r.Ctx, nil)
Secondly, try the combination of:
Increasing the request timeout to a very large value, e.g. 100
Increase both the request delay and random delay
|
gharchive/issue
| 2021-03-11T18:27:28 |
2025-04-01T04:34:23.125223
|
{
"authors": [
"Guilherme-B",
"pdavis156879"
],
"repo": "gocolly/colly",
"url": "https://github.com/gocolly/colly/issues/593",
"license": "apache-2.0",
"license_type": "permissive",
"license_source": "bigquery"
}
|
169897619
|
Implement "My Assets" page
Currently it's just a redirect to the main frontend page.
No junior job tag?
Now yes :P
|
gharchive/issue
| 2016-08-08T11:06:58 |
2025-04-01T04:34:23.136239
|
{
"authors": [
"akien-mga",
"bojidar-bg"
],
"repo": "godotengine/asset-library",
"url": "https://github.com/godotengine/asset-library/issues/18",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
}
|
256429671
|
File heat names in Script Editor are very hard to read.
Duplicate of #11059
Ups.
Opening again. Deleted editor settings tres and created new project with new .godot file and colors didn't changed.
Can't reproduce. Deleting ~/.godot/editor_settings-3.tres is sufficient for me. Is the engine open while you delete the file?
@Noshyaar
Derp. <(^___^)" Thx. I probably had opened Editor...
|
gharchive/issue
| 2017-09-09T10:14:43 |
2025-04-01T04:34:23.159170
|
{
"authors": [
"Noshyaar",
"n-pigeon"
],
"repo": "godotengine/godot",
"url": "https://github.com/godotengine/godot/issues/11095",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
}
|
353546205
|
\n and \t doesn't work in controls, unless if set via GDScript
\n and \t doesn't work in Label, Button, etc. when set via the editor. It does work though via GDScript.
Looks like this has been lost on the way:
https://github.com/godotengine/godot/issues/2967#issuecomment-210496149
That's a duplicate of #2967, no?
Oh, it's still open, my bad, I thought it was closed.
|
gharchive/issue
| 2018-08-23T20:44:12 |
2025-04-01T04:34:23.161164
|
{
"authors": [
"akien-mga",
"char0xff"
],
"repo": "godotengine/godot",
"url": "https://github.com/godotengine/godot/issues/21334",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
}
|
392176212
|
GDScript: floating-point calculation inconsistent
Godot version:
Godot Engine v3.0.6.stable.official.8314054 (installed from Steam)
OS/device including version:
Windows 10 version 1803 (OS internal version 17134.471)
Issue description:
The gdcript calculation leads to inconsistent result.
Steps to reproduce:
Minimal reproduction project:
var v1 = Vector2(309.063385, 558.316772)
var v2 = Vector2(308.78009, 550.69281)
var v3 = Vector2(309.601715, 565.934143)
var a = Basis(
Vector3(v1.x, v1.y, 1),
Vector3(v2.x, v2.y, 1),
Vector3(v3.x, v3.y, 1)
).determinant()
var b = v1.x * (v2.y - v3.y) - v2.x * (v1.y - v3.y) + v3.x * (v1.y - v2.y)
print("a: ", a, " b: ", b)
Output:
a: 1.95459 b: 1.946248
The variables a and b should be the same. I realized that their real types are float, but they still should be the same.
Then, I calculated b in c++ using float, and it's
b: 1.94604
So both the results of gdscript are wrong. Did I miss something?
I calculated b using your script in Python and it's...
>>> v1x = 309.063385
>>> v1y = 558.316772
>>> v2x = 308.78009
>>> v2y = 550.69281
>>> v3x = 309.601715
>>> v3y = 565.934143
>>> b = v1x * (v2y - v3y) - v2x * (v1y - v3y) + v3x * (v1y - v2y)
>>> b
1.9462443460151917
After looking into the source code (version 3.0.6.stable), I found that real number (TYPE_REAL) is calculated as double (b in this case), while Vector2, Vector3, Basis are calculated as float(a in this case).
But float calculation still couldn't lead to a. It's off too much.
Difference is caused by swapped rows/cols in Basis class.
#include <cstdio>
int main() {
float v1x = 309.063385;
float v1y = 558.316772;
float v1z = 1.0;
float v2x = 308.780090;
float v2y = 550.692810;
float v2z = 1.0;
float v3x = 309.601715;
float v3y = 565.934143;
float v3z = 1.0;
//your calculation
float det1 = v1x * (v2y * v3z - v3y * v2z) - v2x * (v1y * v3z - v3y * v1z) + v3x * (v1y * v2z - v2y * v1z);
//what's happening inside Godot's determinant() function
float det2 = v1x * (v2y * v3z - v2z * v3y) - v1y * (v2x * v3z - v2z * v3x) + v1z * (v2x * v3y - v2y * v3x);
printf("%f %f\n", det1, det2);
}
Output (float):
1.946045 1.954590
Same code with double:
1.946244 1.946244
Excessively precise value to compare:
1.946244346015180135873251857346095438627549094323665457295646774582564830780029296875b0
Mathematically determinant should be the same, but looks like it affects float precision loss somehow.
OK, thanks @bruvzg . I didn't notice that their calculation orders are different.
|
gharchive/issue
| 2018-12-18T14:17:12 |
2025-04-01T04:34:23.168936
|
{
"authors": [
"Zylann",
"bruvzg",
"tinyleolin"
],
"repo": "godotengine/godot",
"url": "https://github.com/godotengine/godot/issues/24449",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
}
|
402948401
|
BT_DEBUG is not defined in debug builds
Godot 3.1 ab843b16984d8b62a1100d9b77d8b0dcdc4252bb
I found out that in debug builds of Godot, Bullet may be asserting with btAssert, except it expands to no-op, because BT_DEBUG is not defined.
It seems only iOS builds (so far) keep the asserts defined regardless (due to how Bullet was written rather than the SCons script).
The result is, for example, someone testing on iOS found crashes that may be happening silently in general with heightmap terrains, where parameters are sent as garbage because set_shape_data was not called yet. There was garbage because like every other shape, the heightmap one doesn't initialize any of its members in its constructor...
Is there a reason why BT_DEBUG was not defined?
Temporarily disabled by #25500 as it exposed bugs and we're too close to the 3.1 release to handle them :)
Will be fixed anew after 3.1, when we'll revert #25500.
While working on #34210 I noticed that this requires DEBUG to be enabled not just BT_DEBUG.
|
gharchive/issue
| 2019-01-25T00:00:06 |
2025-04-01T04:34:23.171937
|
{
"authors": [
"Zylann",
"akien-mga",
"madmiraal"
],
"repo": "godotengine/godot",
"url": "https://github.com/godotengine/godot/issues/25301",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
}
|
414562637
|
3D meshes do not render on some Android phones with GLES2
Godot version:
84b3d44 (from #26299) (applications don't even render anything on this phone prior to this PR)
OS/device including version:
Samsung Galaxy S3
GPU ARM Mali-400 MP4
Running LineageOS 14.1 / Android 7.1.2
Issue description:
When running the attached demo using GLES2, I see only text and not the 3D untextured meshes.
Here's how it looks like on desktop or on my higher end phone:
Note: I haven't investigated yet whether the issue is with any 3D mesh, or due to the combination of 3D in 2D via subviewports.
Steps to reproduce:
Download attached demo
Export it on an affected Android device (debug build), run it
Minimal reproduction project:
multitouch_cubes.zip
Note: I haven't investigated yet whether the issue is with any 3D mesh, or due to the combination of 3D in 2D via subviewports.
Confirmed that it seems to be with anything 3D, if I modify the attached project to use CubeScene.tscn as the main scene, I get a black screen on the Samsung Galaxy S3. Even if I remove the WorldEnvironment completely.
Nothing relevant in adb logcat:
02-26 12:32:33.238 21232 21248 I godot : Android modules: Nothing to load, aborting
02-26 12:32:33.243 21232 21249 E : Device driver API match
02-26 12:32:33.243 21232 21249 E : Device driver API version: 29
02-26 12:32:33.243 21232 21249 E : User space API version: 29
02-26 12:32:33.243 21232 21249 E : mali: REVISION=Linux-r3p2-01rel3 BUILD_DATE=Tue Jul 22 19:59:34 KST 2014
02-26 12:32:33.250 21232 21249 I OpenGLRenderer: Initialized EGL, version 1.4
02-26 12:32:33.250 21232 21249 D OpenGLRenderer: Swap behavior 1
02-26 12:32:33.310 21232 21248 W GodotView: creating OpenGL ES 2.0 context :
02-26 12:32:33.373 2362 2396 I ActivityManager: Displayed org.godotengine.interactivecubes/org.godotengine.godot.Godot: +1s394ms
02-26 12:32:33.464 21282 21282 W IInputConnectionWrapper: reportFullscreenMode on inexistent InputConnection
02-26 12:32:33.609 2503 2503 I Choreographer: Skipped 69 frames! The application may be doing too much work on its main thread.
02-26 12:32:33.633 21232 21243 I art : Background sticky concurrent mark sweep GC freed 18869(802KB) AllocSpace objects, 0(0B) LOS objects, 54% free, 3MB/6MB, paused 1.638ms total 265.715ms
02-26 12:32:33.756 21232 21248 I godot : OpenGL ES 2.0 Renderer: Mali-400 MP
02-26 12:32:34.229 21232 21248 I libOpenSLES: Emulating old channel mask behavior (ignoring positional mask 0x3, using default mask 0x3 based on channel count of 2)
02-26 12:32:34.233 2016 1078 W AudioFlinger: createTrack_l(): mismatch between requested flags (00000104) and output flags (00000002)
02-26 12:32:34.240 21232 21248 D AudioTrack: Client defaulted notificationFrames to 1056 for frameCount 6348
02-26 12:32:34.244 2016 2145 D audio_hw_primary: select_output_device: AUDIO_DEVICE_OUT_SPEAKER
02-26 12:32:34.523 2362 2396 I art : Starting a blocking GC Explicit
02-26 12:32:34.709 2362 2396 I art : Explicit concurrent mark sweep GC freed 50119(3MB) AllocSpace objects, 3(60KB) LOS objects, 33% free, 16MB/24MB, paused 2.493ms total 177.462ms
02-26 12:32:34.828 2016 2145 I AudioFlinger: BUFFER TIMEOUT: remove(4096) from active list on thread 0x414031c0
02-26 12:32:34.879 21232 21265 E AudioTrack: did not receive expected priority boost on time
02-26 12:32:34.879 21232 21265 W AudioTrack: releaseBuffer() track 0x42c2e400 disabled due to previous underrun, restarting
Still running CubeScene.tscn directly, I enabled "Debug OpenGL" in the preset, and now the game crashes instantly on start, with this adb logcat output:
02-26 12:36:32.112 2362 2758 I ActivityManager: Start proc 21491:org.godotengine.interactivecubes/u0a104 for activity org.godotengine.interactivecubes/org.godotengine.godot.Godot
02-26 12:36:32.136 21491 21491 I art : Late-enabling -Xcheck:jni
02-26 12:36:32.481 2362 30888 I ActivityManager: Config changes=480 {1.0 238mcc6mnc [en_GB,da_DK,de_DE,fr_FR] ldltr sw360dp w640dp h336dp 320dpi nrml long land finger -keyb/v/h -nav/h s.51}
02-26 12:36:32.588 2362 30888 W ActivityManager: Slow operation: 417ms so far, now at attachApplicationLocked: after mServices.attachApplicationLocked
02-26 12:36:32.588 2362 2396 W art : Long monitor contention with owner Binder:2362_1C (30888) at void com.android.server.am.ActivityManagerService.attachApplication(android.app.IApplicationThread)(ActivityManagerService.java:6882) waiters=0 in void com.android.server.am.ActivityManagerService$LocalService.notifyAppTransitionCancelled() for 406ms
02-26 12:36:32.600 2362 2443 I InputReader: Reconfiguring input devices. changes=0x00000004
02-26 12:36:32.601 2362 2443 I InputReader: Device reconfigured: id=11, name='sec_touchscreen', size 720x1280, orientation 1, mode 1, display id 0
02-26 12:36:32.602 2362 2443 I InputReader: Reconfiguring input devices. changes=0x00000004
02-26 12:36:32.647 21491 21506 D libEGL : loaded /system/lib/egl/libEGL_mali.so
02-26 12:36:32.664 21491 21506 D libEGL : loaded /system/lib/egl/libGLESv1_CM_mali.so
02-26 12:36:32.800 21491 21506 D libEGL : loaded /system/lib/egl/libGLESv2_mali.so
02-26 12:36:32.928 21491 21510 E : Device driver API match
02-26 12:36:32.928 21491 21510 E : Device driver API version: 29
02-26 12:36:32.928 21491 21510 E : User space API version: 29
02-26 12:36:32.928 21491 21510 E : mali: REVISION=Linux-r3p2-01rel3 BUILD_DATE=Tue Jul 22 19:59:34 KST 2014
02-26 12:36:32.954 21491 21510 I OpenGLRenderer: Initialized EGL, version 1.4
02-26 12:36:32.954 21491 21510 D OpenGLRenderer: Swap behavior 1
02-26 12:36:33.006 2362 2536 I WindowManager: setAppOrientation token: Token{73e01e6 ActivityRecord{6e28841 u0 org.godotengine.interactivecubes/org.godotengine.godot.Godot t1513}} requestedOrientation: 0
02-26 12:36:33.009 21491 21509 I godot : Android modules: Nothing to load, aborting
02-26 12:36:33.013 21491 21509 W GodotView: creating OpenGL ES 2.0 context :
02-26 12:36:33.014 21491 21509 E GodotView: After eglCreateContext: EGL error: 0x3004
02-26 12:36:33.017 21491 21509 E AndroidRuntime: FATAL EXCEPTION: GLThread 2615
02-26 12:36:33.017 21491 21509 E AndroidRuntime: Process: org.godotengine.interactivecubes, PID: 21491
02-26 12:36:33.017 21491 21509 E AndroidRuntime: java.lang.RuntimeException: createContext failed: EGL_SUCCESS
02-26 12:36:33.017 21491 21509 E AndroidRuntime: at android.opengl.GLSurfaceView$EglHelper.throwEglException(GLSurfaceView.java:1211)
02-26 12:36:33.017 21491 21509 E AndroidRuntime: at android.opengl.GLSurfaceView$EglHelper.throwEglException(GLSurfaceView.java:1202)
02-26 12:36:33.017 21491 21509 E AndroidRuntime: at android.opengl.GLSurfaceView$EglHelper.start(GLSurfaceView.java:1052)
02-26 12:36:33.017 21491 21509 E AndroidRuntime: at android.opengl.GLSurfaceView$GLThread.guardedRun(GLSurfaceView.java:1412)
02-26 12:36:33.017 21491 21509 E AndroidRuntime: at android.opengl.GLSurfaceView$GLThread.run(GLSurfaceView.java:1259)
02-26 12:36:33.029 2362 32284 W ActivityManager: Force finishing activity org.godotengine.interactivecubes/org.godotengine.godot.Godot
02-26 12:36:33.129 2362 2396 I WindowManager: Screen frozen for +952ms due to Window{41efe79 u0 Starting org.godotengine.interactivecubes}
02-26 12:36:33.197 2362 11564 W InputMethodManagerService: Focus gain on non-focused client com.android.internal.view.IInputMethodClient$Stub$Proxy@21a74da (uid=10104 pid=21491)
02-26 12:36:33.200 2362 20095 I OpenGLRenderer: Initialized EGL, version 1.4
02-26 12:36:33.200 2362 20095 D OpenGLRenderer: Swap behavior 1
That is EGL_BAD_ATTRIBUTE 0x3004
Thanks!
For the reference, it's still valid in 8889089c (to rule out a potential issue in my own unmerged patch 84b3d44).
does it still happen?
Tested on 41708eb8 and it's worse now, I get a black screen again.
057d3e87 regressed and #24952 is broken again.
I tested with 41708eb + a local revert of 057d3e8 that broke rendering, and I can confirm that this issue is still valid (text and clear color work, 3D meshes don't).
Still reproducible in 834a984b with the project in OP.
Did some more testing by focusing on the CubeScene.tscn. In the OP's MRP, it doesn't render at all on my Samsung S3.
I tried to modify it a bit, changing/removing the WorldEnvironment, but that didn't fix it.
Then I added a Sprite3D to the scene and tadaa... both the Sprite3D and the MeshInstance become visible. So it sounds like Sprite3D triggers a code branch that makes the MeshInstance visible.
Updated MRP: multitouch_cubes.zip
Exported as is, the main scene renders fine. Delete the Sprite3D, reexport and then the cube doesn't render anymore.
Can you still reproduce this given all the changes to the GLES2 backend since this was made?
Mixed results... by default I get a black screen, but toying a bit with the WorldEnvironment I can now get the cubes to display without having to add a Sprite3D.
It seems like Background Mode "Clear Color" and "Custom Color" trigger a black screen (even with non-black custom color) which hampers testing this issue, but if I use a ProceduralSky or Canvas mode it seems to work fine.
I'll do more tests and open a new issue for the Bg Mode bugs.
|
gharchive/issue
| 2019-02-26T11:25:09 |
2025-04-01T04:34:23.183213
|
{
"authors": [
"akien-mga",
"clayjohn",
"hpvb",
"reduz"
],
"repo": "godotengine/godot",
"url": "https://github.com/godotengine/godot/issues/26303",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
}
|
421841425
|
Classes recursive references lead to "script could not be loaded"
Godot version: 3.1
OS/device including version: Ubuntu 18.04
Issue description:
Say, you have the following:
# one.gd
extends Node2D
class_name One
func use_two(var two: Two):
pass
# two.gd
extends Node
class_name Two
func use_one(var one: One):
pass
The project will fail on startup with error "Class 'One' was found in global scope but its script could not be loaded"
If we remove one of type annotations : One or : Two from functions above, it will launch without any complaint
Note: same behavior was discovered for case when we replace class_name-s with preload logic, but the message is slightly different (something about recursive references)
Minimal reproduction project:
repro_recursive.tar.gz
Partial duplicate of #21461, but that issue is long and confusing, so it's probably best to keep this new one to focus on the class_name cross-referencing issue.
Note: same behavior was discovered for case when we replace class_name-s with preload logic, but the message is slightly different (something about recursive references)
That one is expected, preload involves loading the whole script, and thus resolving all its dependencies, leading to recursive references. But for class_name, it should be solvable somehow with a system akin to forward declares.
Pretty much the same as #21461. Both will be solved together.
This bug is also present when class references one and the other preloads the other.
So
`
one.gd
extends Node
class_name One
var two = preload("res://two.gd")
two.gd
extends Node
func use_one(var one: One):
pass
`
gives the same error when the program is executed.
What's the status on this bug?
@GaidamakUA @vnen has started working on a GDScript refactor which will eliminate this issue. It should be finished by the time Godot 4.0 is released.
Duplicate of https://github.com/godotengine/godot/issues/21461 (same cause).
|
gharchive/issue
| 2019-03-16T19:02:10 |
2025-04-01T04:34:23.191877
|
{
"authors": [
"Calinou",
"GaidamakUA",
"Mastermori",
"akien-mga",
"smedelyan",
"vnen"
],
"repo": "godotengine/godot",
"url": "https://github.com/godotengine/godot/issues/27136",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
}
|
427457099
|
Multimesh Bug in GLES3
Godot version:
Godot 3.1.0-Stable
OS/device including version:
Windows 10
Radeon RX 570 (19.3.2)
Issue description:
I found this problem while programming on my game. I wanted to use multiple Multimesh nodes and shortly after that I wondered why it wasn't properly working (only one of the Multimesh-Nodes was visible and sometimes it completely bugged out). After I've reduced the Multimesh nodes to one in the entire project it suddenly worked, without touching any GDScript-Code. I only removed the other Multimesh nodes which shared the same script. So after some investigation I found out that this problem doesn't occur in the GLES2 renderer. So my assumption is that in GLES3 the code (or shader?) interferes with each other which can causes wrong rendering or other glitches. Anyway I provided a simple sample project which hopefully makes it more clear what I mean. In the sample project I just randomly show multiple instances of the assigned shapes and sometimes by doing so it glitches out by showing the shapes at the wrong position or other weird stuff.
Here you can see that the Prism overlaps with the Cubes which shouldn't be the case (I only change the Y-Axis in the generation so this shouldn't happen)
Sometimes it also shows black lines instead of anything(Maybe the scaling of the Matrix bugs out in the Shader of the Multimesh?)
Steps to reproduce:
I think the easiest way is to just open the sample project I've provided
Minimal reproduction project:
MultiMeshBug.zip
I am not able to reproduce the described issue. Is there anything you do in addition to running the sample project to spot the issue? I ran it a bunch of time, both in GLES2 and GLES3 and I couldn't reproduce the bug.
Linux 16.04 intel integrated graphics v3.1
@clayjohn, nothing special just running the project. Maybe this is driver related and not Godot related? I will boot up a Linux tomorrow and see if it happens there aswell
RX 470, Win 7 64 bit.
I see it cycle through multiple different MultiMesh generations One of which looks like your first image, did not see anything like your second image though.
Alright I gave it a go on my installed Linux as well (Fedora) and it didn't bugged out there. So I guess this is only Windows related?
@Megalomaniak
To be fair, the black thing on the second image is pretty hard to spot since it only appears for a few ms on my machine. I actually had to record the game window and stop it in the right position to make a screenshot of it ^^
The project gave a few random layouts of items for a few seconds, and then the game crashed with no info but a typical ERROR: _get_socket_error: Socket error: 10054.
Windows 10 64 bits, GeForce GTX 750 Ti/PCIe/SSE2
@Zylann And with GLES 2 everything is fine for you too?
@Nophlock Right, I only ran it for a short period too.
@Megalomaniak GLES2 behaves the same but doesn't crashes.
@Zylann So visually you see the following with GLES2 as well as 3 using the GeForce GTX 750Ti:
?
@Megalomaniak I see the same shapes yes.
OK, so there is a difference between nVidia and AMD GPUs then.
@Megalomaniak it seems that windows vs linux is also an important factor.
Same behavior on another Windows computer on GLES3 with an nVidia GeForce GTX 1060 6GB/PCIe/SSE2: random shapes, some unexpected, then crash.
The example project also crashes on my Windows machine. It's a driver crash when calling glBindBuffer(GL_ARRAY_BUFFER, <invalid buffer id>).
I went looking around and found this:
https://github.com/godotengine/godot/blob/master/drivers/gles3/rasterizer_storage_gles3.cpp#L4454-L4457
Adding multimesh->buffer = 0; to reset the buffer id when deleting the buffer fixes the crash. Even though I didn't dig deep enough to find out exactly why, it makes sense and I'll submit a PR for others to test.
|
gharchive/issue
| 2019-03-31T22:49:16 |
2025-04-01T04:34:23.202556
|
{
"authors": [
"MadEqua",
"Megalomaniak",
"Nophlock",
"Zylann",
"clayjohn"
],
"repo": "godotengine/godot",
"url": "https://github.com/godotengine/godot/issues/27568",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
}
|
141381795
|
Better Collada Exporter - export only deform bones
Hello,
Testing the Better Collada Exporter for Blender, I notice all of the bones in a rig get exported to Godot. A desired way would be to only export those bones that directly influence the mesh (i.e. deform bones). In Blender, bones can be flagged as being 'deform' bones and e.g. the FBX exporter has an option to only export those. It would be nice and very much needed for Better Collada to support this.
A rig is usually made from many bones. There are deform bones which directly deform the mesh, then there are control bones that are used to animate, and also helper bones. Together they make a rig comfortable to use and animate. However, for games, you only need the deform bones, the rest can and should be discarded.
Moved to https://github.com/godotengine/collada-exporter/issues/14.
|
gharchive/issue
| 2016-03-16T19:46:32 |
2025-04-01T04:34:23.205270
|
{
"authors": [
"Lamoot",
"akien-mga"
],
"repo": "godotengine/godot",
"url": "https://github.com/godotengine/godot/issues/4074",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
}
|
815455511
|
HTML5 export, not all objects exported
Godot version: 3.2.3
OS/device including version: running godot on Linux (Lubuntu/Ubuntu)
Issue description: When export to HTML5 and run in (firefox) browser, the game kind of runs but no red ball is shown. When I run the game in Godot or when I generate executable for Linux, it works correctly, the red ball is shown:
Steps to reproduce: clone the https://gitlab.com/isokissa/games/zivac, export game to HTML5, start the local server and goto localhost:, or go to https://isokissa.gitlab.io/games/zivac/ to see the game which does not show the red ball. When you run within Godot, or exported to Linux executable, there is a red ball
Minimal reproduction project: https://gitlab.com/isokissa/games/zivac (it is just a beginning of the project: look at the arena/arena scene, where Ball scene is instantiated. That Ball scene from Arena is not shown when run in HTML5 export)
I can confirm this on Chromium 87 and Firefox 86 (Linux).
@Calinou using sprite did not help. It is again visible when running as application, but not in browser.
@Calinou apparently it depends on the shape of the browser screen. If it's wide so that black areas are shown on the side, then the ball does not show. If the windows is narrow, so that black areas are on top and bottom, then the ball is shown, but at wrong place (I guess also in wide case it is in wrong place, but out of screen, so that's the reason why it is not shown)
This means, the problem is that HTML5 export, this code
func ball_spawn():
var middle: Vector2 = OS.window_size / 2
get_node("Ball").position = middle
does not put Ball in the middle of the window as when ran standalone (Ball's position (0,0) is in the origin)
This is not a bug, you're using the wrong variable, OS.window_size gives you the size of the container the game exists within not the size of the game's viewport or in other words the size including the black bars. the correct variable you wanted would be get_tree().get_root().get_size(), this give you want the game's main viewport size after stretch mode has been applied / after the black bars have been applied.
|
gharchive/issue
| 2021-02-24T12:59:51 |
2025-04-01T04:34:23.211595
|
{
"authors": [
"Calinou",
"Kitsee",
"isokissa"
],
"repo": "godotengine/godot",
"url": "https://github.com/godotengine/godot/issues/46372",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
}
|
963351811
|
XMLParser handles white space outside tags incorrectly/inconsistently.
Godot version
v3.3.stable.official
System information
Linux
Issue description
XMLParser treats "trailing" whitespace differently depending on whether it's two or three+ characters long.
This occurs here: https://github.com/godotengine/godot/blob/49309ac5fb2d8b33cf80c1e4b6eb0d0130f7a24a/core/io/xml_parser.cpp#L45-L46
Origin
This behaviour appears to originate in the original irrXML implementation. (See, e.g. https://sourceforge.net/p/g3d/code/HEAD/tree/G3D10/external/assimp.lib/contrib/irrXML/CXMLReaderImpl.h#l257.)
The why doesn't appear to be documented but I'm guessing it was maybe intended for CR+LF end-of-line handling.
Impact
However, as a result, it handles single-character EOL situation incorrectly--or at least differently depending on e.g. whether subsequent line indent is one tab character or multiple space characters.
Based on this comment https://github.com/godotengine/godot/issues/23895#issuecomment-442432191 this behaviour (as well as being confusing when first encountered) is probably not compliant with the specification.
Background
I encountered this issue when my code worked with one file but didn't work with another and it wasn't obvious what the difference was between the two. Initially I suspected that it may be line-ending related but that wasn't the case so I was stumped until I looked at the parsing implementation in Godot and discovered the "special case" handling.
Potential fix
While it would break compatibility, going forward it would seem best if this special case handling was removed or at least changed to only affect end-of-line related characters. (Depending on what the XML specification says.)
Other impact
It seems this special case handling also means it's impossible to "round trip" certain XML files by reading the file in and then immediately writing it out again because the initial reading is "lossy" in an undetectable way.
(Semi-) Related issues
https://github.com/godotengine/godot/issues/23895
https://github.com/godotengine/godot/issues/31682
https://github.com/godotengine/godot/issues/48779 & https://github.com/godotengine/godot/pull/48806
https://github.com/godotengine/godot/issues/28768
Steps to reproduce
Run reproduction script.
Observe output:
----------------------------------------
buffer content: "<xml><element /></xml>"
node type: 1
XMLParser.NODE_ELEMENT
'xml'
node type: 1
XMLParser.NODE_ELEMENT
'element'
node type: 2
XMLParser.NODE_ELEMENT_END
'xml'
----------------------------------------
buffer content: "<xml> <element /></xml>"
node type: 1
XMLParser.NODE_ELEMENT
'xml'
node type: 1
XMLParser.NODE_ELEMENT
'element'
node type: 2
XMLParser.NODE_ELEMENT_END
'xml'
----------------------------------------
buffer content: "<xml> <element /></xml>"
node type: 1
XMLParser.NODE_ELEMENT
'xml'
node type: 1
XMLParser.NODE_ELEMENT
'element'
node type: 2
XMLParser.NODE_ELEMENT_END
'xml'
----------------------------------------
buffer content: "<xml> <element /></xml>"
node type: 1
XMLParser.NODE_ELEMENT
'xml'
node type: 3
XMLParser.NODE_TEXT
0x20
0x20
0x20
node type: 1
XMLParser.NODE_ELEMENT
'element'
node type: 2
XMLParser.NODE_ELEMENT_END
'xml'
Note that the last example suddenly shows an additional NODE_TEXT node which was missing from the previous two examples which also included whitespace characters (but fewer than 3).
Minimal reproduction project
extends Control
# Note: Oddly, Godot treats "trailing" whitespace differently
# depending on whether it's two or three+ characters long.
#
# See: <https://github.com/godotengine/godot/blame/49309ac5fb2d8b33cf80c1e4b6eb0d0130f7a24a/core/io/xml_parser.cpp#L45-L46>
#
# This behaviour appears to originate in the original
# `irrXML` implementation.
#
# See, e.g.: <https://sourceforge.net/p/g3d/code/HEAD/tree/G3D10/external/assimp.lib/contrib/irrXML/CXMLReaderImpl.h#l257>
#
# The *why* doesn't appear to be documented but I'm
# guessing it was maybe intended for CR+LF end-of-line
# handling.
#
# However, as a result, it handles single-character EOL
# situation incorrectly--or at least differently depending
# on e.g. whether subsequent line indent is one tab character
# or multiple space characters.
var test_xml_buffer_1 = "<xml>\r<element /></xml>"
var test_xml_buffer_2 = "<xml>\r\t<element /></xml>"
var test_xml_buffer_3 = "<xml>\r <element /></xml>"
var test_xml_buffer_4 = "<xml>\r <element /></xml>"
# While I encountered this issue with content that included non-printing end-of-line
# characters (shown above) this set of test data uses only spaces because that
# makes it easier to see the differences between each example:
var test_xml_buffer_5 = "<xml><element /></xml>"
var test_xml_buffer_6 = "<xml> <element /></xml>"
var test_xml_buffer_7 = "<xml> <element /></xml>"
var test_xml_buffer_8 = "<xml> <element /></xml>"
func parse_xml(test_xml_buffer) -> void:
print("-".repeat(40))
print("buffer content: \"%s\"" % test_xml_buffer)
var xml_parser: XMLParser = XMLParser.new()
var err = xml_parser.open_buffer(test_xml_buffer.to_utf8())
if err != OK:
push_error("error: %s" % err)
return
while true:
err = xml_parser.read()
if err != OK:
push_error("error: %s" % err)
return
var node_type = xml_parser.get_node_type()
var node_type_str = ""
var node_name = ""
var node_data = ""
prints("node type:", node_type)
node_type_str = [
"XMLParser.NODE_NONE",
"XMLParser.NODE_ELEMENT",
"XMLParser.NODE_ELEMENT_END",
"XMLParser.NODE_TEXT",
# ...
][node_type]
prints(" ", node_type_str)
match node_type:
XMLParser.NODE_ELEMENT, XMLParser.NODE_ELEMENT_END:
node_name = xml_parser.get_node_name()
prints(" ", "'%s'" % node_name)
XMLParser.NODE_TEXT:
node_data = xml_parser.get_node_data()
for i in range(node_data.length()):
prints(" ", "0x%02x" % ord(node_data.substr(i,1)))
_:
push_error("Unhandled node type.")
print()
func _ready() -> void:
parse_xml(test_xml_buffer_5)
parse_xml(test_xml_buffer_6)
parse_xml(test_xml_buffer_7)
parse_xml(test_xml_buffer_8)
FWIW the XML specification addresses the handling of end-of-line characters here: https://www.w3.org/TR/REC-xml/#sec-line-ends
To summarize, there are two issues:
If the text node contains only one or two whitespace characters, the text node is ignored.
The end-of-line handling required by the specification is not implemented.
Input
Expected
Actual
<xml><element /></xml>
no text
no text
<xml>\t<element /></xml>
text(0x09)
no text❌
<xml>\t\t<element /></xml>
text(0x09 0x09)
no text❌
<xml>\t\t\t<element /></xml>
text(0x09 0x09 0x09)
text(0x09 0x09 0x09)
<xml>\t\t\r<element /></xml>
text(0x09 0x09 0x0A)
text(0x09 0x09 0x0D)❌
<xml>\t\t\r\n<element /></xml>
text(0x09 0x09 0x0A)
text(0x09 0x09 0x0D 0x0A)❌
Just noticed the latest version of the parsing code in irrXML/Irrlicht Engine has been modified to be (in part):
//[...]
// By default xml preserves all whitespace. But Irrlicht dropped some whitespace by default
// in the past which did lead to OS dependent behavior. We just ignore all whitespace for now
// as it's the closest to fixing behavior without breaking downward compatibility too much.
if ( IgnoreWhitespaceText )
{
char_type* p = start;
for(; p != end; ++p)
if (!isWhiteSpace(*p))
break;
if (p == end)
return false;
}
//[...]
May be related #81896
|
gharchive/issue
| 2021-08-08T04:29:30 |
2025-04-01T04:34:23.227445
|
{
"authors": [
"elenakrittik",
"follower",
"timothyqiu"
],
"repo": "godotengine/godot",
"url": "https://github.com/godotengine/godot/issues/51380",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
}
|
159786622
|
Support shadows with DynamicFonts
Currently, DynamicFonts can't cast any shadows, not even sharp ones. It would be nice if they could be set to cast a sharp shadow at least, with customizable X/Y offset and color.
Something like this (from another engine):
I think Label has the option for that
On Sat, Jun 11, 2016 at 3:25 PM, Hugo Locurcio notifications@github.com
wrote:
Currently, DynamicFonts can't cast any shadows, not even sharp ones. It
would be nice if they could be set to cast a sharp shadow at least, with
customizable X/Y offset and color.
Something like this (from another engine): [image: shot0001]
https://cloud.githubusercontent.com/assets/180032/15986801/83915e9e-3012-11e6-9e42-50f22093db80.jpg
—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub
https://github.com/godotengine/godot/issues/5159, or mute the thread
https://github.com/notifications/unsubscribe/AF-Z29H1GeQDRAqBYYySU4R5GOnQyWBMks5qKv1_gaJpZM4IznDv
.
I think Label has the option for that
I can confirm, it also has outline options. Closing.
|
gharchive/issue
| 2016-06-11T18:24:47 |
2025-04-01T04:34:23.232828
|
{
"authors": [
"Calinou",
"reduz",
"volzhs"
],
"repo": "godotengine/godot",
"url": "https://github.com/godotengine/godot/issues/5159",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
}
|
1554984556
|
Engine silently includes export_presets.cfg into project binary if *.cfg is in the non-resource export preset
Godot version
3.x
System information
Ubuntu 22.10 LTS
Issue description
Engine silently includes export_presets.cfg if you mark to include all *.cfg files. If there is no error on engine side, it would be worth to show warning.
export_presets.cfg holds important data such as:
username
256-bit AES key (if encryption enabled)
https://github.com/godotengine/godot-proposals/issues/6163
Steps to reproduce
example for *.apk and *.ipa
include *.cfg files in export window
export
open binary with favorite archive manager
Minimal reproduction project
n/a
Just silently ignore export_presets.cfg would be my solution ;)
I think there should be a warning in the console when exporting a project ("export_presets.cfg was skipped for security reasons. Add it to exclude filters to suppress this warning.").
Adding a file to the exclude filter to include it in the exported PCK feels strange. I personally can't see why you'd ever want export_presets.cfg to be included in the exported PCK.
I don't suggest including the export_presets.cfg to .pck/.exe/.... I just suggest explicitly informing the user when and why we do not follow the specified export rules in some exceptional situations. The user can correct the wrong config, and then there will be no warning.
This may sound strange, but I think quiet implicit behavior is bad.
Other possible solutions:
Add export_presets.cfg to the default exclude filters (when creating a new preset).
Add this example in the hint below the field (instead of duplicate from include filters field).
Display a warning that export_presets.cfg will be added to the export (we can expand the list in the future). But the warning should be displayed in the export window, not in the console.
(Hard) Scan all export files for secrets (passwords, private keys).
|
gharchive/issue
| 2023-01-24T13:26:22 |
2025-04-01T04:34:23.240562
|
{
"authors": [
"Calinou",
"Zireael07",
"boruok",
"dalexeev"
],
"repo": "godotengine/godot",
"url": "https://github.com/godotengine/godot/issues/71970",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
}
|
1583250066
|
Editor won't reload scripts on external change
Godot version
v4.0.rc1.official [8843d9ad3]
System information
Windows 10
Issue description
I have the following option selected: text_editor/behavior/files/auto_reload_scripts_on_external_change, but when editing and saving the script file in VSCode, the internal Editor doesn't seem to pick up the change.
Steps to reproduce
Create a script, open it in internal editor, then modify it in external editor (I only tested VSCode), and the internal editor won't reload the new script.
Minimal reproduction project
N/A
Duplicate of #72825
|
gharchive/issue
| 2023-02-13T23:55:39 |
2025-04-01T04:34:23.242922
|
{
"authors": [
"KoBeWi",
"fisheye36"
],
"repo": "godotengine/godot",
"url": "https://github.com/godotengine/godot/issues/73260",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
}
|
1654534801
|
Dereferencing a funcref from the function it calls leads to a heap-use-after-free
Godot version
3.5.1.stable
System information
Arch Linux (rolling release)
Issue description
Dereferencing a funcref from the function it calls can cause an error in AddressSanitizer in debug/release-debug builds. It might be possible for this to cause more subtle heap corruption issues in release builds--I suspect it is behind a rare/unreproducible crash in my project.
I think what happens in the below example is that when foobar dereferences the funcref, the funcref is freed. Then when control flow returns from foobar to FuncRef.call_func the this pointer is invalid. It seems weird that function calls do not implicitly hold a reference to the object they're on.
Note that this isn't unique to FuncRefs - you can implement a FuncRef in pure GDScript and it will have the same bug.
==16930==ERROR: AddressSanitizer: heap-use-after-free on address 0x612000254d70 at pc 0x0000066b3b05 bp 0x7fff849419c0 sp 0x7fff849419b0
WRITE of size 4 at 0x612000254d70 thread T0
#0 0x66b3b04 in std::__atomic_base<unsigned int>::fetch_sub(unsigned int, std::memory_order) /usr/include/c++/12.2.1/bits/atomic_base.h:628
#1 0x66b3b04 in SafeNumeric<unsigned int>::decrement() core/safe_refcount.h:84
#2 0x66b3b04 in SafeRefCount::unref() core/safe_refcount.h:177
#3 0x66b3b04 in _ObjectDebugLock::~_ObjectDebugLock() core/object.cpp:53
#4 0x66b3b04 in Object::call(StringName const&, Variant const**, int, Variant::CallError&) core/object.cpp:924
#5 0x68b8c7a in Variant::call_ptr(StringName const&, Variant const**, int, Variant*, Variant::CallError&) core/variant_call.cpp:1229
#6 0x12f7960 in GDScriptFunction::call(GDScriptInstance*, Variant const**, int, Variant::CallError&, GDScriptFunction::CallState*) modules/gdscript/gdscript_function.cpp:1048
#7 0x122592e in GDScriptInstance::call_multilevel(StringName const&, Variant const**, int) modules/gdscript/gdscript.cpp:1214
#8 0x29b9039 in Node::_notification(int) scene/main/node.cpp:58
#9 0x2cb073f in Node::_notificationv(int, bool) scene/main/node.h:47
#10 0x2cb073f in CanvasItem::_notificationv(int, bool) scene/2d/canvas_item.h:163
#11 0x2cb073f in Control::_notificationv(int, bool) scene/gui/control.h:47
#12 0x66ab532 in Object::notification(int, bool) core/object.cpp:927
#13 0x2a4c801 in SceneTree::_notify_group_pause(StringName const&, int) scene/main/scene_tree.cpp:1105
#14 0x2a733d1 in SceneTree::idle(float) scene/main/scene_tree.cpp:608
#15 0xb08189 in Main::iteration() main/main.cpp:2298
#16 0xaaf46f in OS_X11::run() platform/x11/os_x11.cpp:3987
#17 0xa4e5dc in main platform/x11/godot_x11.cpp:59
#18 0x7f1c6d03c78f (/usr/lib/libc.so.6+0x2378f)
#19 0x7f1c6d03c849 in __libc_start_main (/usr/lib/libc.so.6+0x23849)
#20 0xa64634 in _start (/media/scratch/coding/godot/godot-builder/godot/bin/godot.x11.opt.debug.64s+0xa64634)
0x612000254d70 is located 48 bytes inside of 272-byte region [0x612000254d40,0x612000254e50)
freed by thread T0 here:
#0 0x7f1c6dcbe672 in __interceptor_free /usr/src/debug/gcc/gcc/libsanitizer/asan/asan_malloc_linux.cpp:52
#1 0x6753795 in void memdelete<Reference>(Reference*) core/os/memory.h:118
#2 0x6753795 in void memdelete<Reference>(Reference*) core/os/memory.h:110
#3 0x6753795 in Ref<Reference>::unref() core/reference.h:258
#4 0x6753795 in RefPtr::unref() core/ref_ptr.cpp:85
previously allocated by thread T0 here:
#0 0x7f1c6dcbfa89 in __interceptor_malloc /usr/src/debug/gcc/gcc/libsanitizer/asan/asan_malloc_linux.cpp:69
#1 0x6b16b2f in Memory::alloc_static(unsigned long, bool) core/os/memory.cpp:75
#2 0x12f66ed in GDScriptFunction::call(GDScriptInstance*, Variant const**, int, Variant::CallError&, GDScriptFunction::CallState*) modules/gdscript/gdscript_function.cpp:1116
#3 0x123621b in GDScriptInstance::_ml_call_reversed(GDScript*, StringName const&, Variant const**, int) modules/gdscript/gdscript.cpp:1229
Steps to reproduce
Put this script on a node and then play the scene, using a release_debug build with asan on:
var fr
func _ready():
fr = funcref(self, "foobar")
func foobar():
fr = null
func _process(_delta):
if fr:
fr.call_func()
Minimal reproduction project
N/A
I've run into this issue at least a dozen separate times now in unrelated parts of my codebase, in multiple projects, in Godot 3 and in Godot 4 as well. I seriously can't be the only one encountering this :laughing:
Updated example that works with Godot 4.3:
godot-callable-crash.zip
Build Godot with asan, run the callable_example scene and observe the asan error:
==90026==ERROR: AddressSanitizer: heap-use-after-free on address 0x5140001a8ca0 at pc 0x636b24b87682 bp 0x7ffd10c69a30 sp 0x7ffd10c69a28
WRITE of size 4 at 0x5140001a8ca0 thread T0
#0 0x636b24b87681 in std::__atomic_base<unsigned int>::fetch_sub(unsigned int, std::memory_order) /usr/bin/../lib64/gcc/x86_64-pc-linux-gnu/14.1.1/../../../../include/c++/14.1.1/bits/atomic_base.h:641:16
#1 0x636b24b87681 in SafeNumeric<unsigned int>::decrement() /media/scratch/coding/godot/godot-builder/godot/./core/templates/safe_refcount.h:87:16
#2 0x636b24b87681 in SafeRefCount::unref() /media/scratch/coding/godot/godot-builder/godot/./core/templates/safe_refcount.h:207:16
#3 0x636b24b87681 in _ObjectDebugLock::~_ObjectDebugLock() /media/scratch/coding/godot/godot-builder/godot/core/object/object.cpp:55:20
#4 0x636b24b87681 in Object::callp(StringName const&, Variant const**, int, Callable::CallError&) /media/scratch/coding/godot/godot-builder/godot/core/object/object.cpp:814:1
#5 0x636b24383b0a in Variant::callp(StringName const&, Variant const**, int, Variant&, Callable::CallError&) /media/scratch/coding/godot/godot-builder/godot/core/variant/variant_call.cpp:1211:27
#6 0x636b18891d65 in GDScriptFunction::call(GDScriptInstance*, Variant const**, int, Callable::CallError&, GDScriptFunction::CallState*) /media/scratch/coding/godot/godot-builder/godot/modules/gdscript/gdscript_vm.cpp:1780:12
#7 0x636b184598e9 in GDScriptInstance::callp(StringName const&, Variant const**, int, Callable::CallError&) /media/scratch/coding/godot/godot-builder/godot/modules/gdscript/gdscript.cpp:2032:22
#8 0x636b1df52251 in bool Node::_gdvirtual__ready_call<false>() /media/scratch/coding/godot/godot-builder/godot/scene/main/node.h:355:2
#9 0x636b1df52251 in Node::_notification(int) /media/scratch/coding/godot/godot-builder/godot/scene/main/node.cpp:217:4
#10 0x636b1f798bd1 in Node::_notificationv(int, bool) /media/scratch/coding/godot/godot-builder/godot/./scene/main/node.h:50:2
#11 0x636b1f798bd1 in CanvasItem::_notificationv(int, bool) /media/scratch/coding/godot/godot-builder/godot/./scene/main/canvas_item.h:45:2
#12 0x636b1f798bd1 in Node2D::_notificationv(int, bool) /media/scratch/coding/godot/godot-builder/godot/scene/2d/node_2d.h:37:2
#13 0x636b24b7edc1 in Object::notification(int, bool) /media/scratch/coding/godot/godot-builder/godot/core/object/object.cpp
#14 0x636b1df5ad36 in Node::_propagate_ready() /media/scratch/coding/godot/godot-builder/godot/scene/main/node.cpp:272:3
#15 0x636b1df5aca2 in Node::_propagate_ready() /media/scratch/coding/godot/godot-builder/godot/scene/main/node.cpp:263:12
#16 0x636b1df6c109 in Node::_set_tree(SceneTree*) /media/scratch/coding/godot/godot-builder/godot/scene/main/node.cpp:3171:4
#17 0x636b17b3417b in OS_LinuxBSD::run() /media/scratch/coding/godot/godot-builder/godot/platform/linuxbsd/os_linuxbsd.cpp:950:13
#18 0x636b17b17295 in main /media/scratch/coding/godot/godot-builder/godot/platform/linuxbsd/godot_linuxbsd.cpp:85:6
#19 0x7b167e034e07 (/usr/lib/libc.so.6+0x25e07) (BuildId: 476816ac63052a8d67cb2a7e2b93ee9bd72e417b)
#20 0x7b167e034ecb in __libc_start_main (/usr/lib/libc.so.6+0x25ecb) (BuildId: 476816ac63052a8d67cb2a7e2b93ee9bd72e417b)
#21 0x636b179dd2e4 in _start (/media/scratch/coding/godot/godot-builder/godot/bin/godot.linuxbsd.editor.x86_64.llvm.san+0x4dc62e4) (BuildId: 94136b4b522f8ae4)
0x5140001a8ca0 is located 96 bytes inside of 392-byte region [0x5140001a8c40,0x5140001a8dc8)
freed by thread T0 here:
#0 0x636b17acb492 in free.part.0 (/media/scratch/coding/godot/godot-builder/godot/bin/godot.linuxbsd.editor.x86_64.llvm.san+0x4eb4492) (BuildId: 94136b4b522f8ae4)
#1 0x636b2432d184 in Variant::_clear_internal() /media/scratch/coding/godot/godot-builder/godot/core/variant/variant.cpp:1389:6
#2 0x636b184598e9 in GDScriptInstance::callp(StringName const&, Variant const**, int, Callable::CallError&) /media/scratch/coding/godot/godot-builder/godot/modules/gdscript/gdscript.cpp:2032:22
previously allocated by thread T0 here:
#0 0x636b17acc4c9 in malloc (/media/scratch/coding/godot/godot-builder/godot/bin/godot.linuxbsd.editor.x86_64.llvm.san+0x4eb54c9) (BuildId: 94136b4b522f8ae4)
#1 0x636b23b772df in Memory::alloc_static(unsigned long, bool) /media/scratch/coding/godot/godot-builder/godot/core/os/memory.cpp:75:14
#2 0x636b23b772df in operator new(unsigned long, char const*) /media/scratch/coding/godot/godot-builder/godot/core/os/memory.cpp:40:9
#3 0x636b1842d697 in GDScriptNativeClass::instantiate() /media/scratch/coding/godot/godot-builder/godot/modules/gdscript/gdscript.cpp:99:9
#4 0x636b1842d697 in GDScript::_new(Variant const**, int, Callable::CallError&) /media/scratch/coding/godot/godot-builder/godot/modules/gdscript/gdscript.cpp:222:29
This sometimes leads to a crash in non-asan builds (in editor, debug, and in release template builds), but it's very difficult to reproduce that way.
It happens because a RefCounted object's refcount can go to zero, causing the object to be freed, while one of its methods is still somewhere on the stack. When the stack unwinds to that method, there's a heap-use-after-free.
The pattern of use that tends to hit this bug in real code is one where there's an object that emits a signal when it's finished - and where the consumer of that signal clears the sole reference to that object (whether intentionally, incidentally, or indirectly). The other example in that project is more along these lines, and demonstrates just how insidious this bug can be:
func _ready() -> void:
var promise = $JobQueue.post(some_slow_operation)
$TextureRect.texture = await promise.fulfilled
When I look at this, I don't tend to think that when _ready returns it's going to (1) free the promise, and then (2) jump back into some method on the promise and crash, lol. That's some spooky action at a distance.
CC @godotengine/gdscript @godotengine/core
This bug seems caused by OBJ_DEBUG_LOCK in Object::call().
Likely this is holding onto a reference to the object and then trying to change a ref count after deletion.
Removing the line OBJ_DEBUG_LOCK seems to remove the bug.
Still investigating.
struct _ObjectDebugLock {
Object *obj; // I AM THE PROBLEM HERE, RAW POINTER TO OBJECT
ObjectID id;
_ObjectDebugLock(Object *p_obj) {
obj = p_obj;
obj->_lock_index.ref();
if (obj)
{
id = obj->get_instance_id();
}
}
~_ObjectDebugLock() {
if (id)
{
if (ObjectDB::get_instance(id))
{
obj->_lock_index.unref();
}
}
}
};
Actually the bodge fix might be not so bad after all.
The FuncRef is a reference, so has a ref count, but the problem is that ObjectDebugLock works with all Objects (which don't seem to have a refcount on cursory inspection). So there's no easy way to defer the deletion.
We could test whether the object is a Reference, and if so, increase the ref count around the call. But although that fixes this case, it doesn't fix the general case for non-reference objects being deleted. Maybe this can't happen but it would be nice to close that too.
Looking up instance ids is probably super slow but it seems debug only.
Maybe there's a more elegant way though.
It seems we crossed our ways. I've made a fix based on Variant as the responsible for validity check.
It seems we crossed our ways. I've made a fix based on Variant as the responsible for validity check.
That looks good I'll check in a second. :+1:
You are more familiar with Object etc than me.
Akien poked me this morning to have a look at these bugs.
There's also the problem that if we did increase the refcount to prevent this (debug) bug, we would probably need to do it in release too because the lifetime would change subtly, and could lead to further hard to debug bugs in the future. So maybe the bodge is better.
Another issue would be the predelete handler. When the object's refcount reaches 0, it frees the object, triggering the predelete_handler and NOTIFICATION_PREDELETE. That could then do a method call and re-increment the refcount. I think this would cause an infinite loop unless how the predelete handler works was also changed somehow?
This bug seems caused by OBJ_DEBUG_LOCK in Object::call().
I think the object debug lock might not be the only problem. It's just the only way the issue manifests in the example project.
So the relevant functions in the less_direct_example are:
func _ready() -> void:
var promise = $JobQueue.post(some_slow_operation)
$TextureRect.texture = await promise.fulfilled
#...
func fulfill(value) -> void:
result = value
fulfilled.emit(result)
The fulfill function does nothing after emitting the "fulfilled" signal, so fixing the object debug lock would fix this example. But there could be anything there after emit returns. If it assigned a value to a property or called another method wouldn't this access unallocated memory (in release builds, too)?
func fulfill(value) -> void:
result = value
fulfilled.emit(result)
# now self has been freed
result = null # oops, write-after-free
I think the object debug lock might not be the only problem. It's just the only way the issue manifests in the example project.
I know next to zero about the await command, and I may be wrong, but this does sound like a separate issue, and should probably be reported in separate issue with MRP.
|
gharchive/issue
| 2023-04-04T20:28:23 |
2025-04-01T04:34:23.258022
|
{
"authors": [
"RandomShaper",
"akien-mga",
"lawnjelly",
"tcoxon"
],
"repo": "godotengine/godot",
"url": "https://github.com/godotengine/godot/issues/75658",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
}
|
214458242
|
Visual Scripting broken on latest builds?
I noticed that having a visual script attached to any node, even if it doesn't contain any complicated logic, will cause the game window to close immediately after running the project without any error messages in the debugger. This must be a new issue, since i remember visual script working just fine a couple of weeks back.
Steps to reproduce:
create a new project and add a node to it
attach a visual script to it and add some logic (or none at all)
run the project
Seen on MacOS el capitan 10.11.6
Can't reproduce this anymore, so seems fixed.
Can't even try to reproduce this, because the new "Multiscript"-Entry in the latest build won't let me create a new Vscript and i can't select the "visual script" entry directly. I tried an older project with an existing visual script and I'm getting the same behavior as before (project window closing instantly)
Can't even try to reproduce this, because the new "Multiscript"-Entry in the latest build won't let me create a new Vscript and i can't select the "visual script" entry directly.
cc @karroffel
@Hinsbart I dont't think this is fixed. I just compiled a fresh built of godot master (with multiscript disabled for testing purposes) and can still reproduce this. I also haven't seen any commits lately, that would have adressed this issue. I'm appending my console output here:
running: /opt/godot-bin/godot3 -path /home/alex/Godot/vs_test -rdebug localhost:6007 -epid 16925 -p 0x0 res://Node2D.tscn
ERROR: load: No loader found for resource: res://test.vs
At: core/io/resource_loader.cpp:213.
ERROR: poll: res://Node2D.tscn:3 - Parse Error: [ext_resource] referenced nonexistent resource at: res://test.vs
At: scene/resources/scene_format_text.cpp:182.
ERROR: load: Condition ' err != OK ' is true. returned: RES()
At: core/io/resource_loader.cpp:144.
ERROR: load: Failed loading resource: res://Node2D.tscn
At: core/io/resource_loader.cpp:213.
ERROR: start: Failed loading scene: res://Node2D.tscn
At: main/main.cpp:1460.
shit?: Connection reset by peer
ERROR: read: Server disconnected!
At: drivers/unix/stream_peer_tcp_posix.cpp:275.
@novemberist can you test if VS was broken before the MultiScript PR as well?
@karroffel I can't test it right now, but I am pretty sure that it was, since I've been testing visual script regularly ever since i opened this issue back in march and it hasn't worked ever since.
Found this between lot of spam, trying on windows a build before multiscript
path not recognized
ERROR: No loader found for resource: res://visualscript.vs
Seems like it's still broken and not related to the MultiScript PR.
Awesome. Thank you very much.
|
gharchive/issue
| 2017-03-15T16:59:37 |
2025-04-01T04:34:23.264470
|
{
"authors": [
"Hinsbart",
"akien-mga",
"algorev",
"eon-s",
"karroffel",
"novemberist",
"vnen"
],
"repo": "godotengine/godot",
"url": "https://github.com/godotengine/godot/issues/8040",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
}
|
2010813792
|
Segfault in AnimationMixer::_blend_apply
Godot version
git master [1ba920f]
System information
Arch Linux, KDE 5, Radeon RX6800XT, X11
Issue description
The act of shooting a Plasma Gun in Liblast causes the engine to segfault not long after.
The backtrace point to AnimationMixer.
I a,m seeing a lot of errors related to animations for that weapon, but I don't expect the engine to segfault from that.
Backtrace:
Thread 1 "godot.linuxbsd." received signal SIGSEGV, Segmentation fault.
StringName::hash (this=<optimized out>) at ./core/string/string_name.h:136
136 return _data->hash;
(gdb) bt
#0 StringName::hash (this=<optimized out>) at ./core/string/string_name.h:136
#1 HashMapHasherDefault::hash (p_string_name=...) at ./core/templates/hashfuncs.h:314
#2 HashMap<StringName, ClassDB::ClassInfo, HashMapHasherDefault, HashMapComparatorDefault<StringName>, DefaultTypedAllocator<HashMapElement<StringName, ClassDB::ClassInfo> > >::_hash (p_key=..., this=0x55555c189540 <ClassDB::classes>)
at ./core/templates/hash_map.h:85
#3 HashMap<StringName, ClassDB::ClassInfo, HashMapHasherDefault, HashMapComparatorDefault<StringName>, DefaultTypedAllocator<HashMapElement<StringName, ClassDB::ClassInfo> > >::_lookup_pos (p_key=..., r_pos=<synthetic pointer>: <optimized out>,
this=0x55555c189540 <ClassDB::classes>) at ./core/templates/hash_map.h:106
#4 HashMap<StringName, ClassDB::ClassInfo, HashMapHasherDefault, HashMapComparatorDefault<StringName>, DefaultTypedAllocator<HashMapElement<StringName, ClassDB::ClassInfo> > >::getptr (this=0x55555c189540 <ClassDB::classes>, p_key=...)
at ./core/templates/hash_map.h:300
#5 0x00005555599a0773 in ClassDB::set_property (p_object=p_object@entry=0x555573eb7600, p_property=..., p_value=...,
r_valid=r_valid@entry=0x0) at core/object/class_db.cpp:1208
#6 0x00005555599c1dd4 in Object::set (this=this@entry=0x555573eb7600, p_name=..., p_value=..., r_valid=0x0)
at core/object/object.cpp:257
#7 0x00005555599c39e9 in Object::set_indexed (this=this@entry=0x555573eb7600, p_names=..., p_value=...,
r_valid=r_valid@entry=0x0) at core/object/object.cpp:407
#8 0x0000555558320300 in AnimationMixer::_blend_apply (this=this@entry=0x555573eb7fc0)
at scene/animation/animation_mixer.cpp:1711
#9 0x0000555558334df1 in AnimationMixer::_process_animation (this=0x555573eb7fc0, p_delta=0.017045999999999999,
p_update_only=<optimized out>) at scene/animation/animation_mixer.cpp:940
#10 0x0000555558377d33 in AnimationMixer::_notificationv (p_reversed=false, p_notification=25, this=0x555573eb7fc0)
at scene/animation/animation_mixer.h:44
#11 AnimationPlayer::_notificationv (this=0x555573eb7fc0, p_notification=25, p_reversed=false)
at scene/animation/animation_player.h:39
#12 0x00005555599bcdf0 in Object::notification (this=0x555573eb7fc0, p_notification=25, p_reversed=false)
at core/object/object.cpp:837
#13 0x0000555557b09c91 in SceneTree::_process_group (this=0x5555601349d0, p_group=0x555560134c28, p_physics=false)
at scene/main/scene_tree.cpp:948
#14 0x0000555557b0a372 in SceneTree::_process (this=this@entry=0x5555601349d0, p_physics=p_physics@entry=false)
at scene/main/scene_tree.cpp:1028
#15 0x0000555557b0a952 in SceneTree::process (this=0x5555601349d0, p_time=0.017045999999999999) at scene/main/scene_tree.cpp:508
--Type <RET> for more, q to quit, c to continue without paging--c
#16 0x0000555555b21b65 in Main::iteration () at main/main.cpp:3636
#17 0x0000555555ab5531 in OS_LinuxBSD::run (this=this@entry=0x7fffffffd280) at platform/linuxbsd/os_linuxbsd.cpp:933
#18 0x0000555555aa4f58 in main (argc=<optimized out>, argv=0x7fffffffd858) at platform/linuxbsd/godot_linuxbsd.cpp:74
Steps to reproduce
Clone Liblast repository:
git clone https://codeberg.org/Liblast/Liblast.git
cd Liblast
git lfs install
git pull
Open the game in Godot 4.2 editor to import all media (if they all fail, check LFS again)
After the game is imported, run via GDB automatically starting a game:
GODOT_BINARY="./godot.linuxbsd.editor.x86_64" # replace with your debug godot binary
gdb -ex run --args $GODOT_BINARY --path ./Game/ --host CP1
Once the map loads, click on the "start now" button to avoid waiting 10 sec.
After that, the bots will do their thing and shoot their Plasma Guns causing a crash.
The crash seems to most often happen when shooting that weapon - hitscan weapons don't seem to cause it.
Minimal reproduction project
I am afraid I don't have one just yet.
I found another segfault, and I am wondering if it could be related:
Run this command to reproduce:
GODOT_BINARY="./godot.linuxbsd.editor.x86_64" # replace with your debug godot binary
gdb -ex run --args $GODOT_BINARY --headless --path ./Game/ --dedicated-host CP1
Very quickly this causes a segfault after the game level is loaded:
Thread 1 "godot.linuxbsd." received signal SIGSEGV, Segmentation fault.
HashSet<RendererSceneCull::Instance*, HashMapHasherDefault, HashMapComparatorDefault<RendererSceneCull::Instance*> >::insert (this=this@entry=0x38, p_key=@0x7fffffffb9f8: 0x5555665dd838) at ./core/templates/hash_set.h:408
408 uint32_t pos = _insert(p_key);
(gdb) bt
#0 HashSet<RendererSceneCull::Instance*, HashMapHasherDefault, HashMapComparatorDefault<RendererSceneCull::Instance*> >::insert (this=this@entry=0x38, p_key=@0x7fffffffb9f8: 0x5555665dd838) at ./core/templates/hash_set.h:408
#1 0x0000555559316b14 in RendererSceneCull::instance_geometry_set_lightmap (this=0x55555ca40270, p_instance=..., p_lightmap=..., p_lightmap_uv_scale=..., p_slice_index=0) at servers/rendering/renderer_scene_cull.cpp:1461
#2 0x0000555557ff5de7 in LightmapGI::_assign_lightmaps (this=0x55556a26fb00) at scene/3d/lightmap_gi.cpp:1337
#3 0x00005555599bcdf0 in Object::notification (this=0x55556a26fb00, p_notification=27, p_reversed=false) at core/object/object.cpp:837
#4 0x0000555557ac0d4a in Node::_propagate_ready (this=0x55556a26fb00) at scene/main/node.cpp:227
#5 0x0000555557ac0d24 in Node::_propagate_ready (this=this@entry=0x55555df01b40) at scene/main/node.cpp:222
#6 0x0000555557ac4417 in Node::_set_tree (this=0x55555df01b40, p_tree=0x55555da009d0) at scene/main/node.cpp:2938
#7 0x0000555557ad7396 in Node::_add_child_nocheck (this=0x55555e0bb8e0, p_child=<optimized out>, p_name=..., p_internal_mode=Node::INTERNAL_MODE_DISABLED) at scene/main/node.cpp:1394
#8 0x0000555557aeacbc in call_with_variant_args_helper<__UnexistingClass, Node*, bool, Node::InternalMode, 0ul, 1ul, 2ul> (r_error=..., p_args=0x7fffffffbc80, p_method=<optimized out>, p_instance=<optimized out>) at ./core/variant/binder_common.h:303
#9 call_with_variant_args_dv<__UnexistingClass, Node*, bool, Node::InternalMode> (default_values=..., r_error=..., p_argcount=<optimized out>, p_args=<optimized out>, p_method=<optimized out>, p_instance=<optimized out>) at ./core/variant/binder_common.h:450
#10 MethodBindT<Node*, bool, Node::InternalMode>::call (this=<optimized out>, p_object=<optimized out>, p_args=<optimized out>, p_arg_count=<optimized out>, r_error=...) at ./core/object/method_bind.h:335
#11 0x0000555555e2d700 in GDScriptFunction::call (this=<optimized out>, p_instance=<optimized out>, p_instance@entry=0x55555dfa27b0, p_args=p_args@entry=0x7fffffffc3a8, p_argcount=<optimized out>, r_err=..., p_state=<optimized out>) at modules/gdscript/gdscript_vm.cpp:1796
#12 0x0000555555ce9f2f in GDScriptInstance::callp (this=0x55555dfa27b0, p_method=..., p_args=<optimized out>, p_argcount=<optimized out>, r_error=...) at modules/gdscript/gdscript.cpp:1937
#13 0x00005555599bf01c in Object::callp (this=0x55555e0bb8e0, p_method=..., p_args=0x7fffffffc3a8, p_argcount=1, r_error=...) at core/object/object.cpp:753
#14 0x000055555976337a in Variant::callp (this=<optimized out>, p_method=..., p_args=0x7fffffffc3a8, p_argcount=1, r_ret=..., r_error=...) at core/variant/variant_call.cpp:1168
#15 0x0000555555e404dc in GDScriptFunction::call (this=<optimized out>, p_instance=<optimized out>, p_instance@entry=0x55555dfa27b0, p_args=p_args@entry=0x7fffffffc948, p_argcount=<optimized out>, r_err=..., p_state=<optimized out>) at modules/gdscript/gdscript_vm.cpp:1704
#16 0x0000555555ce9f2f in GDScriptInstance::callp (this=0x55555dfa27b0, p_method=..., p_args=<optimized out>, p_argcount=<optimized out>, r_error=...) at modules/gdscript/gdscript.cpp:1937
#17 0x00005555599bf01c in Object::callp (this=0x55555e0bb8e0, p_method=..., p_args=0x7fffffffc948, p_argcount=1, r_error=...) at core/object/object.cpp:753
#18 0x000055555976337a in Variant::callp (this=<optimized out>, p_method=..., p_args=0x7fffffffc948, p_argcount=1, r_ret=..., r_error=...) at core/variant/variant_call.cpp:1168
#19 0x0000555555e404dc in GDScriptFunction::call (this=<optimized out>, p_instance=<optimized out>, p_instance@entry=0x55555dfa27b0, p_args=p_args@entry=0x0, p_argcount=<optimized out>, r_err=..., p_state=<optimized out>) at modules/gdscript/gdscript_vm.cpp:1704
#20 0x0000555555ce9f2f in GDScriptInstance::callp (this=0x55555dfa27b0, p_method=..., p_args=<optimized out>, p_argcount=<optimized out>, r_error=...) at modules/gdscript/gdscript.cpp:1937
#21 0x0000555557ad631b in Node::_gdvirtual__ready_call<false> (this=0x55555e0bb8e0) at scene/main/node.h:322
#22 Node::_notification (this=0x55555e0bb8e0, p_notification=<optimized out>) at scene/main/node.cpp:187
#23 0x000055555806550d in Node::_notificationv (p_reversed=false, p_notification=13, this=0x55555e0bb8e0) at ./scene/main/node.h:49
#24 Node3D::_notificationv (this=0x55555e0bb8e0, p_notification=13, p_reversed=<optimized out>) at scene/3d/node_3d.h:52
#25 0x00005555599bcdf0 in Object::notification (this=0x55555e0bb8e0, p_notification=13, p_reversed=false) at core/object/object.cpp:837
#26 0x0000555557ac0d86 in Node::_propagate_ready (this=this@entry=0x55555e0bb8e0) at scene/main/node.cpp:231
#27 0x0000555557ac4417 in Node::_set_tree (this=0x55555e0bb8e0, p_tree=0x55555da009d0) at scene/main/node.cpp:2938
#28 0x0000555557ad7396 in Node::_add_child_nocheck (this=0x55555da01560, p_child=<optimized out>, p_name=..., p_internal_mode=Node::INTERNAL_MODE_DISABLED) at scene/main/node.cpp:1394
#29 0x0000555557aeacbc in call_with_variant_args_helper<__UnexistingClass, Node*, bool, Node::InternalMode, 0ul, 1ul, 2ul> (r_error=..., p_args=0x7fffffffcea0, p_method=<optimized out>, p_instance=<optimized out>) at ./core/variant/binder_common.h:303
#30 call_with_variant_args_dv<__UnexistingClass, Node*, bool, Node::InternalMode> (default_values=..., r_error=..., p_argcount=<optimized out>, p_args=<optimized out>, p_method=<optimized out>, p_instance=<optimized out>) at ./core/variant/binder_common.h:450
#31 MethodBindT<Node*, bool, Node::InternalMode>::call (this=<optimized out>, p_object=<optimized out>, p_args=<optimized out>, p_arg_count=<optimized out>, r_error=...) at ./core/object/method_bind.h:335
#32 0x00005555599beeb7 in Object::callp (this=0x55555da01560, p_method=..., p_args=0x7fffffffd000, p_argcount=<optimized out>, r_error=...) at core/object/object.cpp:775
#33 0x000055555970bdc4 in Callable::callp (this=0x55555d7994c8, p_arguments=0x7fffffffd000, p_argcount=1, r_return_value=..., r_call_error=...) at core/variant/callable.cpp:69
#34 0x00005555599b73a2 in CallQueue::_call_function (this=this@entry=0x55555c5f6580, p_callable=..., p_args=p_args@entry=0x55555d7994e0, p_argcount=1, p_show_error=true) at core/object/message_queue.cpp:219
#35 0x00005555599ba90d in CallQueue::flush (this=0x55555c5f6580) at core/object/message_queue.cpp:324
#36 0x0000555557b0a65b in SceneTree::physics_process (this=0x55555da009d0, p_time=0.016666666666666666) at scene/main/scene_tree.cpp:471
#37 0x0000555555b21a74 in Main::iteration () at main/main.cpp:3598
#38 0x0000555555ab5531 in OS_LinuxBSD::run (this=this@entry=0x7fffffffd270) at platform/linuxbsd/os_linuxbsd.cpp:933
#39 0x0000555555aa4f58 in main (argc=<optimized out>, argv=0x7fffffffd848) at platform/linuxbsd/godot_linuxbsd.cpp:7
cc @TokageItLab
I can't tell anything without MRP. Guessing from the stack trace, there possibly could be a problem with the Continuous interpolation of StringName https://github.com/godotengine/godot/pull/84815. But I have no idea yet.
I can't tell anything without MRP.
The full project is open source and reproduces the crash, I just tested to confirm following the steps to reproduce in the OP.
So that can be used to debug, even if it's a bit heavy, at least it lets you get into a debugger and inspect what is wrong.
I can't tell anything without MRP.
Guessing from the stack trace, there possibly could be a problem with the Continuous interpolation of StringName #84815. Or it could be hashmap-related, and the cache could be corrupted when switching animations by method track or something.
In any case, I cannot say for sure since we do not know what you are doing.
Hi! I always try to create a minimal reproduction project for the bugs we find on Liblast, which is made in a big part to test the engine and help with the reporting, testing, feedback etc. However, we are mostly 2 learning devs and reproducing some of this more technical issues is out of my scope, as sometimes i'm unable to separate the problematic part of the game from the other parts without the project breaking in other way or i can not get the same error while trying to reproduce them from scratch because i can not replicate it. So i apology for the not ideal conditions, but please have in mind this is not because i didn't bother to, but because i could not.
While i understand the project is heavy, it is totally open source and so is the toolchain used, so it can be used for this scenarios without any trouble with licensing and having the whole picture available, from the asset sources to the final project, so complex issues can be tracked to any point. For testing purposes, all you need from the repo is contained inside the game folder, so you can get that and skip all the other files, which are a big part of the total size.
Yeah, I totally agree with you, @yosoyfreeman - the size of Liblast as a project allows us to find such weird engine bugs - thanks to the relative complexity, but it also makes it harder to isolate the issue.
@akien-mga I will try to do that, and see what can I get.
I have a test scene that shoots a weapon (with no physical character holding it) - let's see how that goes.
Ah, I had never heard that there was such an open source project. But it takes a lot of effort to isolate the source of a bug from a large project, so the reproduction project is the smaller the better.
BTW, this may be a similar issue?
https://github.com/godotengine/godot/issues/85388
@TokageItLab Here's an MRP:
PlasmaSecondaryCrash MRP.zip
I've ripped out the guts from Liblast and managed to stop all the bleeding for long enough to reproduce this crash.
@unfa Thanks! I tried that, but it crashed even without playing the animation, so the animation may not be the actual cause..
BTW, this may be a similar/related issue?
#85388
I tested this hypothesis, it doesn't fix this crash, but indeed the final crash location might be related. In some logs I've seen StringName::hash(), though gdb doesn't go all the way there for some reason for me now.
The PlasmaSecondaryCrash MRP crashes for me with this stacktrace:
(gdb) bt
#0 0x0000000000000621 in ?? ()
#1 0x0000000009e58bc4 in Object::set (this=0xd911770, p_name=..., p_value=..., r_valid=0x0) at ./core/object/object.cpp:301
#2 0x0000000009e5905e in Object::set_indexed (this=0xd911770, p_names=..., p_value=..., r_valid=0x0) at ./core/object/object.cpp:407
#3 0x000000000870371c in AnimationMixer::_blend_apply (this=0xdd62320) at ./scene/animation/animation_mixer.cpp:1711
#4 0x00000000086fc5d9 in AnimationMixer::_process_animation (this=0xdd62320, p_delta=0.016666666666666666, p_update_only=false) at ./scene/animation/animation_mixer.cpp:940
#5 0x0000000008705014 in AnimationMixer::_notification (this=0xdd62320, p_what=25) at ./scene/animation/animation_mixer.cpp:2034
#6 0x0000000008731a92 in AnimationMixer::_notificationv (this=0xdd62320, p_notification=25, p_reversed=false) at ./scene/animation/animation_mixer.h:44
#7 0x00000000087405d5 in AnimationPlayer::_notificationv (this=0xdd62320, p_notification=25, p_reversed=false) at ./scene/animation/animation_player.h:39
#8 0x0000000009e5abda in Object::notification (this=0xdd62320, p_notification=25, p_reversed=false) at ./core/object/object.cpp:837
#9 0x0000000007e096e6 in SceneTree::_process_group (this=0xd39ea10, p_group=0xd39ec68, p_physics=false) at ./scene/main/scene_tree.cpp:948
#10 0x0000000007e09c85 in SceneTree::_process (this=0xd39ea10, p_physics=false) at ./scene/main/scene_tree.cpp:1028
#11 0x0000000007e07bc4 in SceneTree::process (this=0xd39ea10, p_time=0.016666666666666666) at ./scene/main/scene_tree.cpp:508
#12 0x000000000579a28c in Main::iteration () at main/main.cpp:3636
#13 0x0000000005713a0e in OS_LinuxBSD::run (this=0x7fffffffcfb0) at platform/linuxbsd/os_linuxbsd.cpp:933
#14 0x000000000570c480 in main (argc=1, argv=0x7fffffffd598) at platform/linuxbsd/godot_linuxbsd.cpp:74
@unfa Thanks! I tried that, but it crashed even without playing the animation, so the animation may not be the actual cause..
What stacktrace do you get? There might be several bugs, but at least the crash originating in AnimationMixer::_blend_apply seems fairly easy to replicate, and even if it's crashing in StringName::hash(), it is likely due to invalid data being passed by AnimationMixer.
Since the backtrace isn't super helpful to really understand what's causing the crash exactly, I added some good old printf:
diff --git a/core/object/object.cpp b/core/object/object.cpp
index 40df13849b..6437ebca05 100644
--- a/core/object/object.cpp
+++ b/core/object/object.cpp
@@ -298,6 +298,7 @@ void Object::set(const StringName &p_name, const Variant &p_value, bool *r_valid
#endif
// Something inside the object... :|
+ print_line(vformat("Object: %s\tProperty: %s\tValue: %s", to_string(), p_name, p_value.operator String()));
bool success = _setv(p_name, p_value);
if (success) {
if (r_valid) {
That spams a lot, but lets me see what object/property/value combo it's crashing on (which gdb seemed to have trouble extracting, or I don't know how to use it :D):
Object: Explosion:<Node3D#60414756587> Property: damage_amount Value: 25
Object: Explosion:<Node3D#60414756587> Property: push_force Value: 12
Object: Explosion:<Node3D#60414756587> Property: blast_radius Value: 2.5
Object: Explosion:<Node3D#60414756587> Property: blast_duration Value: 0.1
Object: <AnimationPlayer#60817409893> Property: libraries Value: { "": <AnimationLibrary#-9223371986724452699> }
Object: Halo:<Node3D#60834187110> Property: SoundClip Value: res://Assets/SFX/Weapons_BulletHitWall_01.wav
Object: Halo:<Node3D#60834187110> Property: RandomizePitch Value: 0.1
WARNING: Node 'SoundPlayer/AudioStreamPlayer3D' was modified from inside an instance, but it has vanished.
at: instantiate (./scene/resources/packed_scene.cpp:231)
Object: <AnimationPlayer#60884518761> Property: libraries Value: { "": <AnimationLibrary#-9223371986640566614> }
================================================================
handle_crash: Program crashed with signal 11
Reproduced a few times, it seems pretty consistent (with different Node IDs, but somehow the same numbers of AnimationLibrary).
The warning also appears before in the logs, but seems to consistently appear before the last set_indexed on libraries.
Here's the full output:
output.log
Tested again with the full Liblast project (and #85388 merged locally as it seemed possibly related), and I still get the crash.
With my tweaked build using the above diff, I get these logs:
setv: Object: Flash:<MeshInstance3D#1793031418085> Property: shader_uniforms/Energy Value: (0, 0, 0, 1)
script_instance->set: Object: -10:<CharacterBody3D#1784173045998> Property: global_transform Value: [X: (-0.998903, 0, -0.046833), Y: (0, 1, 0), Z: (0.046833, 0, -0.998903), O: (24.17453, 8.001047, -22.72682)]
script_instance->set: Object: <RefCounted#-9223367204194477496> Property: aim Value: (0.005752, -0.002225)
script_instance->set: Object: <RefCounted#-9223367204194477496> Property: index Value: 293
script_instance->set: Object: <RefCounted#-9223367204194477496> Property: frame Value: 3053
script_instance->set: Object: -10:<CharacterBody3D#1784173045998> Property: global_transform Value: [X: (-0.999156, 0, -0.041087), Y: (0, 1, 0), Z: (0.041087, 0, -0.999156), O: (24.17453, 8.001047, -22.72682)]
Thread 1 "godot-git" received signal SIGSEGV, Segmentation fault.
0x0000007400000070 in ?? ()
(gdb) bt
#0 0x0000007400000070 in ?? ()
#1 0x0000000009e58b24 in Object::set (this=0x20152d50, p_name=..., p_value=..., r_valid=0x0) at ./core/object/object.cpp:245
#2 0x0000000009e59418 in Object::set_indexed (this=0x20152d50, p_names=..., p_value=..., r_valid=0x0) at ./core/object/object.cpp:412
#3 0x000000000870379c in AnimationMixer::_blend_apply (this=0x20153370) at ./scene/animation/animation_mixer.cpp:1711
#4 0x00000000086fc659 in AnimationMixer::_process_animation (this=0x20153370, p_delta=0.13266700000000001, p_update_only=false) at ./scene/animation/animation_mixer.cpp:940
#5 0x0000000008705094 in AnimationMixer::_notification (this=0x20153370, p_what=25) at ./scene/animation/animation_mixer.cpp:2034
#6 0x0000000008731b12 in AnimationMixer::_notificationv (this=0x20153370, p_notification=25, p_reversed=false) at ./scene/animation/animation_mixer.h:44
#7 0x0000000008740655 in AnimationPlayer::_notificationv (this=0x20153370, p_notification=25, p_reversed=false) at ./scene/animation/animation_player.h:39
#8 0x0000000009e5af94 in Object::notification (this=0x20153370, p_notification=25, p_reversed=false) at ./core/object/object.cpp:842
#9 0x0000000007e09766 in SceneTree::_process_group (this=0xd9cabe0, p_group=0xd9cae38, p_physics=false) at ./scene/main/scene_tree.cpp:948
#10 0x0000000007e09d05 in SceneTree::_process (this=0xd9cabe0, p_physics=false) at ./scene/main/scene_tree.cpp:1028
#11 0x0000000007e07c44 in SceneTree::process (this=0xd9cabe0, p_time=0.13266700000000001) at ./scene/main/scene_tree.cpp:508
#12 0x000000000579a30c in Main::iteration () at main/main.cpp:3636
#13 0x0000000005713a8e in OS_LinuxBSD::run (this=0x7fffffffcf70) at platform/linuxbsd/os_linuxbsd.cpp:933
#14 0x000000000570c500 in main (argc=5, argv=0x7fffffffd558) at platform/linuxbsd/godot_linuxbsd.cpp:74
That object.cpp line in my build is this:
It's starting to sound related to #85155.
Overall this kind of crash is likely some dangling pointers to invalid data or memory corruption.
I tested the PlasmaSecondaryCrash_noanim.zip project with ASAN build on Windows 10 on current master (eda44bfe109d7f15c5a965db2d65c8a17ce4b7d0). The crash (heap-use-after-free) happens here because the object this pointer has been previously freed:
https://github.com/godotengine/godot/blob/eda44bfe109d7f15c5a965db2d65c8a17ce4b7d0/core/object/object.cpp#L226
The more relevant call is a few stack levels below in AnimationMixer::_blend_apply(). It crashes because t->object has been freed previously and now points to invalid (freed) memory:
https://github.com/godotengine/godot/blob/eda44bfe109d7f15c5a965db2d65c8a17ce4b7d0/scene/animation/animation_mixer.cpp#L1711
More detailed stack report below, uncollapse to see it. Seems like one potential hint is that the memory was allocated in Object* ClassDB::creator<OmniLight3D>(). Sometimes it changes to creator<GPUParticlesAttractorSphere3D>().
ASAN crash report
ERROR: AddressSanitizer: heap-use-after-free on address 0x12b85b6509fa at pc 0x7ff62b63bf3a bp 0x00df5a9fb020 sp 0x00df5a9fb068
WRITE of size 1 at 0x12b85b6509fa thread T0
#0 0x7ff62b63bf39 in Object::set(StringName const&, Variant const&, bool*) E:/Repositories/godot/core\object/object.cpp:226:10
#1 0x7ff62b6400f6 in Object::set_indexed(Vector<StringName> const&, Variant const&, bool*) E:/Repositories/godot/core\object/object.cpp:407:3
#2 0x7ff636480476 in AnimationMixer::_blend_apply() E:/Repositories/godot/scene\animation/animation_mixer.cpp:1711:16
#3 0x7ff636450e21 in AnimationMixer::_process_animation(double, bool) E:/Repositories/godot/scene\animation/animation_mixer.cpp:940:3
#4 0x7ff636490e87 in AnimationMixer::_notification(int) E:/Repositories/godot/scene\animation/animation_mixer.cpp:2034:5
#5 0x7ff6364cc61c in AnimationMixer::_notificationv(int, bool) E:/Repositories/godot/scene\animation/animation_mixer.h:44:2
#6 0x7ff636560aac in AnimationPlayer::_notificationv(int, bool) E:/Repositories/godot/scene\animation/animation_player.h:39:2
#7 0x7ff62b63a99c in Object::notification(int, bool) E:/Repositories/godot/core\object/object.cpp:837:3
#8 0x7ff62e5f6dd1 in SceneTree::_process_group(SceneTree::ProcessGroup*, bool) E:/Repositories/godot/scene\main/scene_tree.cpp:948:8
#9 0x7ff62e5e3e0b in SceneTree::_process(bool) E:/Repositories/godot/scene\main/scene_tree.cpp:1028:8
#10 0x7ff62e5ea1d3 in SceneTree::process(double) E:/Repositories/godot/scene\main/scene_tree.cpp:508:2
#11 0x7ff629b34c74 in Main::iteration() E:/Repositories/godot/main/main.cpp:3636:44
#12 0x7ff6298e21b4 in OS_Windows::run() E:/Repositories/godot/platform\windows/os_windows.cpp:1474:7
#13 0x7ff6298c2c2f in widechar_main(int, wchar_t**) E:/Repositories/godot/platform\windows/godot_windows.cpp:180:6
#14 0x7ff6298c30dd in _main() E:/Repositories/godot/platform\windows/godot_windows.cpp:204:11
#15 0x7ff6298c335f in main E:/Repositories/godot/platform\windows/godot_windows.cpp:223:9
#16 0x7ff6298c1314 in __tmainCRTStartup /home/runner/work/llvm-mingw/llvm-mingw/mingw-w64/mingw-w64-crt/build-x86_64/../crt/crtexe.c:267:15
#17 0x7ff6298c1155 in .l_startw /home/runner/work/llvm-mingw/llvm-mingw/mingw-w64/mingw-w64-crt/build-x86_64/../crt/crtexe.c:157:9
#18 0x7ff8b8a07343 (C:\WINDOWS\System32\KERNEL32.DLL+0x180017343)
#19 0x7ff8ba7e26b0 (C:\WINDOWS\SYSTEM32\ntdll.dll+0x1800526b0)
0x12b85b6509fa is located 122 bytes inside of 1528-byte region [0x12b85b650980,0x12b85b650f78)
freed by thread T0 here:
#0 0x7ffffd842671 in free /home/runner/work/llvm-mingw/llvm-mingw/llvm-project/compiler-rt/lib/asan/asan_malloc_win.cpp:82:3
#1 0x7ff62ab730ca in Memory::free_static(void*, bool) E:/Repositories/godot/core\os/memory.cpp:168:3
#2 0x7ff629b44eb9 in void memdelete<Object>(Object*) E:/Repositories/godot/./core/os/memory.h:112:2
#3 0x7ff62e5e8b9d in SceneTree::_flush_delete_queue() E:/Repositories/godot/scene\main/scene_tree.cpp:1345:4
#4 0x7ff62e5ea69b in SceneTree::process(double) E:/Repositories/godot/scene\main/scene_tree.cpp:516:2
#5 0x7ff629b34c74 in Main::iteration() E:/Repositories/godot/main/main.cpp:3636:44
#6 0x7ff6298e21b4 in OS_Windows::run() E:/Repositories/godot/platform\windows/os_windows.cpp:1474:7
#7 0x7ff6298c2c2f in widechar_main(int, wchar_t**) E:/Repositories/godot/platform\windows/godot_windows.cpp:180:6
#8 0x7ff6298c30dd in _main() E:/Repositories/godot/platform\windows/godot_windows.cpp:204:11
#9 0x7ff6298c335f in main E:/Repositories/godot/platform\windows/godot_windows.cpp:223:9
#10 0x7ff6298c1314 in __tmainCRTStartup /home/runner/work/llvm-mingw/llvm-mingw/mingw-w64/mingw-w64-crt/build-x86_64/../crt/crtexe.c:267:15
#11 0x7ff6298c1155 in .l_startw /home/runner/work/llvm-mingw/llvm-mingw/mingw-w64/mingw-w64-crt/build-x86_64/../crt/crtexe.c:157:9
#12 0x7ff8b8a07343 (C:\WINDOWS\System32\KERNEL32.DLL+0x180017343)
#13 0x7ff8ba7e26b0 (C:\WINDOWS\SYSTEM32\ntdll.dll+0x1800526b0)
previously allocated by thread T0 here:
#0 0x7ffffd842791 in malloc /home/runner/work/llvm-mingw/llvm-mingw/llvm-project/compiler-rt/lib/asan/asan_malloc_win.cpp:98:3
#1 0x7ff62ab7186b in Memory::alloc_static(unsigned long long, bool) E:/Repositories/godot/core\os/memory.cpp:75:14
#2 0x7ff62ab716dc in operator new(unsigned long long, char const*) E:/Repositories/godot/core\os/memory.cpp:40:9
#3 0x7ff62c53dc4a in Object* ClassDB::creator<OmniLight3D>() E:/Repositories/godot/./core/object/class_db.h:144:10
#4 0x7ff62b584ace in ClassDB::instantiate(StringName const&) E:/Repositories/godot/core\object/class_db.cpp:377:10
#5 0x7ff62e912f22 in SceneState::instantiate(SceneState::GenEditState) const E:/Repositories/godot/scene\resources/packed_scene.cpp:237:18
#6 0x7ff62e921ec1 in PackedScene::instantiate(PackedScene::GenEditState) const E:/Repositories/godot/scene\resources/packed_scene.cpp:1946:19
#7 0x7ff62e9c4cd4 in void call_with_variant_args_retc_helper<__UnexistingClass, Node*, PackedScene::GenEditState, 0ull>(__UnexistingClass*, Node* (__UnexistingClass::*)(PackedScene::GenEditState) const, Variant const**, Variant&, Callable::CallError&, IndexSequence<0ull>) E:/Repositories/godot/./core/variant/binder_common.h:806:10
#8 0x7ff62e9c47e5 in void call_with_variant_args_retc_dv<__UnexistingClass, Node*, PackedScene::GenEditState>(__UnexistingClass*, Node* (__UnexistingClass::*)(PackedScene::GenEditState) const, Variant const**, int, Variant&, Callable::CallError&, Vector<Variant> const&) E:/Repositories/godot/./core/variant/binder_common.h:567:2
#9 0x7ff62e9c22ca in MethodBindTRC<Node*, PackedScene::GenEditState>::call(Object*, Variant const**, int, Callable::CallError&) const E:/Repositories/godot/./core/object/method_bind.h:583:3
#10 0x7ff6429ee572 in GDScriptFunction::call(GDScriptInstance*, Variant const**, int, Callable::CallError&, GDScriptFunction::CallState*) E:/Repositories/godot/modules\gdscript/gdscript_vm.cpp:1794:21
#11 0x7ff63c2542a5 in GDScriptInstance::callp(StringName const&, Variant const**, int, Callable::CallError&) E:/Repositories/godot/modules\gdscript/gdscript.cpp:1937:21
#12 0x7ff62b64afe0 in Object::callp(StringName const&, Variant const**, int, Callable::CallError&) E:/Repositories/godot/core\object/object.cpp:753:26
#13 0x7ff630df56ae in Variant::callp(StringName const&, Variant const**, int, Variant&, Callable::CallError&) E:/Repositories/godot/core\variant/variant_call.cpp:1168:27
#14 0x7ff6429ea6df in GDScriptFunction::call(GDScriptInstance*, Variant const**, int, Callable::CallError&, GDScriptFunction::CallState*) E:/Repositories/godot/modules\gdscript/gdscript_vm.cpp:1704:12
#15 0x7ff63c2542a5 in GDScriptInstance::callp(StringName const&, Variant const**, int, Callable::CallError&) E:/Repositories/godot/modules\gdscript/gdscript.cpp:1937:21
#16 0x7ff62b64afe0 in Object::callp(StringName const&, Variant const**, int, Callable::CallError&) E:/Repositories/godot/core\object/object.cpp:753:26
#17 0x7ff630df56ae in Variant::callp(StringName const&, Variant const**, int, Variant&, Callable::CallError&) E:/Repositories/godot/core\variant/variant_call.cpp:1168:27
#18 0x7ff6429ea6df in GDScriptFunction::call(GDScriptInstance*, Variant const**, int, Callable::CallError&, GDScriptFunction::CallState*) E:/Repositories/godot/modules\gdscript/gdscript_vm.cpp:1704:12
#19 0x7ff63c2542a5 in GDScriptInstance::callp(StringName const&, Variant const**, int, Callable::CallError&) E:/Repositories/godot/modules\gdscript/gdscript.cpp:1937:21
#20 0x7ff62b64afe0 in Object::callp(StringName const&, Variant const**, int, Callable::CallError&) E:/Repositories/godot/core\object/object.cpp:753:26
#21 0x7ff62b1313bf in Callable::callp(Variant const**, int, Variant&, Callable::CallError&) const E:/Repositories/godot/core\variant/callable.cpp:69:25
#22 0x7ff62b656742 in Object::emit_signalp(StringName const&, Variant const**, int) E:/Repositories/godot/core\object/object.cpp:1127:15
#23 0x7ff62ea65efd in Node::emit_signalp(StringName const&, Variant const**, int) E:/Repositories/godot/scene\main/node.cpp:3606:17
#24 0x7ff62e6210ef in Error Object::emit_signal<Node*>(StringName const&, Node*) E:/Repositories/godot/./core/object/object.h:922:10
#25 0x7ff6374befc7 in Area3D::_body_inout(int, RID const&, ObjectID, int, int) E:/Repositories/godot/scene\3d/area_3d.cpp:248:6
#26 0x7ff637507e5d in void call_with_variant_args_helper<Area3D, int, RID const&, ObjectID, int, int, 0ull, 1ull, 2ull, 3ull, 4ull>(Area3D*, void (Area3D::*)(int, RID const&, ObjectID, int, int), Variant const**, Callable::CallError&, IndexSequence<0ull, 1ull, 2ull, 3ull, 4ull>) E:/Repositories/godot/./core/variant/binder_common.h:303:2
#27 0x7ff63750767c in void call_with_variant_args<Area3D, int, RID const&, ObjectID, int, int>(Area3D*, void (Area3D::*)(int, RID const&, ObjectID, int, int), Variant const**, int, Callable::CallError&) E:/Repositories/godot/./core/variant/binder_common.h:417:2
SUMMARY: AddressSanitizer: heap-use-after-free E:/Repositories/godot/core\object/object.cpp:226:10 in Object::set(StringName const&, Variant const&, bool*)
Shadow bytes around the buggy address:
0x12b85b650700: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
0x12b85b650780: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
0x12b85b650800: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
0x12b85b650880: 00 00 00 fa fa fa fa fa fa fa fa fa fa fa fa fa
0x12b85b650900: fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa
=>0x12b85b650980: fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd[fd]
0x12b85b650a00: fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd
0x12b85b650a80: fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd
0x12b85b650b00: fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd
0x12b85b650b80: fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd
0x12b85b650c00: fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd
Shadow byte legend (one shadow byte represents 8 application bytes):
Addressable: 00
Partially addressable: 01 02 03 04 05 06 07
Heap left redzone: fa
Freed heap region: fd
Stack left redzone: f1
Stack mid redzone: f2
Stack right redzone: f3
Stack after return: f5
Stack use after scope: f8
Global redzone: f9
Global init order: f6
Poisoned by user: f7
Container overflow: fc
Array cookie: ac
Intra object redzone: bb
ASan internal: fe
Left alloca redzone: ca
Right alloca redzone: cb
==12152==ABORTING
Thanks @bitsawer! I could work around the issue in #85461, which should be "good enough" for 4.2 hopefully. At least it fixes Liblast.
I'll repurpose this issue to focus on the fact that AnimationMixer stores Object pointers in its TrackCache, and as seen here, this is really unsafe. We should remove object from the TrackCache and always get the object pointed from the stored object_id. But that's refactoring work for 4.3 (maybe 4.2.x cherry-pick if we find more crashes like this).
Thanks a lots @bitsawer @akien-mga! I don't remember how long track->object has existed (it probably already existed before 4.0), but a quick look at the code suggests that eliminating track->object is relatively straightforward.
Another crash likely caused by the same t->object problem: #85572.
We should aim to fix this sooner than later, ideally for 4.2.1.
|
gharchive/issue
| 2023-11-25T22:20:56 |
2025-04-01T04:34:23.293381
|
{
"authors": [
"Calinou",
"TokageItLab",
"akien-mga",
"bitsawer",
"unfa",
"yosoyfreeman"
],
"repo": "godotengine/godot",
"url": "https://github.com/godotengine/godot/issues/85365",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
}
|
2014383396
|
Class names aren't processed on plugin
Godot version
4.1.3-stable
System information
Windows 10
Issue description
Due to dependency issues, global_script_class_cache.cfg isn't being populated after installing plugin.
This is a regression issue in 4.1.3 (and probably 4.1.2 and 4.1.1) but already fixed in 4.2.
Steps to reproduce
This is easiest to reproduce with the TiltFive plugin, but also happens to other plugins such as XR Tools. You do not need a tilt five to reproduce the issue, but you do need Windows.
Create a new project
Open the asset library and search for TiltFive
Install the TiltFive plugin (Godot may crash, just reopen)
Errors should already be visible but they will disappear on restart (but aren't really gone)
Enable the TiltFive plugin, it's accessing classes again and thus new errors
Create a new scene and make this your startup scene
Add a Node3D to the scene, assign the res://addons/tiltfive/T5Manager.gd script to this node
Save your scene
Every time you load up the scene these errors are reported:
When you open up one of the effected scripts such as T5ProjectSettings you'll see that the class_name remains unresolved:
Opening up global_script_class_cache.cfg you will see none of the class names are added:
list=[]
The workaround is to edit each script that has class_name and rename the class, and then rename it back so it gets added.
Minimal reproduction project
n/a
Note also that the same issues are present when cloning an existing project that uses the tiltfive plugin, or when you delete the .godot folder to start from scratch (easy way to reproduce this on a working project).
Here is a full project that can be cloned and opened in Godot: https://github.com/GodotVR/TiltFiveTools
This is a regression issue in 4.1.3 (and probably 4.1.2 and 4.1.1) but already fixed in 4.2.
If it's already present in 4.1.x then there is no point marking it as a regression, it's just a bug introduced some time ago. 🙂
Yup, definately worked in 4.0, and as far as I'm aware in 4.1, broke sometime later, but seems to be fixed in 4.2
We're getting reports of this issue occurring in 4.2-stable, but I'm unable to reproduce:
@YuriSizov @akien-mga now that the Godot XR game jam has started I'm getting reports from people that this issue does exist in Godot 4.2 as well, so either we reverted something that fixed it, or there is still a condition where this happens.
I'm going to try and reproduce it myself to verify the claims, but it has been reported with the TiltFive plugin.
One theory I have, but I could be barking up the wrong tree here, is that I wonder if when a plugin is disabled, we stop the scripts in that plugin from being evaluated. As for TiltFive you actually install two plugins that have a dependency, and we can't enable the plugin if the plugin script can't be run, but it can't be run without the classes being registered, and we got ourselves a catch-22 situation.
@Malcolmnixon what stands out in that error log are the errors on classes T5Origin3D and T5Camera3D which are all classes implemented in gdextension. This feels like the extension not being loaded.
@Malcolmnixon what stands out in that error log are the errors on classes T5Origin3D and T5Camera3D which are all classes implemented in gdextension. This feels like the extension not being loaded.
I was thinking the same, but then they responded with:
threw a couple of scripts into autoload (Plugin.gd and T5Interface.gd as the second didn't fix the problem alone) and it still throws errors but now I at least have it to where I finally have it generating my first holographic cube! (kinda neat!)
Now to get back to the meat of what I need to start on for now
This suggests they're cognizant about the plugin enabling, and there's enough running that they can get the headset to work; so I'm still stumped.
|
gharchive/issue
| 2023-11-28T12:58:09 |
2025-04-01T04:34:23.306110
|
{
"authors": [
"BastiaanOlij",
"Malcolmnixon",
"YuriSizov"
],
"repo": "godotengine/godot",
"url": "https://github.com/godotengine/godot/issues/85467",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
}
|
358318503
|
Make _sort_list_on_update true before opening files, fixes 14400
Looks like this issue was due to _sort_list_on_update not being set back to true before _update_script_names() is called in the appropriate functions. I have also changed the deferred calls to immediate, which keeps the indices relevant.
closes #14400
Apologies for edits, mis-hit enter key while opening PR.
Thanks!
|
gharchive/pull-request
| 2018-09-08T18:38:08 |
2025-04-01T04:34:23.307976
|
{
"authors": [
"RyanStein",
"akien-mga"
],
"repo": "godotengine/godot",
"url": "https://github.com/godotengine/godot/pull/21878",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
}
|
542238407
|
Fixes tab height in Batch Rename dialog
Applies editor scale to minimum container height
Before this fix, tabs in Batch Rename dialog have different heights on a HiDPI monitor, which is different from the behavior on a normal DPI monitor.
Also, on a HiDPI monitor, the "Preview" section will be pushed outside the dialog if you:
Switch to the second tab
Click "Advanced Options" to close the tab area
Click "Advanced Options" again to open the tab area
Switch to the first tab
Thanks!
|
gharchive/pull-request
| 2019-12-25T02:40:12 |
2025-04-01T04:34:23.310512
|
{
"authors": [
"akien-mga",
"timothyqiu"
],
"repo": "godotengine/godot",
"url": "https://github.com/godotengine/godot/pull/34598",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
}
|
704946572
|
Move gltf to a module for future proposals.
Support for proposal https://github.com/godotengine/godot-proposals/issues/885
The goal of this pr is to try to make my work in progress https://github.com/godot-extended-libraries/gltf mergable. Also, this is how the new FBX importer works, so having the gltf2 importer be a module makes sense. The gltf2 doesn't have any internal dependencies so it can be moved out.
Thoughts on moving the collada and the escn importers out too?
I'm not sure I see much point in doing this. But let's see what @reduz thinks about it.
I can make a PR and move collada to a module too if you want @akien-mga we could do them all?
You should see with @reduz what he thinks of the proposal first.
Since there has been no movement, I'll work on implementing the whole proposal https://github.com/godotengine/godot/pull/34193. I was hoping this would be an easier merge.
|
gharchive/pull-request
| 2020-09-19T17:35:28 |
2025-04-01T04:34:23.313604
|
{
"authors": [
"RevoluPowered",
"akien-mga",
"fire"
],
"repo": "godotengine/godot",
"url": "https://github.com/godotengine/godot/pull/42198",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
}
|
994116273
|
Add editor keyboard shortcut for Mono Build solution button
Related issue: #39633
This adds a keyboard shortcut for the Build button in Godot Mono 3.x. Default is Alt+B (since other suggestions from the issue were taken by the script editor bookmark shortcuts).
This exposes the ED_GET_SHORTCUT function from modules\mono\editor\editor_internal_calls.cpp to the GodotTools C# solution, via a new GodotTools.Internals.Globals function: GodotTools.Internals.Globals::internal_EditorShortcut - allowing the GodotSharpEditor.cs EditorPlugin to retrieve editor shortcuts, which it does to attach a new script_editor/build_mono_solution shortcut definition to the mono Build button.
I couldn't get these changes to compile against master; firstly I got an error because ShortCut is now Shortcut, but then I got errors related to the mono/editor/code_completion.cpp file (unchanged):
modules\mono\editor\code_completion.cpp(126): error C3536: '<begin>$L1': cannot be used before it is initialized
modules\mono\editor\code_completion.cpp(126): error C3536: '<end>$L1': cannot be used before it is initialized
[Initial build] modules\mono\editor\code_completion.cpp(126): error C2100: illegal indirection
[Initial build] modules\mono\editor\code_completion.cpp(126): error C2440: 'initializing': cannot convert from 'int' to 'const KeyValue<StringName,ProjectSettings::AutoloadInfo> &'
[Initial build] modules\mono\editor\code_completion.cpp(126): note: Reason: cannot convert from 'int' to 'const KeyValue<StringName,ProjectSettings::AutoloadInfo>'
modules\mono\editor\code_completion.cpp(126): note: No constructor could take the source type, or constructor overload resolution was ambiguous
I'm unsure what the status of Mono is in Godot 4? So, I made these changes against 3.x since that's what I'm using, let me know if this needs to be moved to master, and if I need to change anything to do so.
Additionally, I've added the keyboard shortcut under the script_editor shortcut setting group, since that seemed most relevant, however this shortcut will be available even in non-Mono builds; is there a way to conditionally add the shortcut only in a Mono build?
Cheers.
Additionally, I've added the keyboard shortcut under the script_editor shortcut setting group, since that seemed most relevant, however this shortcut will be available even in non-Mono builds; is there a way to conditionally add the shortcut only in a Mono build?
I think the ED_SHORTCUT should be inside the mono module but I'm not sure where exactly.
For the implementation of passing ED_SHORTCUT I followed the style of the other methods for getting editor settings, however I only wrote a getter. I assume the function can also set editor shortcuts, allowing the Mono project to add it from C#, though I haven't tried, so I'll look into that; that would indeed solve it.
I didn't mean to set the editor shortcut from C#, just to have the ED_SHORTCUT macro inside the modules/mono directory since the shortcut only applies to the mono module.
I see, for clarity's sake, you mean to move the call to register the shortcut i.e.:
ED_SHORTCUT("script_editor/build_mono_solution", TTR("Build Mono Solution"), KEY_MASK_ALT | KEY_B);
Into one of the cpp files in modules/mono? And probably it'd be good to change the script_editor category to something more appropriate like mono as well.
I see, for clarity's sake, you mean to move the call to register the shortcut into one of the cpp files in modules/mono?
Yes, but I'm not sure where exactly.
And probably it'd be good to change the script_editor category to something more appropriate like mono as well.
I guess that would make sense.
After a bit of trial and error, I think this is about the right place for it:
https://github.com/godotengine/godot/blob/6841c89e1374c87e857a4da6820194f2ce803ffc/modules/mono/csharp_script.cpp#L1183
At this point the checks for Mono being enabled are done, and the cs editor plugin is about to be enabled. Any later than this in the process, and the plugin will try and lookup the shortcut before it's been registered. I have tested on a non-mono build, and the shortcut doesn't get registered as expected.
I fixed the shortcut hint label:
And the shortcut now sits in a new "mono" category of shortcuts:
I haven't got round to testing it with Godot 4 yet, but I will pull down your fix, @raulsntos, and see what happens, when I get a chance later. Thanks.
Btw, with my initial PR a review was auto requested from the @godotengine/script-editor group, as the original commit registered the editor shortcut in the script_editor_plugin.cpp file. I've since amended the PR so that the shortcut is added in the mono module - so I think that review request can be removed/dismissed.
Thanks!
|
gharchive/pull-request
| 2021-09-12T10:02:11 |
2025-04-01T04:34:23.324364
|
{
"authors": [
"akien-mga",
"lewiji",
"raulsntos"
],
"repo": "godotengine/godot",
"url": "https://github.com/godotengine/godot/pull/52595",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
}
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.