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 |
|---|---|---|---|---|---|
348284143 | max() arg is an empty sequence
Traceback (most recent call last):
File "Test.py", line 17, in
best_model = max(models, key=os.path.getctime)
ValueError: max() arg is an empty sequence
The model was trained with an error of 0.00022
but testing the model throws this error
Please use the issue template when reporting issues. It gives us the needed information to help solve your problem. For example, what version of keras / tensorflow / python / numpy / pandas are you using?
Probably a duplicate of this one. https://github.com/Microsoft/AutonomousDrivingCookbook/issues/63.
| gharchive/issue | 2018-08-07T11:41:01 | 2025-04-01T06:37:13.222124 | {
"authors": [
"mitchellspryn",
"vighnesh-balaji"
],
"repo": "Microsoft/AutonomousDrivingCookbook",
"url": "https://github.com/Microsoft/AutonomousDrivingCookbook/issues/59",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
147215736 | Cannot mv a file from inside of /tmp to non-/tmp and non-/ directory, for example, /root
Cannot move any file in /tmp to /root, for example.
My operation log is here.
=== === === === === ===
root@localhost:~# pwd
/root
root@localhost:~# touch /tmp/test
root@localhost:~# /bin/mv /tmp/test .
/bin/mv: cannot move ‘/tmp/test’ to a subdirectory of itself, ‘./test’
root@localhost:~#
=== === === === === ===
And I have found '/bin/mv /tmp/test /' works fine.
And '/bin/mv /tmp/test /tmp/test2' works fine too.
Thank you for the bug.
We have a fix internally on one of our dev branches. Should hit the insiders builds before too long.
root@RUSSALEX-DESK:~# pwd
/root
root@RUSSALEX-DESK:~# touch /tmp/test
root@RUSSALEX-DESK:~# mv /tmp/test .
root@RUSSALEX-DESK:~# ls | grep test
test
| gharchive/issue | 2016-04-10T11:17:35 | 2025-04-01T06:37:13.226114 | {
"authors": [
"russalex",
"ryoon"
],
"repo": "Microsoft/BashOnWindows",
"url": "https://github.com/Microsoft/BashOnWindows/issues/104",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
148019892 | Can't Install both Oracle JDK and OpenJDK due to so many reason.
Pattern 1:
add-apt-repository ppa:webupd8team/java
apt-get update
apt-get install oracle-java8-installer
172032K ........ ........ ........ ........ ........ ........ 98% 3.43M 0s
175104K ........ ........ ........ ....... 100% 4.51M=41s
2016-04-12 18:22:29 (4.17 MB/s) - ‘jdk-8u77-linux-x64.tar.gz’ saved [181365687/181365687]
Download done.
Removing outdated cached downloads...
sha256sum mismatch jdk-8u77-linux-x64.tar.gz
Oracle JDK 8 is NOT installed.
dpkg: error processing package oracle-java8-installer (--configure):
subprocess installed post-installation script returned error exit status 1
Setting up gsfonts (1:8.11+urwcyr1.0.7~pre44-4.2ubuntu1) ...
Setting up x11-common (1:7.7+1ubuntu8.1) ...
dmesg: klogctl failed: Invalid argument
/bin/df: cannot read table of mounted file systems: No such file or directory
initctl: Unable to connect to Upstart: Failed to connect to socket /com/ubuntu/upstart: No such file or directory
runlevel:/var/run/utmp: No such file or directory
Setting up X socket directories... [ OK ]
Processing triggers for ureadahead (0.100.0-16) ...
Setting up xfonts-encodings (1:1.0.4-1ubuntu1) ...
Setting up xfonts-utils (1:7.7+1) ...
Setting up gsfonts-x11 (0.22) ...
Processing triggers for libc-bin (2.19-0ubuntu6.7) ...
Errors were encountered while processing:
oracle-java8-installer
E: Sub-process /usr/bin/dpkg returned an error code (1)
Pattern 2:
wget --no-cookies --no-check-certificate --header "Cookie: gpw_e24=http%3A%2F%2Fwww.oracle.com%2F; oraclelicense=accept-securebackup-cookie" "http://download.oracle.com/otn-pub/java/jdk/8u77-b03/jdk-8u77-linux-x64.tar.gz"
tar xvfz jdk-8u77-linux-x64.tar.gz
jdk1.8.0_77/jre/lib/fontconfig.SuSE.10.properties.src
jdk1.8.0_77/jre/lib/fontconfig.SuSE.11.bfc
jdk1.8.0_77/jre/COPYRIGHT
jdk1.8.0_77/jre/THIRDPARTYLICENSEREADME-JAVAFX.txt
jdk1.8.0_77/jre/Welcome.html
jdk1.8.0_77/jre/README
jdk1.8.0_77/README.html
tar: jdk1.8.0_77/jre/lib/amd64/server/libjsig.so: Cannot create symlink to ‘../libjsig.so’: Invalid argument
tar: Exiting with failure status due to previous errors
root@localhost:/tmp#
root@localhost:/tmp/jdk1.8.0_77/bin#
root@localhost:/tmp/jdk1.8.0_77/bin# ./java -version
<---- In this time, Hung up no response
Pattern 3:
apt-get install default-jdk
Setting up policykit-1-gnome (0.105-1ubuntu4) ...
Setting up gconf-service (3.2.6-0ubuntu2) ...
Setting up libgnome2-bin (2.32.1-4ubuntu1) ...
Setting up libgconf2-4:amd64 (3.2.6-0ubuntu2) ...
Setting up openjdk-7-jre-headless:amd64 (7u95-2.6.4-0ubuntu0.14.04.2) ...
the java command requires a mounted proc fs (/proc).
dpkg: error processing package openjdk-7-jre-headless:amd64 (--configure):
subprocess installed post-installation script returned error exit status 1
dpkg: dependency problems prevent configuration of default-jre-headless:
default-jre-headless depends on openjdk-7-jre-headless (>= 7~u3-2.1.1); however:
Package openjdk-7-jre-headless:amd64 is not configured yet.
dpkg: error processing package default-jre-headless (--configure):
dependency problems - leaving unconfigured
dpkg: dependency problems prevent configuration of openjdk-7-jre:amd64:
openjdk-7-jre:amd64 depends on openjdk-7-jre-headless (= 7u95-2.6.4-0ubuntu0.14.04.2); however:
Package openjdk-7-jre-headless:amd64 is not configured yet.
dpkg: error processing package openjdk-7-jre:amd64 (--configure):
dependency problems - leaving unconfigured
dpkg: dependency problems prevent configuration of default-jre:
default-jre depends on default-jre-headless (= 2:1.7-51); however:
Package default-jre-headless is not configured yet.
dpkg: error processing package default-jre (--configure):
dependency problems - leaving unconfigured
Setting up gconf-service-backend (3.2.6-0ubuntu2) ...
dmesg: klogctl failed: Invalid argument
Setting up gconf2 (3.2.6-0ubuntu2) ...
/bin/df: cannot read table of mounted file systems: No such file or directory
No apport report written because the error message indicates its a followup error from a previous failure.
No apport report written because MaxReports is reached already
No apport report written because MaxReports is reached already
update-alternatives: using /usr/bin/gconftool-2 to provide /usr/bin/gconftool (gconftool) in auto mode
(gconftool-2:8236): GConf-WARNING **: Client failed to connect to the D-BUS daemon:
Unable to autolaunch a dbus-daemon without a $DISPLAY for X11
Setting up libgnomevfs2-common (1:2.24.4-1ubuntu6) ...
Setting up libgnomevfs2-0:amd64 (1:2.24.4-1ubuntu6) ...
Setting up libgnome2-common (2.32.1-4ubuntu1) ...
Setting up libgnome2-0:amd64 (2.32.1-4ubuntu1) ...
dpkg: dependency problems prevent configuration of openjdk-7-jdk:amd64:
openjdk-7-jdk:amd64 depends on openjdk-7-jre (= 7u95-2.6.4-0ubuntu0.14.04.2); however:
Package openjdk-7-jre:amd64 is not configured yet.
dpkg: error processing package openjdk-7-jdk:amd64 (--configure):
dependency problems - leaving unconfigured
dpkg: dependency problems prevent configuration of default-jdk:
default-jdk depends on default-jre (= 2:1.7-51); however:
Package default-jre is not configured yet.
default-jdk depends on openjdk-7-jdk (>= 7~u3-2.1.1); however:
Package openjdk-7-jdk:amd64 is not configured yet.
dpkg: error processing package default-jdk (--configure):
dependency problems - leaving unconfigured
dpkg: dependency problems prevent configuration of ca-certificates-java:
ca-certificates-java depends on openjdk-7-jre-headless (>= 7~u3-2.1.1~pre1-1) | java6-runtime-headless; however:
Package openjdk-7-jre-headless:amd64 is not configured yet.
Package java6-runtime-headless is not installed.
Package oracle-java8-installer which provides java6-runtime-headless is not configured yet.
Package default-jre-headless which provides java6-runtime-headless is not configured yet.
Package openjdk-7-jre-headless:amd64 which provides java6-runtime-headless is not configured yet.
dpkg: error processing package ca-certificates-java (--configure):
dependency problems - leaving unconfigured
dpkg: dependency problems prevent configuration of libatk-wrapper-java:
libatk-wrapper-java depends on default-jre | java2-runtime; however:
Package default-jre is not configured yet.
Package oracle-java8-installer which provides default-jre is not configured yet.
Package java2-runtime is not installed.
Package default-jre which provides java2-runtime is not configured yet.
Package oracle-java8-installer which provides java2-runtime is not configured yet.
Package openjdk-7-jre:amd64 which provides java2-runtime is not configured yet.
dpkg: error processing package libatk-wrapper-java (--configure):
dependency problems - leaving unconfigured
dpkg: dependency problems prevent configuration of libatk-wrapper-java-jni:amd64:
libatk-wrapper-java-jni:amd64 depends on default-jre | java2-runtime; however:
Package default-jre is not configured yet.
Package oracle-java8-installer which provides default-jre is not configured yet.
Package java2-runtime is not installed.
Package default-jre which provides java2-runtime is not configured yet.
Package oracle-java8-installer which provides java2-runtime is not configured yet.
Package openjdk-7-jre:amd64 which provides java2-runtime is not configured yet.
libatk-wrapper-java-jni:amd64 depends on libatk-wrapper-java (>= 0.30.4-4); however:
Package libatk-wrapper-java is not configured yet.
dpkg: error processing package libatk-wrapper-java-jni:amd64 (--configure):
dependency problems - leaving unconfigured
Processing triggers for libc-bin (2.19-0ubuntu6.7) ...
No apport report written because MaxReports is reached already
No apport report written because MaxReports is reached already
No apport report written because MaxReports is reached already
No apport report written because MaxReports is reached already
No apport report written because MaxReports is reached already
Errors were encountered while processing:
oracle-java8-installer
openjdk-7-jre-headless:amd64
default-jre-headless
openjdk-7-jre:amd64
default-jre
openjdk-7-jdk:amd64
default-jdk
ca-certificates-java
libatk-wrapper-java
libatk-wrapper-java-jni:amd64
E: Sub-process /usr/bin/dpkg returned an error code (1)
root@localhost:~#
The first error in "apt-get install default-jdk" output is "the java command requires a mounted proc fs (/proc)." So it can be connected with #105
The second error is due to #9 (tar's current inability to handle symlinks). I believe they've fixed that in the current code, but we won't see it until the next flight. The third issue is a duplicate of #49. Seems they've gotten Oracle's Java running on their internal builds but, again, we have to wait a while. I'm just grateful for all the attention they're pouring into this!
Support for installing Java manually is inbound. See #49.
Everything else is in bugs we're tracking.
I'm going to close this one as a dup of #49.
| gharchive/issue | 2016-04-13T10:45:25 | 2025-04-01T06:37:13.253417 | {
"authors": [
"mphaney",
"niqolay",
"russalex",
"yoshioterada"
],
"repo": "Microsoft/BashOnWindows",
"url": "https://github.com/Microsoft/BashOnWindows/issues/174",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
308524688 | BotBuilder download blob file
Bot Info
.NET
Issue Description
Hello I made a bot and want to download a file from my azure blob storage. But the bot just stops working. There is no Exception, it just hangs at that position. Here is the code:
Code Example
public async static Task<byte[]> download(IDialogContext dialog)
{
try
{
return await new WebClient().DownloadDataTaskAsync(new Uri("https://storageName.blob.core.windows.net/container/file.pdf"));
}
catch (Exception ex)
{
await dialog.PostAsync(ex);
return null;
}
}
Which channel(s) are you using? Some channel only allow for bots to send certain types of files.
| gharchive/issue | 2018-03-26T10:58:10 | 2025-04-01T06:37:13.258452 | {
"authors": [
"JasonSowers",
"mantaplatte"
],
"repo": "Microsoft/BotBuilder",
"url": "https://github.com/Microsoft/BotBuilder/issues/4371",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
303240059 | Fix #798 hide upload button
New prop showUploadButton to hide upload button. It is default to true.
Where should this property be used in frame or within the node code .
Is there anyway to enable upload button using the backhannel feature that, when a specific event was triggered, i show it to the user?
@ernestoguimaraes Yes it's possible.
I assume you are embedding Web Chat using React. If you set the props.showUploadButton to false, the upload button will hide right away. The switch is on-the-fly and does not re-render the control.
So you will listen to backchannel, when you get the signal, you set it in your React code, done. 😃
Is there anyway to disable the upload button when using the webchat iframe?
still im not clear with that..how can i remove the image upload button from the chat withdow?
still im not clear with that..how can i remove the image upload button from the chat window?
window.WebChat.renderWebChat({
...
styleOptions: { hideUploadButton: false }
}, document.getElementById('bot'));
@nmssalman
still im not clear with that..how can i remove the image upload button from the chat window?
window.WebChat.renderWebChat({
...
styleOptions: { hideUploadButton: false }
}, document.getElementById('bot'));
This did not work for me :/
@ernestoguimaraes Yes it's possible.
I assume you are embedding Web Chat using React. If you set the props.showUploadButton to false, the upload button will hide right away. The switch is on-the-fly and does not re-render the control.
So you will listen to backchannel, when you get the signal, you set it in your React code, done. 😃
@compulim Hello William,
Is there a way to do the configuration for a iframe like the following?
@CharingJC iframe Web Chat does not have a way to modify style options. Only default iframe props can be assigned (width, etc).
Greetings
Thank you so much
Get Outlook for iOShttps://aka.ms/o0ukef
From: Corina @.>
Sent: Saturday, March 13, 2021 2:49:07 AM
To: microsoft/BotFramework-WebChat @.>
Cc: CharingJC @.>; Mention @.>
Subject: Re: [microsoft/BotFramework-WebChat] Fix #798 hide upload button (#883)
@CharingJChttps://github.com/CharingJC iframe Web Chat does not have a way to modify style options. Only default iframe props can be assigned (width, etc).
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHubhttps://github.com/microsoft/BotFramework-WebChat/pull/883#issuecomment-797677932, or unsubscribehttps://github.com/notifications/unsubscribe-auth/AR3M2NTEUOJW5WJ4MM4YVULTDJHZHANCNFSM4EUEJHWA.
| gharchive/pull-request | 2018-03-07T20:02:29 | 2025-04-01T06:37:13.268618 | {
"authors": [
"AndrewTilson",
"CharingJC",
"compulim",
"corinagum",
"deepanigi",
"ernestoguimaraes",
"jvanderbiest",
"lumayara",
"nmssalman"
],
"repo": "Microsoft/BotFramework-WebChat",
"url": "https://github.com/Microsoft/BotFramework-WebChat/pull/883",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
182157785 | Support SQL Local DB 2016
In my SQLLocalDB-2016-only setup (this is what you get by installing VS2015 Update 3 on a fresh machine), CC fails to connect to default cache. I tracked down to this; I am not sure if you would consider the current approach tenable. I would rather try to connect to (LocalDb)\MSSQLLocalDB to see what happens. But I think this is how CC worked pre- b4b89e6a and there was a reason it was changed to be like it is now.
I kinda fixed it by creating an empty "12.0" subkey there, but this is just a hotwired workaround, mostly to confirm that I did find the correct spot.
/cc @tom-englert
Thanks! Looks good.
| gharchive/pull-request | 2016-10-11T02:19:37 | 2025-04-01T06:37:13.270457 | {
"authors": [
"SergeyTeplyakov",
"kkm000"
],
"repo": "Microsoft/CodeContracts",
"url": "https://github.com/Microsoft/CodeContracts/pull/464",
"license": "mit",
"license_type": "permissive",
"license_source": "bigquery"
} |
274214134 | Navigate to an App?
Is there a way to navigate to a specific app that was built with the new app model?
At this point there is no API that will open the new apps. You would need to use the generic Document API's or the base driver for this. I've added this as a work item to add these API's but just no timetable at this moment.
Thanks for the feedback
| gharchive/issue | 2017-11-15T16:13:53 | 2025-04-01T06:37:13.273170 | {
"authors": [
"devkeydet",
"dtu11"
],
"repo": "Microsoft/EasyRepro",
"url": "https://github.com/Microsoft/EasyRepro/issues/45",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
297633759 | SetValue results in blank field even with the bugfix
When setting a value in text field in CRM V9 the value seems to blank out again.
Using V9 branch at line 147 I can see there is a note about the bugfix, but looks to be still occurring in our case.
https://github.com/Microsoft/EasyRepro/blob/e34b5cae67393f7cc789313fd60536ae326f4fdb/Microsoft.Dynamics365.UIAutomation.Api/XrmPage.cs#L147
I'll take a look at it.
@KiavashS - I attempted to reproduce this on the latest versions using both Chrome and Internet Explorer and both Single Line of Text and Multiple Lines of Text field types, but have been unsuccessful. The supplied input values are saved successfully. My Dynamics organization is running version v9.0.1.459
Please try again and if you can still reproduce this then reopen the issue and supply the following:
Dynamics org version
Browser type/version
Code used during your test
Thanks,
Tyler
| gharchive/issue | 2018-02-15T23:27:04 | 2025-04-01T06:37:13.276641 | {
"authors": [
"KiavashS",
"TYLEROL",
"dtu11"
],
"repo": "Microsoft/EasyRepro",
"url": "https://github.com/Microsoft/EasyRepro/issues/82",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
400801380 | GraphEngine importing GML
I have a project where I need to import a large road network that is expressed/exported as GML is/are there: -
any existing or plans to create tools that can import terabytes of GML data.
any plans to provide GraphEngine extensions to perform geospatial/routing queries on graph data
@grahamehorner
the importer prototype in the repo is already capable of chewing terabytes of RDF data. So GML support could also be built upon that (relatively easily). contributions are welcomed :)
very interesting idea, but outside the scope of this repo (base GE system).
| gharchive/issue | 2019-01-18T16:38:59 | 2025-04-01T06:37:13.278671 | {
"authors": [
"grahamehorner",
"yatli"
],
"repo": "Microsoft/GraphEngine",
"url": "https://github.com/Microsoft/GraphEngine/issues/276",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
435617551 | what's the Sharding strategy of GraphEngine? Can I use other db backends ?
what's the Sharding strategy of GraphEngine? Can I use other db backends ?
@BenDerPan the default sharding function is simple: (long id) => (int)(id / 8 % Global.CloudStorage.PartitionCount
You can override this with Global.CloudStorage.SetPartitionMethod(.) and align with the other backend.
@yatli thanks, and how many replicas of each shard ? can we config it in trinity.xml ? :)
use "AvailabilityGroup" to specify the availability group id.
see:
https://github.com/Microsoft/GraphEngine/wiki/config-v2#server
thanks, great! close this now.
| gharchive/issue | 2019-04-22T06:35:13 | 2025-04-01T06:37:13.281052 | {
"authors": [
"BenDerPan",
"yatli"
],
"repo": "Microsoft/GraphEngine",
"url": "https://github.com/Microsoft/GraphEngine/issues/288",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
196091853 | Dart - very poor accuracy
When I use dart as a booster I always get very poor performance in term of l2 result for regression task.
Even If I use small drop_rate = 0.01 or big like 0.3.
When I use dart in xgboost on same dataset, with similar setting (same learning rate, similiar num_trees) dart alwasy give me boost for accuracy (small but always).
But here accuracy is poor badly, like there is a bug, not just dart is not suitable for my task.
Can anyone confirm that dart is working for regression task in term of better accuracy?
My setting is as follows (part of the Python code for ramdom search of params):
lr = np.random.choice([0.01, 0.005, 0.0025])
list_count = np.random.choice([250, 500, 750, 1000])
min_in_leaf = np.random.choice([25, 50, 100])
subF = np.random.choice([0.15, 0.22, 0.3, 0.5, 0.66, 0.75])
subR = np.random.choice([0.66, 0.75, 0.83, 0.9])
max_depth = np.random.choice([9, 11, 15, 25, 45, 100, -1])
dart_rate = np.random.choice([0, 0, 0, 0.01, 0.03, 0.1])
max_bin = np.random.choice([63, 127, 255, 511])
lambda_l1 = np.random.choice([0, 1., 10., 100.])
lambda_l2 = np.random.choice([0, 1., 10., 100.])
iterace = 10000
if only_testing:
min_in_leaf = 25
iterace = 10
boost_type = 'gbdt'
if dart_rate > 0:
boost_type = 'dart'
params = {
'task' : 'train',
'boosting_type' : boost_type,
'objective' : 'regression',
'metric' : 'l2',
'max_depth' : int(max_depth),
'num_leaves' : int(list_count),
'min_data_in_leaf' : int(min_in_leaf),
'learning_rate' : lr,
'feature_fraction' : subF,
'bagging_fraction' : subR,
'bagging_freq': 1,
'verbose' : 0,
'nthread' : nthread,
'drop_rate': dart_rate,
'max_bin': max_bin,
'lambda_l1' : lambda_l1,
'lambda_l2' : lambda_l2
}
model = lg.train(
params,
(matrix_learn, target_learn),
num_boost_round = iterace,
valid_datas = (matrix_test, target_test),
early_stopping_rounds = 50
)
What is working in XGB is probably not just drop_rate but also skip_drop. http://xgboost.readthedocs.io/en/latest/tutorials/dart.html
Can you please support this as well?
@wxchan
https://github.com/dmlc/xgboost/blob/master/src/gbm/gbtree.cc#L547-L711
can you check the code of implementation in XGBoost, and figure out why?
@guolinke I read these codes before. It's not hard to be added, but I have no idea where he came upon those parameters. Btw, is this line a bug? weight_drop[idx_drop[i]] *= factor;?.
@gugatr0n1c
Can you provide how you set up skip_rate etc. for xgboost?
Do you know where xgboost gets the ideas of skip_rate, sample_type and normalize_type?
Can you show accuracy comparisons between xgboost and lightgbm (both gbdt & dart) for your task?
@wxchan, I am not familiar with XGBoost.
Following is my guess:
weight_drop is the weights of all trees, not just dropped trees, right? If yes, I think the line you refer to is a bug. it should be weight_drop[idx_drop[i]] *= factor . If this is a bug, I am not sure why XGBoost can preform better....
It seems drop tree logic is based on sum_weight in XGBoost, but LightGBM is just random. Maybe we can adapt this.
@guolinke I can add that option. I will do some investigations on this.
@wxchan
ad1] usually I run random search on parameters, and for xgboost on my datasets best setting is:
skip_rate = 0.3
drop_rate = 0.2
I use default for next two dart parameters.
ad2] actually dart for xgboost was done with my issue here
https://github.com/dmlc/xgboost/issues/809
And it seems to that marugari add skip_rate just like his idea.
On other two setting I do not know the source.
ad3]
my task is regression problem with 500k rows, 400 features, and RMSE on cross validation:
randomforest from scikit: 5.46
deepnet (5 hidden layers, prelu, adam, xavier) from mxnet 5.43
xgboost without dart: 5.421
xgboost with dart: 5.419
lightgbm without dart: 5.418
lightgbm with dart: 5.5 - not a chance to beat randomforest
For all methods I did some random search of parameters and method should be comparable in the sence of RMSE. Speed is best for deepnet - but it is different algorithm (also depends on settings and hardware).
@gugatr0n1c
Can you test two things on your dataset:
For lightgbm dart, set drop_rate to a very small number, such as drop_rate=1/num_iter; because your num_iter is big, each trees may be dropped too many times;
For xgboost dart, set learning rate=1. For dart, learning rate is a different concept from gbdt. In original paper, it's fixed to 1.
It's not so difficult to add skip_rate.
If there are users, I will PR.
@wxchan
ad1] I tried with 0.005, 0.0005, 0.0001 and got similar "bad" results... strange is that early stopping stop training very early (hundreds of iterations with compare two several thousands without dart)
ad2] I can do this tomorrow from work
@marugari where do you get the idea about skip_rate and normalize_type?
another question: when sample_type = weighted in xgboost, dparam.rate_drop * weight_drop.size() * weight_drop[i] / sum_weight (from here) can be bigger than 1, the probability of 'this' tree being selected is calculated as some number bigger than 1 but actually 1. Will it cause the expectation of dropped trees number smaller than rate_drop?
@wxchan Regarding skip_rate, All we have to do is leaving idx_drop empty.
In order to use learning_rate < 1 in dart, I propose the following:
/* dart.hpp L:103 */
if (normalize_type == 1) {
shrinkage_rate_ = learning_rate_ / (1.0 + learning_rate_);
} else {
shrinkage_rate_ = learning_rate_ / (drop_index_.size() + learning_rate_);
}
/* dart.hpp L:119 */
if (normalize_type == 1) {
models_[curr_tree]->Shrinkage(-k / learning_rate_);
} else {
models_[curr_tree]->Shrinkage(-1 / learning_rate_);
}
sample_type is very rough, since I think it's not involved in convergence.
@marugari sorry, what I meant is, is there some paper regarding skip_rate and normalize_type, because they are not shown in dart original paper?
@wxchan sorry. They are my extensions, not published.
@marugari I see, thanks.
If you want to use skip_rate, you can arrange a callback with changing drop_rate.
Just my opinion on this: I think the parameters I add 'here' should be tested, I am not saying they are wrong, they can be wonderful ideas, but I want to be convinced before I add them 'here'. And I totally agree if other user would like to contribute to it.
@wxchan If I set in xgboost learning_rate to 1, accuracy is very bad as well, even worse than here...
@gugatr0n1c in original paper, they only use shrinkage_rate=1/(1+num_drop_trees); as marugari said, he set shrinkage_rate_ = learning_rate_ / (drop_index_.size() + learning_rate_); I think this is the main difference rather than skip_rate or sample_type.
As I wrote I did not used 'sample_type' parameter. I only used 'drop_rate' and 'skip_rate' (together with learning_rate << 1) in xgboost. And it was working nicely. I can not say now the influence of 'skip_rate'. But when I used random search for hyperparameters tuning, it always wanted to set 'skip_rate' to something non zero.
But, yes I agree that the main trick is to adapt shrinkage_rate as written.
But to be honest, if I can use below to simulate 'skip_rate', than it is not necessary to add this parameter, just maybe mention it as an example about callback? Actually it is even better solution, because I can then randomize drop_rate together with skip_rate (this is not even in xgboost) during learning process by some logic in generating drop_rate_list.
-->simulate 0.33 skip_rate
drop_rate_list = np.random.choice([0., 0.3, 0.3], iteration_count).tolist()
lg.train(
params,
data_train,
num_boost_round = iteration_count,
valid_sets = data_valid,
early_stopping_rounds = 50,
callbacks = [lg.reset_parameter(drop_rate = drop_rate_list)]
)
So I believe there are nice things to do: 1] allow change drop_rate in callback, should work now right?, 2] change shrinkage_rate as marugari wrote.
@wxchan Although weight_drop of new trees can be large, leaf values of that are scaled by learning_rate.
I think weight_drop * learning_rate = 1.0.
@marugari oh sorry, I didn't see the if (num_drop == 0) branch. My fault. I also didn't know xgboost calculates learning_rate in another place, I thought it's calculated only in CommitModel.
Then I think it's reasonable to change shrinkage_rate to see if it works. It is actually a combination of shrinkage_rate in gbdt and normalization weight in dart.
@gugatr0n1c can you test again with latest code?
@guolinke I tried, but got error: Segmentation fault, I tried twice recompile from github
boostint_type = 'gbdt' is OK
boostint_type = 'dart' return Segmentation fault
@wxchan can you take a look about this?
@guolinke it seems caused by c_api because cmd line version is fine. I take a look at segfault log: it happened in Dart::Init, seems gbdt_config_ is still null after GBDT::ResetTrainingData.
@gugatr0n1c @wxchan fixed, can you try it again?
@guolinke training seems to be working, give you know about accuracy
@guolinke seems to be not working properly. It almost not converge (even worse than before). Seems to me that first tree is build with very large learning_rate, than additional trees has huge problem to converge.
Tried with dart_rate: 0.3, 0.05, 0.001 and with learning_rate = 0.004 and 0.1 - almost same result on all settings.
@gugatr0n1c OK, I see.
@wxchan , I also try this, the training error is small. seems the learning rate is still wrong.
But all I change is to make learning rate a variable rather than fixed to 1. So if you set learning rate=1, it's same as before. The first tree is built with shrinkage_rate=1, it's also same as before.
I actually test this setting when I implemented dart and found it not work so well. So I didn't add it before.
Can you try to set learning rate to some number near 1 like 0.8 or 0.9? Because shrinkage_rate will divide num_drop in dart, I don't think it should set very small.
@wxchan ,
I think the weight of dropped trees is not correct now.
when learning_rate = 1.0:
weight of new tree: 1/(k+1)
weight of drop tree: 1/(k'+1) * k/(k+1).
sum of these weight is approx to 1.
current implementation:
weight of new tree: lr/(k+lr)
weight of drop tree: lr/(k+lr)* k/(k+lr)
sum of these weights is still approx to 1.
However, I think if we add learning rate to DART, may be sum of these weights should approx to learning_rate.
So, maybe:
weight of new tree: lr/(k+1)
weight of drop tree: lr/(k+1)* k/(k+1)
is better?
@guolinke that is exactly what I have tried before and I think it's better, that's why I said it's a combination of gbdt shrinkage rate (lr term) and dart dropped weight (1/(k+1) term).
Current implementation is according to xgboost. All we know here is xgboost dart works better on some datasets, I think we still need better theory.
By the way, I just fixed that bug of xgboost dart I mentioned it earlier, check out here.
I read the original paper and there is several thing to point out:
1] setting, where DART is beating MART is small number of trees (250), this is not typical in practice (maybe for randomforest, not for boosting on bigger datasets) - table 5., not sure if we can replicate results for 5k trees with small learning_rate (little bit questioning result from paper for bigger datasets).
2] 'core' idea is to built equivalently strong trees, this is interesting when lightgbm is building leaf_wise not level_wise - this can be different from xgboost - as in paper is shown different tree architecture figure 2.
3] Critical part for boosting is very small learning_rate, I do not believe this can substituted just by drop_rate for datasets where we need thousands of trees. Setting learning_rate to 1 kills convergence almost always after max hundreds of iterations.
4] What I believe is useful from the paper is idea of dropping trees. Because it works as next regularizer and this is what marugari did in xgboost implementation. But sure, there is no paper or something to backup this.
So bottom line: we need set learning_rate small and it is nice to have chance to drop trees as regularizer... but sure, it is little bit different against the paper. Up to you guys what to do.
@wxchan
I think xgboost using learning rate at [here](weight of new tree: lr/(k+1)
weight of drop tree: lr/(k+1)* k/(k+1)), not in GBTree.
So the learning rate of DRAT is not 1.0.
@gugatr0n1c can you try the xgboost with latest code? @wxchan just fix bug in xgboost.
@guolinke but weight of the first tree is 1/(num_drop+lr) = 1/lr, it canceled out the lr.
@wxchan , I see.
Maybe we can do some experiments for these different settings.
I think we can choose our own algorithms, not limit to the implementation in xgboost.
I think:
weight of new tree: lr/(k+1)
weight of drop tree: lr/(k+1)* k/(k+1)
will be better. In this case, tree normalization is same as before, just new tree shrinks by lr.
@wxchan I think this may be help
Dropped trees and new trees has similar leaf values.
It seems that shrinking just new trees is improper.
@marugari
In you implementation. if not trees are dropped, add new_tree with weight lr. Total delta weight is lr.
if trees are dropped, add new_tree with lr/(k+lr), and shrink k trees by k/(k+lr). Total delta weight is lr/(k+lr) - k * lr/(k+lr)* w_j . w_j is the previous weight of drop tree j.
w_j can be lr * (k/(k+lr))^i (if skipped) or lr/(k+lr)*(k/(k+lr))^i(not skip), i is drop count of this tree.
I think lr/(k+lr) - k * lr/(k+lr)* w_j is far smaller than lr, since lr/(k+lr) < lr and it minus a positive number. And when k==0(means not drop), it equal to 1, this is not consist with your operation when not trees are dropped.
BTW,
In original paper, delta weight is:
1/(k+1) - k/((k+1)*(k+1)) = 1/((k+1)(k+1))
Do you think let delta wight approx to lr/((k+1)*(k+1)) is better?
This is a little inaccurate. I'm Sorry...
New trees and sum of dropped trees scaled by previous weights has similar leaf values.
Thus, the normalization factor is independent of previous weights.
If weights of dropped trees are small. new trees have also small leaf values.
I don't think the delta weight is suitable characteristic.
@marugari I think i understand your idea now.
You let weight of new tree approx to the sum of dropped trees, and use skip_rate to add new tree and increase total weight?
It seems when skip_rate==0.5 has best accuracy.
My idea is let total weight to be increased at every iteration. I think this may save half of iterations while achieve same accuracy.
@guolinke You are right.
In the case of small learning_rate, skipping iterations are main sources of loss reduction.
I just push a commit at dart branch: https://github.com/Microsoft/LightGBM/commit/518bafd53570ebdf0db84aceefa989dd24a32394
It adds skip_drop and xgboost_dart_mode.
I also try some experiments, and it seems xgboost_dart_mode works well in small dataset.
@gugatr0n1c can you try this with your data?
love to test this, but have some xmas travelling now, so it will take some time..
@gugatr0n1c take you time and merry xmas!
@wxchan Do you have any comments?
It's better now. I think you can merge it to master and update docs for it.
@wxchan , I am not sure should we keep xgboost_dart_mode? And should give it a better name?
@guolinke merry xmas to you as well
I done some tests:
for xDart (xgboost_dart_mode = True) I got great accuracy - this is great for my datasets, beating 'gbdt'
for Dart (xgboost_dart_mode = False) still bad accuracy... (but done limited tests only with very small learning_rate)
But here both xDart and Dart have training very slow: 15mins (gdbt) vs. several hours for xDart, some iterations are very fast ~ 100ms, but some are about 15-20s (related to skip_drop?), DART in xgboost is not so slower to 'gbdt', is there space to speed it up?
@gugatr0n1c Really?
Yes, XGBoost-DART is very slow (espesially late in a training) because it doesn't use the buffer.
However, I seem LightGBM-(x)DART runs as fast as gbdt.
@gugatr0n1c what values of your drop_rate and skip_drop? this speed seems abnormal.
hi, erveryone, I have rencently trianed dart and gbdt model using same training data with some noise. For predictions, in general, the dart model is little better than gbdt model. I set all the same parameter learning_rate=0.0112. I didn't try xgboost so I can't tell something.
@gugatr0n1c I just don't know how to use mxnet to implement "deepnet (5 hidden layers, prelu, adam, xavier) ", can you provide some examples to me? Thank you very much!
@guolinke my setting is this:
learning_rate = 0.006
up to 10k iterations with early_stopping = 50 (stopped around 5k-7k iters with 'gdbt')
drop_rate = skip_drop = 0.3
using xDART
First like 100 iterations are fast, but I can see still that some iterations are very fast some slower
But after 1k iterations it is like 100ms vs 20s
@marugari not detaily tested in xgboost, but never have this feeling that later in the training is difference so big..
@anddelu this is very big offtopic here... send me your email to my twitter PM... @gugatr0n1c
@gugatr0n1c
I know what happen now. I will try to optimize the speed.
The reason is it will drop much tree after trained many iterations, this will increase the computation cost of score.
A quick solution is reset drop_rate smaller during training.
@wxchan if xgboost_dart_mode is better, I think we can just keep it and remove our version.
What do you think?
@guolinke It may depend on different task. Our version is based on original paper. I think it's better keep it for now.
@gugatr0n1c I add an parameter max_drop to limit the number of dropped trees. This can help for the speed. But I am not sure how it will affect to the accuracy.
And if max_drop <= 0, it means not limit.
I will close this issue for now, feel free to open an new issue if it cannot speed up or have much loss on accuracy.
@guolinke thx, I will run some tests after New Year
| gharchive/issue | 2016-12-16T16:00:43 | 2025-04-01T06:37:13.335497 | {
"authors": [
"anddelu",
"gugatr0n1c",
"guolinke",
"marugari",
"wxchan"
],
"repo": "Microsoft/LightGBM",
"url": "https://github.com/Microsoft/LightGBM/issues/126",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
339235727 | Pdb not found / mismatch so I can't debug
.NET Core 2.1.301 (Both windows and linux)
VS 15.7.4
Yes, I use <DebugType>portable</DebugType>
Command used:
Debug.MIDebugLaunch /Executable:"C:/Users/pc/Desktop/shared_linux_vm_folder/ConsoleApp1/Linux_MyApp/bin/Debug/netcoreapp2.1/Linux_MyApp.dll" /OptionsFile:"C:/Users/pc/Desktop/shared_linux_vm_folder/ConsoleApp1/Linux_MyApp/options.xml"
Output Window:
Using arguments
Version : 'vs2015u2'
Location : '/home/pc/.vs-debugger/vs2015u2'
SkipDownloads : 'true'
LaunchClrDbgAfter : 'true'
RemoveExistingOnUpgrade : 'false'
Info: Using clrdbg version '15.0.26022.0'
Info: Last installed version of clrdbg is '15.0.26022.0'
Info: ClrDbg is upto date
Info: Skipping downloads
Info: Launching clrdbg
Loaded '/home/pc/workarea/software/dotnet/shared/Microsoft.NETCore.App/2.1.1/System.Private.CoreLib.dll'. Cannot find or open the symbol file.
Loaded '/mnt/hgfs/shared_linux_vm_folder/ConsoleApp1/Linux_MyApp/bin/Debug/netcoreapp2.1/Linux_MyApp.dll'. Cannot find or open the symbol file.
Loaded '/home/pc/workarea/software/dotnet/shared/Microsoft.NETCore.App/2.1.1/System.Runtime.dll'. Cannot find or open the symbol file.
Loaded '/home/pc/workarea/software/dotnet/shared/Microsoft.NETCore.App/2.1.1/System.Runtime.Extensions.dll'. Cannot find or open the symbol file.
Loaded '/home/pc/workarea/software/dotnet/shared/Microsoft.NETCore.App/2.1.1/System.Console.dll'. Cannot find or open the symbol file.
Loaded '/home/pc/workarea/software/dotnet/shared/Microsoft.NETCore.App/2.1.1/System.Threading.dll'. Cannot find or open the symbol file.
Hello World!
The program '' has exited with code 0 (0x0).
As you can see, it says: "Linux_MyApp.dll'. Cannot find or open the symbol file."
To diagnose this I dumpbin'ed the .dll headers out and it shows a windows path to the pdb, but even when I compile on linux (which then uses a windows path) it still wont find the pdb so it must be a mismatch of the pdb internal guid(?) or something, I'm not sure how to diagnose this any further, much less fix it.
I also tried enabling logging with Debug.MIDebugLog /On /OutputWindow but it shows no further error messages or relevant info.
Weird: I can attach and the symbols will be loaded if I start the app on linux first manually and then attach to it via visual studio. Of course this is too cumbersome to constantly do manually.
It does say though "Breakpoint warning: A copy of C:\Users\pc\Desktop\shared_linux_vm_folder\ConsoleApp1\Linux_MyApp\Program.cs was found in Linux_MyApp.dll, but the current source code is different from the version built into Linux_MyApp.dll." which shouldnt be the case either.
@Meai clrdbg is no longer being maintained, and will not work correctly with .NET Core 2.1 applications. You want to switch over to using vsdbg. Instructions: https://github.com/Microsoft/MIEngine/wiki/Offroad-Debugging-of-.NET-Core-on-Linux---OSX-from-Visual-Studio
| gharchive/issue | 2018-07-08T15:44:03 | 2025-04-01T06:37:13.350373 | {
"authors": [
"Meai",
"gregg-miskelly"
],
"repo": "Microsoft/MIEngine",
"url": "https://github.com/Microsoft/MIEngine/issues/743",
"license": "mit",
"license_type": "permissive",
"license_source": "bigquery"
} |
81033617 | Color Theme for PTVS
Hello,
Is there any good color theme for PTVS? I would prefer a dark theme, but many are not ready for Python syntax. If you like any color theme for working with PTVS please let me know.
Thank you,
Marcio
The built-in Dark theme in Visual Studio is probably the only one that properly handles Python syntax, since that's the one that we can update when we install PTVS.
If you look under Tools, Options, Fonts and Colors you'll find a number of "Python ..." items that can be modified. Other themes will need to set those as well to handle Python syntax, but you can always modify them yourself.
I'm sure the other theme developers would appreciate suggestions on what colours they could use there, but I don't know how best to get in contact with them. Chances are they don't even know about PTVS :)
| gharchive/issue | 2015-05-26T16:19:04 | 2025-04-01T06:37:13.356034 | {
"authors": [
"arqeco",
"zooba"
],
"repo": "Microsoft/PTVS",
"url": "https://github.com/Microsoft/PTVS/issues/252",
"license": "Apache-2.0",
"license_type": "permissive",
"license_source": "github-api"
} |
357668883 | Error when trying to connect to local workspace
I'm having trouble to connect to my local pbix files and the application reports an error message saying:
Analysis Server localhost:60343 is not running in Tabular mode
Both Source and Target files are open and I have verified that I actually can connect to the local workspaces using SSMS
Hi, it currently doesn't work with the local Power BI Desktop AS instance. When we open up the XMLA endpoint on Power BI (soon), it will work with the Power BI service in the first version
Ah!
Thanks for the info Christian!
No problem!
| gharchive/issue | 2018-09-06T14:00:10 | 2025-04-01T06:37:13.358742 | {
"authors": [
"MagAlp",
"christianwade"
],
"repo": "Microsoft/Power-BI-ALM-Toolkit",
"url": "https://github.com/Microsoft/Power-BI-ALM-Toolkit/issues/9",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
113753923 | Change tsconfig.json in TS template to match http://json.schemastore.…
tsconfig.json content should be {compilerOptions...} instead of { tsCompilerOptions.. } as specified in http://json.schemastore.org/tsconfig
@guillaumejenkins can you review this?
This is correct, good catch
:+1:
I was just looking over this PR to see if we should merge it, but it looks like there are more references to tsCompileOptions in the code which might need to be modified as well. @guillaumejenkins can you look into this and verify that after we take the change everything still works? The other references might just be for our own compilation/gulp scripts.
What Jimmy means is that in src\build_config.json, src\gulpmain.ts and src\gulpfile.js, we have an object called "tsCompilerOptions". That object is an option bag that we give to gulp-typescript. Gulp-typescript only cares about the contents of the object, not its name, so nothing is broken by rido's change.
However, for the sake of consistency, we might as well rename our options bag object. I'll do this in a separate change to make sure we do the proper validation.
| gharchive/pull-request | 2015-10-28T06:37:48 | 2025-04-01T06:37:13.365012 | {
"authors": [
"MSLaguana",
"guillaumejenkins",
"nisheetjain",
"rido-min"
],
"repo": "Microsoft/TACO",
"url": "https://github.com/Microsoft/TACO/pull/121",
"license": "mit",
"license_type": "permissive",
"license_source": "bigquery"
} |
72019304 | Enums
regular enums, Const enums, different in lining rules
https://github.com/Microsoft/TypeScript-Handbook/blob/master/pages/Basic Types.md#enum
#71
| gharchive/issue | 2015-04-29T23:47:05 | 2025-04-01T06:37:13.377034 | {
"authors": [
"DanielRosenwasser",
"mhegazy"
],
"repo": "Microsoft/TypeScript-Handbook",
"url": "https://github.com/Microsoft/TypeScript-Handbook/issues/38",
"license": "apache-2.0",
"license_type": "permissive",
"license_source": "bigquery"
} |
171170292 | Typescript crashes when a file starts with a comment and import statements are in it
TypeScript Version: 1.8.10
Code
/*
import { SomeImport } from 'example';
^
this import statement makes typescript crash,
but actually it's a comment..
*/
import { platformBrowserDynamic } from '@angular/platform-browser-dynamic';
import { AppModule } from './app.module';
platformBrowserDynamic().bootstrapModule(AppModule);
Problem
Comments with import statements at the beginning of the file let typescript crash. error is:
ERROR-Message:
zone.js:101 GET http://localhost:3000/traceur 404 (Not Found)
minimal-project-with-bug-typescript.zip
minimal project with this bug - source code
http://plnkr.co/edit/fdGzQPHSBy12rYdPwNNP?p=preview
Expected behavior:
a comment should not affect the behaviour
Actual behavior:
the code in the comment affects the behaviour
That error doesn't look like it comes from TypeScript at compile time. Isn't that a runtime error - a HTTP 404?
Hi @yortus,
I'm not sure if it's caused by typescript or any of the node_modules of the angular2 application, because there are absolutely no useful error logs.
I think the problem is related to SystemJS. When you change the SystemJS transpiler from Traceur to TypeScript you get an similar error message (http://localhost:3000/ts 404 instead of http://localhost:3000/traceur 404). I don't think that the problem is related to TypeScript because the code is compiling fine.
None of these errors come from TypeScript. Stack Overflow may be useful to help you identify a root cause; we cannot due external product support in the TS bug tracker.
| gharchive/issue | 2016-08-15T13:29:35 | 2025-04-01T06:37:13.383028 | {
"authors": [
"RyanCavanaugh",
"phip1611",
"vkennke",
"yortus"
],
"repo": "Microsoft/TypeScript",
"url": "https://github.com/Microsoft/TypeScript/issues/10338",
"license": "Apache-2.0",
"license_type": "permissive",
"license_source": "github-api"
} |
172270882 | Bug: Recursive generic union type synonyms cause poor compiler performance
TypeScript Version: 1.8.10 and 2.1.0-dev.20160818
Code
type T<A> = { a: TSynonym<A> }
| { b: TSynonym<A> }
| { c: TSynonym<A> }
//| { d: TSynonym<A> }
type TSynonym<A> = T<A>
function f<A>(): T<A> {
return g<A>()
}
function g<A>(): T<A> {
// return h<A>()
}
function h<A>(): T<A> {
}
Expected behavior:
Type error, in a reasonable length of time (because g and h do not return a value).
Actual behavior:
This example takes over 7 seconds to (fail to) compile on my machine (a new, maxed-out MacBook Pro) and peaks at over 400MB of memory. Uncommenting the fourth line of T's declaration causes the compiler to crash with an out-of-memory error after about a minute. Uncommenting the call to h doubles the compilation time to 14 seconds and increases peak memory usage to over 1GB.
Removing TSynonym (so that the body of T uses T directly), or removing all the generic types, improves the performance significantly.
My guess is that the type-checker is recursively expanding the definitions of T and TSynonym when it tries to type-check the functions, resulting in a very large quantity of nested union types.
Here is the output of tsc with the --diagnostics flag:
Files: 2
Lines: 19030
Nodes: 98142
Identifiers: 35533
Symbols: 354868
Types: 755366
Memory used: 611459K
I/O read: 0.00s
I/O write: 0.00s
Parse time: 0.14s
Bind time: 0.10s
Check time: 6.73s
Emit time: 0.00s
Total time: 6.97s
Hi,
this issue can still be reproduced with current master (9afb07de5a0bcfba871ec8745d795b4f6359de7c). On my machine, the example takes 8 seconds to compile, and if the fourth line is commented out, node crashes with "JavaScript heap out of memory".
looks like a duplicate of the same underlying issue in https://github.com/Microsoft/TypeScript/issues/10759.
Anonymous types are not cached when doing generic instantiations, causing a lot of these to happen than there should be. the work around is to use named types. e.g.:
interface A<T> {
a: TSynonym<T>
}
interface B<T> {
b: TSynonym<T>
}
interface C<T> {
c: TSynonym<T>
}
interface D<T> {
d: TSynonym<T>
}
type T<V> = A<V> | B<V> | C<V> | D<V>;
type TSynonym<A> = T<A>
function f<A>(): T<A> {
return g<A>()
}
function g<A>(): T<A> {
return h<A>()
}
function h<A>(): T<A> {
}
This has been fixed by the type alias instantiation caching logic that was added in #12114. The OP example now type checks in effectively zero time.
| gharchive/issue | 2016-08-20T13:23:03 | 2025-04-01T06:37:13.389797 | {
"authors": [
"ahejlsberg",
"benjamin-hodgson",
"kpreisser",
"mhegazy"
],
"repo": "Microsoft/TypeScript",
"url": "https://github.com/Microsoft/TypeScript/issues/10451",
"license": "Apache-2.0",
"license_type": "permissive",
"license_source": "github-api"
} |
175729284 | VisualStudio 2015 Update 3 - Comment/uncomment selection does not work for .ts files anymore.
TypeScript Version: 2.0.2 Beta
VisualStudio 2015 Update 3
Expected behavior:
The selected text/line we're on should be commented/uncommented
Works fine in the other file types I've tried it on (html, js, cs, scss)
Actual behavior:
Selected text/line is not commented/uncommented.
Visual Studio says:
"The key combination (Ctrl + K, Ctrl + C) is bound to command (Comment Selection) which is not currently available."
Seems working fine locally. do you see this on other machines? are other features impacted? do you have any extensions?
Thanks for the reply. Might be just my setup then :/
I've tried to disable all extensions that I can disable, and suspended ReSharper without any luck.
I haven't check any other machines, but I'll set up a clean VM and check it out.
Works as expected in the new clean VM.
Sorry for the false alarm.
This can be fixed without reinstalling on clean VM. Just use the solution provided here
https://github.com/Microsoft/TypeScript/issues/7519
Try deleting: %localappdata%\Microsoft\VisualStudio\14.0\ComponentModelCache,
And run devenv /setup from the Developer Command Prompt for VS2015 (as an admin).
I experienced the same CommentSelection command disappear after installation of TypeScript for Visual Studio 2015 tools on 20,11.2016 and fixed using method above.
https://www.microsoft.com/ru-ru/download/confirmation.aspx?id=48593
| gharchive/issue | 2016-09-08T11:25:36 | 2025-04-01T06:37:13.395373 | {
"authors": [
"RodionKulin",
"mhegazy",
"ruant"
],
"repo": "Microsoft/TypeScript",
"url": "https://github.com/Microsoft/TypeScript/issues/10775",
"license": "Apache-2.0",
"license_type": "permissive",
"license_source": "github-api"
} |
205785606 | instanceof with class-like variable RHS erroneously narrows "else"
TypeScript Version: 2.2-rc
Code
class Prop<T> {
arg: T;
}
interface PropClass<T> {
new(): Prop<T>;
prototype: Prop<T>;
}
function check<T>(p: Prop<{}>, specialSubclass: PropClass<T>) {
if (p instanceof specialSubclass) {
p.arg; // ok, arg has type T
} else {
p.arg; // error here: Property 'arg' does not exist on type 'never'.
}
}
Expected behavior:
p should have type Prop<{}> in the else block.
Actual behavior:
p has type never in the else block.
Because specialSubclass is a class-like object with instance type Prop, TypeScript is assuming that instanceof returning false implies the object is not a Prop. This is a bad assumption, because in this kind of code instanceof only returns true for members of some (unknown) subclass.
instanceof should narrow the false case only if the RHS is an actual class reference. If it's a variable of a class-like type, it should only narrow the true case.
This works in nightly.
This works in nightly.
I do not think so.
The Type system in TS is structural. so the checking for instanceof is rather an approximation to what happens at run time (which is nominal).
you could however, try and trick the compiler by using generic type parameters, these will not be narrowed to never since they are not assignable to the instance of your other argument. e.g.: function check<T, U extends Prop<{}>>(p: U, specialSubclass: PropClass<T>).
The Type system in TS is structural. so the checking for instanceof is rather an approximation to what happens at run time (which is nominal).
Right, I understand that due to structural typing {arg:string} is considered the same type as Prop<string>, and so TS acts as if instanceof Prop returns true on it. That's a tad unfortunate but it's actually not the issue here.
My issue is that a value that is (both structurally and nominally) of a class, is getting narrowed away due to instanceof on a subclass returning false. Even if class types were treated 100% nominally this would still be a problem.
If the RHS of instanceof is literally the class A, then it's correct to narrow A away when it returns false (other than the structural/nominal issue). But if it's a variable of type new() => A, it's generally some subclass of A, not A itself - so one can't reasonably assume anything about the object's type from instanceof returning false.
you could however, try and trick the compiler by using generic type parameters, these will not be narrowed to never since they are not assignable to the instance of your other argument. e.g.: function check<T, U extends Prop<{}>>(p: U, specialSubclass: PropClass<T>).
Prop<{}> is already unassignable to Prop<T>. Even though this issue has existed in some form since TS 1.8 (PR 5092), the non-assignability would prevent the false-side narrowing for this particular piece of code, until PR 13230. I don't think it makes any sense that Prop<{}> gets narrowed to Prop<T> on the true side and to never on the false side, or that one should have to make it "even more unassignable" to Prop<T> to prevent that.
The bug is still present in 2.3rc and in master
| gharchive/issue | 2017-02-07T04:47:22 | 2025-04-01T06:37:13.404517 | {
"authors": [
"aluanhaddad",
"jeffreymorlan",
"mhegazy"
],
"repo": "Microsoft/TypeScript",
"url": "https://github.com/Microsoft/TypeScript/issues/13918",
"license": "Apache-2.0",
"license_type": "permissive",
"license_source": "github-api"
} |
217370799 | Duplicate identifier error with readonly/getter, class and merged interface
TypeScript Version: 2.2.1
I'm having a lot of trouble finding an explanation for the errors below:
Code
// Co-named interface and class doesn't like readonly property implementation:
interface Foo {
readonly x: number; // Error: Duplicate identifier 'x'
y: number;
}
class Foo {
get x(): number { // Error: Duplicate identifier 'x'
return 0;
}
y = 1;
}
// Same as above, but different class name + explicit `implements`
class Bar implements Foo {
get x(): number { // No error!
return 0;
}
y = 1;
}
// Duplicating the first example, but explicitly implementing the co-named interface:
interface Baz {
readonly x: number; // Error: Duplicate identifier 'x'
y: number;
}
class Baz implements Baz {
get x(): number { // Error: Duplicate identifier 'x'
return 0;
}
y = 1;
}
Why does TypeScript complain about duplicate identifiers between a readonly field in an interface that merges with a class that implements that field as a getter?
An interface and a class with the same name "merge".
more information can be found at: http://www.typescriptlang.org/docs/handbook/declaration-merging.html
@mhegazy
I have the same question with this. For property and it's getter, type checker should not consider them as different during merging.
Also, there's some other question: when readonly property being merged, the property can not be assigned in constructor.
interface Rect {
width: number;
height: number;
readonly area: number;
}
class Rect implements Rect {
readonly area: number;
constructor(public x: number,public y: number){
this.area = this.width * this.height; // Error: Cannot assign to 'area' because it is a constant or a read-only property.
}
}
| gharchive/issue | 2017-03-27T20:48:57 | 2025-04-01T06:37:13.408528 | {
"authors": [
"axefrog",
"mhegazy",
"uushuo"
],
"repo": "Microsoft/TypeScript",
"url": "https://github.com/Microsoft/TypeScript/issues/14882",
"license": "Apache-2.0",
"license_type": "permissive",
"license_source": "github-api"
} |
277337461 | Type inference for indexed array element
TypeScript Version: 2.6.1
Code
/* EXAMPLE 1 */
var a: (number | undefined)[] = [1, 2, undefined, 4];
var b = a[2];
if (b) {
b; // number
}
if (a[2]) {
var c = a[2]; // SHOULD be number BUT is number | undefined
}
/* EXAMPLE 2 */
function aa(a: (number | undefined)[]) {
for (var i=0; i<a.length; i++) {
if (a[i]) {
var p = a[i]; // SHOULD be number BUT is number | undefined
}
}
}
Part of #9998
(I thought there was something more specific to tracking array index accesses, but couldn't find it)
Duplicate of https://github.com/Microsoft/TypeScript/issues/10530
| gharchive/issue | 2017-11-28T10:44:51 | 2025-04-01T06:37:13.411253 | {
"authors": [
"kitsonk",
"mhchem",
"mhegazy"
],
"repo": "Microsoft/TypeScript",
"url": "https://github.com/Microsoft/TypeScript/issues/20301",
"license": "Apache-2.0",
"license_type": "permissive",
"license_source": "github-api"
} |
284883734 | ES6 Conversion error，
this is my code
[...Array(10).keys()]
The result is on the official website
Array(10).keys().slice();
But it is wrong
Uncaught TypeError: Array(...).keys(...).slice is not a function at <anonymous>:1:18
TypeScript doesn't polyfill methods - if you are running on a runtime without Array#keys, you need to provide the polyfill yourself
Which runtime can run that ES5 output from Typescript? In Node v8.6.0, Firefox, and Chrome:
Array(10).keys() works
[...Array(10).keys()] works
but the transpiled ES5 output from the 2nd one there (Array(10).keys().slice()) causes TypeError: Array(...).keys(...).slice is not a function in Firefox, Chrome, and Node 8, even though the original ES6 does work.
@j127 that code is flagged as an error when targeting ES5
If you run with --target ES5 --lib es6 --downlevelIteration, which is the only way to have that code not issue a type error, then we emit a proper spread helper:
var __read = (this && this.__read) || function (o, n) {
var m = typeof Symbol === "function" && o[Symbol.iterator];
if (!m) return o;
var i = m.call(o), r, ar = [], e;
try {
while ((n === void 0 || n-- > 0) && !(r = i.next()).done) ar.push(r.value);
}
catch (error) { e = { error: error }; }
finally {
try {
if (r && !r.done && (m = i["return"])) m.call(i);
}
finally { if (e) throw e.error; }
}
return ar;
};
var __spread = (this && this.__spread) || function () {
for (var ar = [], i = 0; i < arguments.length; i++) ar = ar.concat(__read(arguments[i]));
return ar;
};
var p = __spread(Array(10).keys());
Thanks. Isn't that still a bug though? (or at least undesired behavior) I have it set to output ES5 in my tsconfig file, and it's compiling working JavaScript to non-working JavaScript.
Isn't that still a bug though?
If target is es5 then we can't just emit the spread as-is, because spreads are not legal ES6. What do you propose we do differently?
Sorry if I'm confused here. I'm not understanding why it would transpile code to Array(10).keys().slice(). Is there any ES5 interpreter that can run that code?
The output doesn't seem to match the target. If you're writing ES6/Typescript and transpiling to ES5, it seems like the transpired version should work in ES5 browsers. Node 9.5, Chrome 64, and Firefox 59 can't run that "ES5" snippet.
Under the constraints given, TS can't emit correct code, so it issues an error and tries to do the best it can.
What would you expect to be emitted instead?
What would you expect to be emitted instead?
I expect to write ES6 code and have it transpiled to code that runs in the most common ES5 interpreters. :)
Right now, I'm writing valid ES6 and Typescript doesn't transpile it to working ES5. The exact output doesn't matter -- just that it runs.
I don't know the best way to do it, but this is one way that works:
// ES6
[...Array(10).keys()]
// => [ 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 ]
// working ES5
Array.apply(null, Array(10)).map(function (_, idx) {return idx;});
// => [ 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 ]
I expect to write ES6 code and have it transpiled to code that runs in the most common ES5 interpreters. :)
I think your expectation has to be "TypeScript emits code that runs in the specified runtime, or issues an error when it can't". TypeScript issued an error in this case; you can't expect code to work if TS has told you (in one way or another) that it can't emit correct code.
Your proposed emit only works for the exact special case of spreading Array.keys; there are many other iterator-returning functions that can't be turned into a map call.
Thanks, I see the error now when running it from the command line. I missed it when compiling from VS Code. I only saw the broken output file, assuming that it was supposed to work. Sorry for my confusion.
| gharchive/issue | 2017-12-28T10:13:16 | 2025-04-01T06:37:13.420933 | {
"authors": [
"RyanCavanaugh",
"j127",
"lwpersonal"
],
"repo": "Microsoft/TypeScript",
"url": "https://github.com/Microsoft/TypeScript/issues/20916",
"license": "Apache-2.0",
"license_type": "permissive",
"license_source": "github-api"
} |
354287363 | OpenExternalProjects command doesn't work in latest master (typescript@next)
We use integration with the typescript service. After a0190e397abd2ca7fbff948f869f6e9573c0b117
(editorServices.ts line 2615 I guess) the typescript service command "OpenExternalProjects" doesn't work properly. The open project always stands empty because of "pendingReload".
The test case:
Call OpenExternalProjects for some tsconfig.json
Open file under the tsconfig (but tsconfig must be not in the same or parent directory). e.g.
project/config/tsconfig.json:
{ "include": [ "../foo.ts", ] }
project/foo.ts
class Foo {}
Expected: file is open as part of the project
Actual: file is open under "inferred" project
//cc @sheetalkamat
seems the problem was resolved in the current typescript@next.
| gharchive/issue | 2018-08-27T11:50:17 | 2025-04-01T06:37:13.424676 | {
"authors": [
"anstarovoyt"
],
"repo": "Microsoft/TypeScript",
"url": "https://github.com/Microsoft/TypeScript/issues/26686",
"license": "Apache-2.0",
"license_type": "permissive",
"license_source": "github-api"
} |
377629878 | React StatelessComponent regression
TypeScript Version: 3.2.0-dev.20181103
Code
import React = require("react");
class C<P> extends React.Component<{ component: React.ComponentType<P> } & Partial<P>> {}
declare const myStatelessComponent: React.StatelessComponent;
<C component={myStatelessComponent} />
Expected behavior:
No error, as in ts3.1.
Actual behavior:
src/a.tsx:4:4 - error TS2322: Type 'StatelessComponent<{}>' is not assignable to type 'ComponentType<{ component: {}; }>'.
Type 'StatelessComponent<{}>' is not assignable to type 'StatelessComponent<{ component: {}; }>'.
Types of property 'propTypes' are incompatible.
Type 'ValidationMap<{}> | undefined' is not assignable to type 'ValidationMap<{ component: {}; }> | undefined'.
Type 'ValidationMap<{}>' is not assignable to type 'ValidationMap<{ component: {}; }>'.
Property 'component' is missing in type 'ValidationMap<{}>'.
4 <C component={myStatelessComponent} />
~~~~~~~~~
src/a.tsx:2:38
2 class C<P> extends React.Component<{ component: React.ComponentType<P> } & Partial<P>> {}
~~~~~~~~~
The expected type comes from property 'component' which is declared here on type 'IntrinsicAttributes & IntrinsicClassAttributes<C<{ component: {}; }>> & Readonly<{ children?: ReactNode; }> & Readonly<{ component: ComponentType<{ component: {}; }>; } & Partial<{ component: {}; }>>'
Found 1 error.
Discovered in redux-form on DefinitelyTyped.
The error is correct - you want myStatelessComponent to be a React.StatelessComponent<any> or React.StatelessComponent<{component: React.StatelessComponent<any>}> and not a React.StatelessComponent<{}>. Why, you ask? Because C claims that it takes a component which takes the same props it does. If you pass a stateless component that only takes {}, it does not take the component property that the component itself is passed in thru. 😄
The error being on the propTypes member rather than somewhere more useful is something the react .d.ts authors know of and are looking into.
| gharchive/issue | 2018-11-05T23:38:42 | 2025-04-01T06:37:13.429045 | {
"authors": [
"andy-ms",
"weswigham"
],
"repo": "Microsoft/TypeScript",
"url": "https://github.com/Microsoft/TypeScript/issues/28351",
"license": "Apache-2.0",
"license_type": "permissive",
"license_source": "github-api"
} |
140207859 | New Website Logo: LiveBet
Please fill the following out:
Name: LiveBet
Homepage url: http://www.livebet.com/
Brand Guidelines/Licensing: Feel free to use the logo in white or black background
Logo: http://www.livebet.com/a/img/LiveBet_600x220.png
Example:
Name: Microsoft
Homepage url: https://microsoft.com
Brand Guidelines/Licensing: https://www.microsoft.com/en-us/legal/intellectualproperty/trademarks/usage/logo.aspx
Logo: https://news.microsoft.com/microsoft-news-center-photos/msft_logo_rgb_c-gray/
Should be in the next site update in the next few days.
| gharchive/issue | 2016-03-11T15:12:11 | 2025-04-01T06:37:13.432947 | {
"authors": [
"mhegazy",
"michalas"
],
"repo": "Microsoft/TypeScript",
"url": "https://github.com/Microsoft/TypeScript/issues/7482",
"license": "Apache-2.0",
"license_type": "permissive",
"license_source": "github-api"
} |
260763231 | Fixes to @augments handling
Fixes #18740
We will now get a superclass from @augments even if there is no extends.
Allows us to parse /** @augments A */ without the curly braces, because that seems to be the style.
We will no longer parse arbitrary types, only a class (qualified) name followed by type arguments. So you can't declare to inherit from () => void without parse errors.
Adds errors for:
@augments tag not on a class
@augments A but extends B
@andy-ms <> aren't legal characters on Windows file systems. I can't git pull after this commit:
E:\src\TypeScript>git pull upstream master
From https://github.com/Microsoft/TypeScript
* branch master -> FETCH_HEAD
error: unable to create file tests/baselines/reference/JSDocParsing/DocComments.parsesCorrectly.<> characters.json: Invalid argument
Note: i can just update my .gitignore but FYI
| gharchive/pull-request | 2017-09-26T20:42:34 | 2025-04-01T06:37:13.436207 | {
"authors": [
"andy-ms",
"minestarks"
],
"repo": "Microsoft/TypeScript",
"url": "https://github.com/Microsoft/TypeScript/pull/18775",
"license": "Apache-2.0",
"license_type": "permissive",
"license_source": "github-api"
} |
303203267 | Port generated lib files
Ports changes in https://github.com/Microsoft/TSJS-lib-generator/pull/383
@mhegazy This change broke my code because selectionStart: number is now selectionStart: number | null on interface HTMLInputElement. The same for selectionEnd.
However I don't believe this is correct because MDN doesn't mention that the property is nullable.
When would this return null?
Please file a new ticket and will get it fixed.
@mhegazy Thanks, I created #23094
| gharchive/pull-request | 2018-03-07T18:07:52 | 2025-04-01T06:37:13.438922 | {
"authors": [
"mhegazy",
"styfle"
],
"repo": "Microsoft/TypeScript",
"url": "https://github.com/Microsoft/TypeScript/pull/22381",
"license": "Apache-2.0",
"license_type": "permissive",
"license_source": "github-api"
} |
337641697 | Use addRelatedInfo.
Addresses https://github.com/Microsoft/TypeScript/pull/25359#discussion_r199599823
@DanielRosenwasser Missing semicolon XD
| gharchive/pull-request | 2018-07-02T19:51:41 | 2025-04-01T06:37:13.440108 | {
"authors": [
"DanielRosenwasser",
"weswigham"
],
"repo": "Microsoft/TypeScript",
"url": "https://github.com/Microsoft/TypeScript/pull/25377",
"license": "Apache-2.0",
"license_type": "permissive",
"license_source": "github-api"
} |
160964987 | [Salsa] Adding base indentation for script block formatting and smart indent
[Salsa] Adding base indentation for script block formatting and smart indent
@vladima @mhegazy: can you take a look at this?
can you also add fourslash tests for formatting/indentation that will use BaseIndentSize?
Good idea - will add some tests.
:+1:
| gharchive/pull-request | 2016-06-17T20:05:49 | 2025-04-01T06:37:13.442349 | {
"authors": [
"jramsay",
"vladima"
],
"repo": "Microsoft/TypeScript",
"url": "https://github.com/Microsoft/TypeScript/pull/9236",
"license": "Apache-2.0",
"license_type": "permissive",
"license_source": "github-api"
} |
378795747 | RFC: Background object cache maintenance job
The (shared) object cache is growing to be large. This includes:
Number of loose objects.
Number of pack-files.
Size of pack-files.
There are some things we can do to improve this situation, with varying levels of difficulty and benefit. Here are the four items I believe will have a strong impact on our data size and performance:
I. Delete loose objects that exist in pack-files.
II. Place loose objects into pack-files.
III. Repack some prefetch packs into larger packs.
IV. Delete packs with no objects referenced by the multi-pack-index.
These are numbered by order of implementation, but they work together in an interesting way. If we place loose objects into pack-files (I), then the next time we delete loose objects (II) we will delete those objects that we just put into a pack-file. If we repack some prefetch packs into larger packs (III), the multi-pack-index will use the "newer" pack for all objects in those packs, and we can delete them (IV).
In order to perform this maintenance without disrupting users, we can actually do these interactions in two independent jobs, with their sub-actions in a particular order:
A. Loose Object Maintenance
Delete loose objects that exist in pack-files.
Place loose objects into pack-files.
B. Prefetch Pack Maintenance
Delete packs with no objects referenced by the multi-pack-index.
Repack some prefetch packs into larger packs.
Care must be taken to ensure that Git processes that initialized their object store before the maintenance still succeed. Thus, we should never delete anything that those processes could be reading from. This is why we delete loose objects only if they are referenced by pack-files. This is why we delete prefetch packs only if they have no objects referenced by the multi-pack-index. Since a second run of these jobs could delete objects that were repacked in the previous run, we need to space our job runs such that no Git processes span both runs.
Loose Object Job
The loose object job can save users who have many loose object downloads. These loose objects are populated by several user activities: running git fetch will get loose commits and trees for the ref tips that are not covered by prefetch packs, hydrating a file will trigger a loose blob, and many other. Older enlistments have many loose commits and trees due to the behavior of git fetch before the --no-show-forced-updates option.
Delete loose objects that appear in a pack-file
By running git prune-packed -q with the environment variable GIT_OBJECT_DIRECTORY set to the object cache, Git will delete the loose objects that exist in a pack-file. This will use the multi-pack-index to test containment, so is only bound by the file-deletion operations.
Pack loose objects that do not appear in a pack-file
We can use git pack-objects -q <object-dir>/pack/loose-<timestamp> to create a pack-file filled with a list of loose objects. We can pass the list of loose objects to this command through stdin. We can choose to batch this operation by limiting the number of objects we pack at a time, if we want. We may also want to skip this step if the number of loose objects is low (< 1,000? < 10,000?)
By listing a timestamp in the pack-file name, we can store the "most-recent" time this job was run.
Note: While we are thinking about packing loose objects, we can consider deleting some loose blobs. It is easier to delete loose blobs than packed blobs. If we want to delete "stale" blobs, then we should add a step to this job.
Prefetch Pack Job
We store a list of prefetch packs as <object-dir>/pack/prefetch-<timestamp>-<sha>.[pack|idx]. These are downloaded at a regular cadence by the post-fetch job, and could have a new pack-file for every hour. In addition to these prefetch packs, some daily prefetch packs will appear and occasionally a very large prefetch pack covering everything more than 30 days old. This large prefetch pack appears on initial clone and sometimes again if a user does not prefetch for 30 days (this occurs less often now that background prefetch exists).
Delete pack-files with no objects referenced by the multi-pack-index
The multi-pack-index stores a single copy of every packed object. When an object appears in multiple pack-files covered by the multi-pack-index, we select the pack-file with the most-recent modified time as reported by the filesystem. For example, if someone downloads a second large prefetch pack when prefetching for the first time in 31 days, then that large prefetch pack is referenced and all older prefetch packs are "stale" and not referenced by the multi-pack-index.
We can extend the git multi-pack-index builtin to delete pack-files that have no objects referenced by the multi-pack-index. This should happen at the same time as we update the multi-pack-index since we don't want to have this list of pack-files in the multi-pack-index if they don't exist. We also don't want to remove them from the multi-pack-index while the pack-files still exist, because then a Git process will start referencing those pack-files in the packed_git linked list.
Repack the prefetch pack-files
We can reduce the number of prefetch packs and improve file locality by combining prefetch packs into new "repacked" pack-files. Use git show-index <object-dir>/pack/prefetch-<timestamp[i]>-<hash[i]>.idx to list the objects in the prefetch packs and pipe that output to git pack-objects -q <object-dir>/pack/repacked-<timestamp>.
We should refrain from packing the most-recent prefetch packs, as we don't want to get in the way of the prefetch machinery for determining the most-recent timestamp. We could, however, start repacking the prefetch packs from oldest-to-newest, stopping in the most-recent week, or when the prefetch packs are large enough to satisfy a batch. Some stopping conditions could include "pack-file size is above 1GB".
Running the Jobs
We already have a post-fetch job that is automatically queued in the background. We should rename this job to be a "maintenance" job that could trigger several actions. The maintenance job should run every 15 minutes (as the post-fetch job currently does) but then it should check each "sub-job" for a timing condition to run again.
For example:
The post-fetch job doesn't trigger if the most-recent prefetch pack timestamp is within the last 70 minutes.
The loose object job doesn't trigger if the most-recent "loose" pack timestamp is within the last 24 hours.
The prefetch repack job doesn't trigger if the most-recent "repacked" pack timestamp is within the last 7 days or the most-recent "repacked" pack is modified within the last 24 hours.
Each job that may delete data should take a file-based lock on the object cache. This could be a "maintenance lock".
In addition to this background activity, we can enable manually running these jobs as a verb, say gvfs maintenance.
Secondary Concerns
We will want to track some data improvements using logging. We could report the number of loose objects before and after the loose object job, for instance. We could report the number and total size of pack-files, possibly grouped by the prefix; we expect pack-files with prefixes including "pack", "prefetch", "loose", and "repacked".
Mentioning @jrbriggs, @sanoursa, @kewillford, @wilbaker, @jamill, @nickgra, @alameenshah, @jeschu1, @mjcheetham for feedback. Thanks!
Please use this PR review for comments.
Use #492 for further design discussion. The problems this design tries to solve are covered in #85.
Reopening, as this is rewritten to be about the new cleanup steps to add in the future. @jeschu1 is working on the loose object cleanup step next, I believe.
| gharchive/issue | 2018-11-08T15:55:29 | 2025-04-01T06:37:13.458994 | {
"authors": [
"derrickstolee"
],
"repo": "Microsoft/VFSForGit",
"url": "https://github.com/Microsoft/VFSForGit/issues/488",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
370321131 | Shouldn't this say "you can't use WSL commands here"?
My Windows build number: Microsoft Windows [Version 10.0.16299.665]
What I'm doing and what's happening: I'm trying to install themes for Hyper.JS, my preferred client for accessing WSL.
What's wrong / what should be happening instead: I get this error: Warning! Due to WSL limitations, you can use CLI commands here. Please use Hyper CLI on cmd, PowerShell or GitBash/CygWin.rShell or GitBash/CygWin.
This is confusing to me. Can or can't I use CLI commands here? I think it's trying to say that I can't, and I think this error message should be changed to: Warning! Due to WSL limitations, you can't use CLI commands here.
This error message is owned by the Hyper.js folks, not us.
| gharchive/issue | 2018-10-15T20:01:13 | 2025-04-01T06:37:13.462518 | {
"authors": [
"benhillis",
"mike-u"
],
"repo": "Microsoft/WSL",
"url": "https://github.com/Microsoft/WSL/issues/3617",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
210162563 | Pori app gets confused and starts to render seemingly-random images
Navigate around a bit and you'll quickly see various images getting swapped out for others.
This looks like a regression caused by the CG refactor. Thanks, @DHowett-MSFT for help confirming.
Disabling the caching behavior in DisplayTexture.mm (CachedDisplayTextureForImage, etc.) works around the issue.
| gharchive/issue | 2017-02-24T21:28:47 | 2025-04-01T06:37:13.464644 | {
"authors": [
"jaredhms"
],
"repo": "Microsoft/WinObjC",
"url": "https://github.com/Microsoft/WinObjC/issues/2096",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
430578518 | Missing WCAG link
Describe the bug
A clear and concise description of what the bug is.
To Reproduce
Steps to reproduce the behavior:
Go to any page
Run Accessibility Insights for Web > Assessment
Go to Multimedia > Audio Description > more info and examples
missing link for WCAG 1.2.3
Expected behavior
Add a link in "WCAG Success criteria" for this requirement.
Name: Understanding Success Criterion 1.2.3: Audio Description or Media Alternative (Prerecorded)
link: https://www.w3.org/WAI/WCAG21/Understanding/audio-description-or-media-alternative-prerecorded
Screenshots
.
Extension (please complete the following information)
NA
Are you willing to submit a PR?
If asked, will you be willing to submit a PR to fix this bug? No
Did you search for similar existing issues?
Yes
@LiLoDavis I do not think there needs to be more links for sufficient techniques or additional guidance?
@ferBonnin Agreed.
| gharchive/issue | 2019-04-08T17:48:16 | 2025-04-01T06:37:13.473430 | {
"authors": [
"LiLoDavis",
"ferBonnin"
],
"repo": "Microsoft/accessibility-insights-web",
"url": "https://github.com/Microsoft/accessibility-insights-web/issues/507",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
438054169 | Add assignees for "Release" issues
Added few more Bot rules for "Release" specific issues.
@bryanmacfarlane - Can you please review the changes done to issues-rules.yml for the bot rules.
👍
Can you please explicitly approve the PR. Currently it is blocked as there is no approver.
Thanks @bryanmacfarlane / @sachinma for the review.
| gharchive/pull-request | 2019-04-28T12:59:54 | 2025-04-01T06:37:13.477000 | {
"authors": [
"bryanmacfarlane",
"kmkumaran"
],
"repo": "Microsoft/azure-pipelines-tasks",
"url": "https://github.com/Microsoft/azure-pipelines-tasks/pull/10245",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
239682227 | make Azure SDK versions as properties
@yungez @aovechki @xscript
fix #12
@ZhijunZhao,
Thanks for your contribution as a Microsoft full-time employee or intern. You do not need to sign a CLA.
Thanks,
Microsoft Pull Request Bot
| gharchive/pull-request | 2017-06-30T05:40:06 | 2025-04-01T06:37:13.478394 | {
"authors": [
"ZhijunZhao",
"msftclas"
],
"repo": "Microsoft/azure-spring-boot-starters",
"url": "https://github.com/Microsoft/azure-spring-boot-starters/pull/14",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
310609820 | define state manager behavior for endOfConversation and deleteUserData
It'd be useful for the SDK to provide an option to automatically clear ConversationState on receiving an endOfConversation activity, and automatically clear UserState when receiving deleteUserData.
If we don't provide this as a built-in option of the state manager, then we should provide a sample showing how to do it as middleware.
document endOfConversation through a sample. this is on the v4 Sample list.
@dandriscoll does any channel send 'deleteUserData' events?
| gharchive/issue | 2018-04-02T21:00:10 | 2025-04-01T06:37:13.479772 | {
"authors": [
"DeniseMak",
"sgellock"
],
"repo": "Microsoft/botbuilder-js",
"url": "https://github.com/Microsoft/botbuilder-js/issues/159",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
219775896 | https://github.com/Microsoft/cppwinrt/blob/master/10.0.14393.0/Samples/CL/main.cpp report error "c3861:'initialize': identifier not found",
Sample https://github.com/Microsoft/cppwinrt/blob/master/10.0.14393.0/Samples/CL/main.cpp has error "c3861:'initialize': identifier not found", since ‘initialize’ have renamed to init_apartment, this issue fixed with revision ‘74c56cb35’, however, the latest revision ‘532b09e’ for WinRT looks report this issue again. and we can’t find the revision ‘74c56cb35’, does anyone revert that fix?
Here is details for revision ‘74c56cb35’:
D:\WinRT\src>git show 74c56cb35fdf7ebcf13f9b99c2469387cadfe617
commit 74c56cb35fdf7ebcf13f9b99c2469387cadfe617
Author: Brent Rector brent.rector@microsoft.com
Date: Wed Mar 8 13:28:40 2017 -0800
Update main.cpp
diff --git a/10.0.14393.0/Samples/CL/main.cpp b/10.0.14393.0/Samples/CL/main.cpp
index 9b33abd..8164f51 100644
--- a/10.0.14393.0/Samples/CL/main.cpp
+++ b/10.0.14393.0/Samples/CL/main.cpp
@@ -6,5 +6,5 @@ using namespace winrt;
int main()
{
initialize();
init_apartment();
}
Sorry this happened again Phoebe. Our publishing process needs a bit of work... :disappointed:
I new release has just been dropped that should solve your problem :)
Thank you Kenny and Kevin! WinRT is clean now.
| gharchive/issue | 2017-04-06T03:22:06 | 2025-04-01T06:37:13.502589 | {
"authors": [
"KevinWelton",
"PhoebeHui",
"kennykerr"
],
"repo": "Microsoft/cppwinrt",
"url": "https://github.com/Microsoft/cppwinrt/issues/163",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
156706939 | update distributed word embedding
add communicator to get and add parameter
communicator own a memory_manager
implement pipeline
trainer get datablock from LoadBlockThread
use the datablock to save parameter
refine SaveEmbedding function
Please update the submodule of multiverso as well
| gharchive/pull-request | 2016-05-25T09:44:44 | 2025-04-01T06:37:13.504540 | {
"authors": [
"chivee",
"xuehui1991"
],
"repo": "Microsoft/distributed_word_embedding",
"url": "https://github.com/Microsoft/distributed_word_embedding/pull/4",
"license": "mit",
"license_type": "permissive",
"license_source": "bigquery"
} |
181956959 | Adding ability to publish VSIX
Allows our CI build to publish updates to vsixgallery
Fixes nuget warnings while packaging nuget package because libraries are not stable releases
Are you handling version numbering for CD builds? Right now, versions are taken from the version.txt files, but to do CD, we would want at least the revision number changing on each deployment
Are you handling version numbering for CD builds? Right now, versions are taken from the version.txt files, but to do CD, we would want at least the revision number changing on each deployment
We don't use version.txt for ApiPort.Vsix builds. But, in restore.ps1, we increment the "revision" number by default, so takes into account the version that exists in the vsix manifest.
We should consolidate the versioning strategy, but that'll work for now
| gharchive/pull-request | 2016-10-10T08:10:57 | 2025-04-01T06:37:13.507594 | {
"authors": [
"conniey",
"twsouthwick"
],
"repo": "Microsoft/dotnet-apiport",
"url": "https://github.com/Microsoft/dotnet-apiport/pull/348",
"license": "mit",
"license_type": "permissive",
"license_source": "bigquery"
} |
124130212 | Enabled proper SSL Offloading
A redirect loop occurred when trying to access a secure site in hybris.
This changes allow a proper SSL offloading
Second try to get pull request
| gharchive/pull-request | 2015-12-28T21:35:51 | 2025-04-01T06:37:13.513832 | {
"authors": [
"nploner"
],
"repo": "Microsoft/microsoft-deployment-accelerator-for-hybris-on-azure",
"url": "https://github.com/Microsoft/microsoft-deployment-accelerator-for-hybris-on-azure/pull/6",
"license": "mit",
"license_type": "permissive",
"license_source": "bigquery"
} |
115421839 | Run tests from deployment dir
Depends on Pull Request #341.
Build items are generated into a new Output directory
No dependencies are copied in the Output directory
Tests are run from a separate deployment directory. Target specific dependencies are copied into it.
A target specific deployment directory is created ( Specified by TargetPlatformIdentifier). Currently it has the same contents as the test deployment directory. In the future it won't have the test dependencies.
Mkay, I changed the build script such that it does not invoke the msbuild script with parameters. Instead, the deploydependencies targets copy to both deployment and test directories.
@ValMenn I did two copy tasks and removed the incorrect batching.
LGTM :shipit:
| gharchive/pull-request | 2015-11-06T02:36:12 | 2025-04-01T06:37:13.516406 | {
"authors": [
"cdmihai",
"dsplaisted"
],
"repo": "Microsoft/msbuild",
"url": "https://github.com/Microsoft/msbuild/pull/336",
"license": "mit",
"license_type": "permissive",
"license_source": "bigquery"
} |
171018836 | Microsoft
Issue #
Bug
Fix
Testing
Nodejstool
Hi @vnnd, I'm your friendly neighborhood Microsoft Pull Request Bot (You can call me MSBOT). Thanks for your contribution!
This seems like a small (but important) contribution, so no Contribution License Agreement is required at this point. Real humans will now evaluate your PR.
TTYL, MSBOT;
| gharchive/pull-request | 2016-08-13T17:35:12 | 2025-04-01T06:37:13.522387 | {
"authors": [
"msftclas",
"vnnd"
],
"repo": "Microsoft/nodejstools",
"url": "https://github.com/Microsoft/nodejstools/pull/1204",
"license": "apache-2.0",
"license_type": "permissive",
"license_source": "bigquery"
} |
335318451 | (Merge #760 into release branch)[Pylon] Refine and fix bugs
Merge #760 into release branch.
@hwuu please take a look.
Coverage increased (+0.01%) to 50.984% when pulling cc8a6490eac69db743bff19b7d6ee70780a9d728 on hwuu/pylon into aac6f30342811f7648f3c9d89b63312bf0b738b0 on pai-0.6.y.
Coverage increased (+0.01%) to 50.984% when pulling cc8a6490eac69db743bff19b7d6ee70780a9d728 on hwuu/pylon into aac6f30342811f7648f3c9d89b63312bf0b738b0 on pai-0.6.y.
Coverage increased (+0.01%) to 50.984% when pulling cc8a6490eac69db743bff19b7d6ee70780a9d728 on hwuu/pylon into aac6f30342811f7648f3c9d89b63312bf0b738b0 on pai-0.6.y.
| gharchive/pull-request | 2018-06-25T09:02:12 | 2025-04-01T06:37:13.527254 | {
"authors": [
"coveralls",
"hao1939"
],
"repo": "Microsoft/pai",
"url": "https://github.com/Microsoft/pai/pull/766",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
214541644 | Search in documentation doesn't work in Edge
Try searching anything in the documentation https://pxt.microbit.org/projects
Does nothing.
Works on IE11\Firefox & chrome
Works for me. Can you try the latest? or perhaps it's a different repro.
Works for me.
@abchatra still repro'ing?
Does nothing in Safari on my iPhone and iPad with iOS 9. Works for me on macOS.
Repros for me. Not for Richard. I could not create a small repro. May be I have weird settings in edge
Repros for me on IOS. @abchatra are you using the sidebar TOC for this? I.e: did you have to pull out the TOC using the hamburger icon or not?
From: Abhijith Chatra notifications@github.com
Sent: Saturday, March 18, 2017 7:44:02 AM
To: Microsoft/pxt
Cc: Sam El-Husseini; Assign
Subject: Re: [Microsoft/pxt] Search in documentation doesn't work in Edge (#1652)
Repros for me. Not for Richard. I could not create a small repro. May be I have weird settings in edge
—
You are receiving this because you were assigned.
Reply to this email directly, view it on GitHubhttps://github.com/Microsoft/pxt/issues/1652#issuecomment-287550570, or mute the threadhttps://github.com/notifications/unsubscribe-auth/AP6rzKN3Tlg6QP33sjfS_BKwzjBa_qHUks5rm-2ygaJpZM4MemGp.
Richard is on Android.
Get Outlook for iOShttps://aka.ms/o0ukef
From: Sam El-Husseini notifications@github.com
Sent: Saturday, March 18, 2017 8:36:31 AM
To: Microsoft/pxt
Cc: Peli de Halleux; Comment
Subject: Re: [Microsoft/pxt] Search in documentation doesn't work in Edge (#1652)
Repros for me on IOS. @abchatra are you using the sidebar TOC for this? I.e: did you have to pull out the TOC using the hamburger icon or not?
From: Abhijith Chatra notifications@github.com
Sent: Saturday, March 18, 2017 7:44:02 AM
To: Microsoft/pxt
Cc: Sam El-Husseini; Assign
Subject: Re: [Microsoft/pxt] Search in documentation doesn't work in Edge (#1652)
Repros for me. Not for Richard. I could not create a small repro. May be I have weird settings in edge
—
You are receiving this because you were assigned.
Reply to this email directly, view it on GitHubhttps://github.com/Microsoft/pxt/issues/1652#issuecomment-287550570, or mute the threadhttps://github.com/notifications/unsubscribe-auth/AP6rzKN3Tlg6QP33sjfS_BKwzjBa_qHUks5rm-2ygaJpZM4MemGp.
—
You are receiving this because you commented.
Reply to this email directly, view it on GitHubhttps://github.com/Microsoft/pxt/issues/1652#issuecomment-287553825, or mute the threadhttps://github.com/notifications/unsubscribe-auth/AD-4KUo-zAsczeojqfTa5nD_413D99qYks5rm_n_gaJpZM4MemGp.
Chrome on iOS treats the form submission as a pop up.
Get Outlook for iOShttps://aka.ms/o0ukef
From: Sam El-Husseini notifications@github.com
Sent: Saturday, March 18, 2017 8:36:31 AM
To: Microsoft/pxt
Cc: Peli de Halleux; Comment
Subject: Re: [Microsoft/pxt] Search in documentation doesn't work in Edge (#1652)
Repros for me on IOS. @abchatra are you using the sidebar TOC for this? I.e: did you have to pull out the TOC using the hamburger icon or not?
From: Abhijith Chatra notifications@github.com
Sent: Saturday, March 18, 2017 7:44:02 AM
To: Microsoft/pxt
Cc: Sam El-Husseini; Assign
Subject: Re: [Microsoft/pxt] Search in documentation doesn't work in Edge (#1652)
Repros for me. Not for Richard. I could not create a small repro. May be I have weird settings in edge
—
You are receiving this because you were assigned.
Reply to this email directly, view it on GitHubhttps://github.com/Microsoft/pxt/issues/1652#issuecomment-287550570, or mute the threadhttps://github.com/notifications/unsubscribe-auth/AP6rzKN3Tlg6QP33sjfS_BKwzjBa_qHUks5rm-2ygaJpZM4MemGp.
—
You are receiving this because you commented.
Reply to this email directly, view it on GitHubhttps://github.com/Microsoft/pxt/issues/1652#issuecomment-287553825, or mute the threadhttps://github.com/notifications/unsubscribe-auth/AD-4KUo-zAsczeojqfTa5nD_413D99qYks5rm_n_gaJpZM4MemGp.
Yeah it seems IOS Safari doesn't like that it's target _blank. Removed that for the sidebar TOC.
Fixed.
It looks still repro on Edge,Firefox,IE & Chrome,that search in documentation doesn't work.
Clicking on the search icon isn't how that search works. Although come to think of it maybe it should.
Search works on ENTER.
Bing & google work on clicking search icon
Yeah I agree @abchatra, we should enable that option.
I shall use this issue to track that work.
| gharchive/issue | 2017-03-15T22:10:41 | 2025-04-01T06:37:13.545470 | {
"authors": [
"Helen09",
"abchatra",
"martinwork",
"microsoftsam",
"pelikhan"
],
"repo": "Microsoft/pxt",
"url": "https://github.com/Microsoft/pxt/issues/1652",
"license": "mit",
"license_type": "permissive",
"license_source": "bigquery"
} |
284398117 | Shaking simulator, block "pick random" only works once
Repro steps:
• Navigating to https://makecode.microbit.org/beta
• Drag blocks "on shake", "show number" and "pick random 0 to 4"
• Shake the simulator and observe
Expected Result:
Shaking simulator should be same as clicking "SHAKE" button, block "pick ramdom" always works
Actual Result:
When shaking the simulator, after block "pick random" works once and the number doesn't change anymore
I have moved this issue to https://github.com/Microsoft/pxt-microbit/issues
| gharchive/issue | 2017-12-25T03:16:43 | 2025-04-01T06:37:13.549163 | {
"authors": [
"Helen09"
],
"repo": "Microsoft/pxt",
"url": "https://github.com/Microsoft/pxt/issues/3672",
"license": "mit",
"license_type": "permissive",
"license_source": "bigquery"
} |
390888637 | Deprecated no-reserved-keywords
PR checklist
[x] Addresses an existing issue: fixes #652
[x] New feature, bugfix, or enhancement
[x] Includes tests
[x] Documentation update
Overview of change:
Marks the rule as deprecated per #652.
:shipit: :shipit: :shipit:
@JoshuaKGoldberg, are you planning on a release today? (Just to know if I temporarily tslint:disable in my code or not :D)
:shipit: :shipit: :shipit:
Lol, forgot about that piece of history...
I wasn't planning on it, since this is the only user-facing change that goes in it. It's probably better to just disable no-reserved-keywords. Sorry!
Not even worth a teeeeny tiny v6.0.1? 🙏😁
No worries if not.
Haha sorry no, it'd be a breaking change.
Really? I didn't realize it was a breaking change, I thought that was a simple deprecation for now.
The added console.warn can be surprisingly confusing. I had a coworker new to the area completely disable TSLint locally because we'd bumped to a version and forgotten to disable no-unused-variables and they'd interpreted the warning as TSLint erroring out. It took us a little while to realize why they had so many "fix lint" commits in their PRs!
That makes sense, thanks!
| gharchive/pull-request | 2018-12-13T22:29:35 | 2025-04-01T06:37:13.580449 | {
"authors": [
"JoshuaKGoldberg",
"astorije"
],
"repo": "Microsoft/tslint-microsoft-contrib",
"url": "https://github.com/Microsoft/tslint-microsoft-contrib/pull/674",
"license": "mit",
"license_type": "permissive",
"license_source": "bigquery"
} |
331494336 | Add Compiler option -D_DISABLE_EXTENDED_ALIGNED to disable error C2338
cartographer and mongo-cxx-driver failed with error C2338, it should fix form the error message provides the instructions that add -D_DISABLE_EXTENDED_ALIGNED option to disable this error.
Delete the useless 'static.patch' in mongo-cxx-driver because it did not applied anymore(it did not apply in previous).
Looks good, thanks!
| gharchive/pull-request | 2018-06-12T08:48:53 | 2025-04-01T06:37:13.582126 | {
"authors": [
"LarryIII",
"alexkaratarakis"
],
"repo": "Microsoft/vcpkg",
"url": "https://github.com/Microsoft/vcpkg/pull/3709",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
425904558 | "View Commit in GitHub" doesn't work on Windows \Mac
OS: All
Build Version: 20190327.12
Regression: Not a Regression
Repro Steps:
Make sure you have signed in GitHub.
Right click one function app-->Select "Configure Deployment Source..." --> Select "GitHub".
Right click one deployment and select "View Commit in GitHub".
Expect:
Navigate to the GitHub --> Commit.
Actual:
"View Commit in GitHub" doesn't work.
More Info:
This issue dos not reproduce on Linux.
I tested this on Windows and it doesn't seem to reproduce for me.
| gharchive/issue | 2019-03-27T11:26:42 | 2025-04-01T06:37:13.587293 | {
"authors": [
"nturinski",
"v-dantch"
],
"repo": "Microsoft/vscode-azurefunctions",
"url": "https://github.com/Microsoft/vscode-azurefunctions/issues/1130",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
306697606 | File must be saved before symbols appear in "Go to symbol in file"
If I open a new document and set the language to Python, then symbols will not appear in "Go to symbol in file" unless the file is saved.
Please make it so that we don't have to save the file before we can get this functionality.
@DonJayamanne is this even possible? I'm thinking this isn't possible as ctags won't be able to find the file on disk otherwise to index.
is this even possible?
Yes.
For Go to symbol in file we use jedi, so we should be able to do this.
We use ctags for Go to Symbol in workspace
To help manage our issues and to better communicate what the team plans to work on we are closing issues that we don't plan to work on but would accept a pull request from a volunteer for. To be clear, closing this issue does not mean we won't consider a pull request for this enhancement as outlined in our contributing guide, just that the development team has no plans to work on it themselves.
| gharchive/issue | 2018-03-20T01:44:23 | 2025-04-01T06:37:13.619245 | {
"authors": [
"DonJayamanne",
"brettcannon",
"garyking"
],
"repo": "Microsoft/vscode-python",
"url": "https://github.com/Microsoft/vscode-python/issues/1121",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
341738498 | vscode can't format python file
Environment data
VS Code version: Code 1.25.1 (1dfc5e557209371715f655691b1235b6b26a06be, 2018-07-11T15:43:53.668Z)
Extension version (available under the Extensions sidebar): 2018.6.0
OS and version: Windows_NT x64 10.0.16299
Python version (& distribution if applicable, e.g. Anaconda): python 3.7 x64
Type of virtual environment used (N/A | venv | virtualenv | conda | ...): N/A
Relevant/affected Python packages and their versions:
asn1crypto 0.24.0
autopep8 1.3.5
certifi 2018.4.16
cffi 1.11.5
chardet 3.0.4
click 6.7
crypto 1.4.1
cryptography 2.2.2
cssselect 1.0.3
Django 2.0.7
django-filter 1.1.0
djangorestframework 3.8.2
Flask 1.0.2
idna 2.7
itsdangerous 0.24
Jinja2 2.10
lxml 4.2.3
Markdown 2.6.11
MarkupSafe 1.0
Naked 0.1.31
numpy 1.14.5
pip 10.0.1
protobuf 3.6.0
pycodestyle 2.4.0
pycparser 2.18
PyMySQL 0.9.2
pyquery 1.4.0
pytz 2018.5
PyYAML 3.13
requests 2.19.1
setuptools 39.0.1
shellescape 3.4.1
six 1.11.0
SQLAlchemy 1.2.9
urllib3 1.23
virtualenv 16.0.0
Werkzeug 0.14.1
yapf 0.22.0
Actual behavior
When I format the code with Alt + Shift +F or right click and select "format file", vscode just shows Formatting with autopep8 but doing nothing to my code.
Expected behavior
Should format my code.
Steps to reproduce:
open a .py source code.
add config: "python.formatting.autopep8Args": ["--ignore E501"]
Alt + Shift +F -> doesn't format my code.
remove config: "python.formatting.autopep8Args": ["--ignore E501"]
Alt + Shift +F -> code formatted.
Logs
Output for Python in the Output panel (View→Output, change the drop-down the upper-right of the Output panel to Python)
Formatting with autopep8 failed.
Error: usage: autopep8 [-h] [--version] [-v] [-d] [-i] [--global-config filename]
[--ignore-local-config] [-r] [-j n] [-p n] [-a]
[--experimental] [--exclude globs] [--list-fixes]
[--ignore errors] [--select errors] [--max-line-length n]
[--line-range line line] [--hang-closing]
[files [files ...]]
autopep8: error: unrecognized arguments: d:\test.py
Output from Console under the Developer Tools panel (toggle Developer Tools on under Help)
console.ts:136 [Extension Host] ApplicationInsights:Sender
[Object]
t.log @ console.ts:136
t._logExtensionHostMessage @ extensionHost.ts:393
(anonymous) @ extensionHost.ts:210
emitTwo @ events.js:106
emit @ events.js:194
process.nextTick @ internal/child_process.js:766
_combinedTickCallback @ internal/process/next_tick.js:73
_tickCallback @ internal/process/next_tick.js:104
Please change ["--ignore E510"] to ["--ignore", "E510"] in your configuration. Basically your error about the file's path not being acceptable is because the way you wrote out the arguments trips up autopep8. If I'm wrong then let us know and we can re-open the issue.
@brettcannon
Changing the configuration doesn't work.
But I notice a new warning in the output panel of vscode.
It may help.
Diagnostic Code: InvalidEnvironmentPathVariableDiagnostic, Message: The environment variable 'Path' seems to have some paths containing characters (';', '"' or ';;'). The existence of such characters are known to have caused the Python extension to not load. If the extension fails to load please modify your paths to remove these characters.
Starting the classic analysis engine.
My PATH is
C:\Users\rick>echo %path%
C:\Windows\system32;C:\Windows;C:\Windows\System32\Wbem;C:\Windows\System32\WindowsPowerShell\v1.0\;C:\Windows\System32\OpenSSH\;C:\Program Files (x86)\NVIDIA Corporation\PhysX\Common;C:\Program Files\Git\cmd;C:\Program Files\nodejs\;C:\Program Files\Microsoft VS Code\bin;C:\Users\rick\AppData\Local\Programs\Python\Python37\Scripts\;C:\Users\rick\AppData\Local\Programs\Python\Python37\;C:\Users\rick\AppData\Local\Microsoft\WindowsApps;;C:\Program Files\Microsoft VS Code\bin;C:\Users\rick\AppData\Roaming\npm
| gharchive/issue | 2018-07-17T02:03:07 | 2025-04-01T06:37:13.632356 | {
"authors": [
"brettcannon",
"ladyrick"
],
"repo": "Microsoft/vscode-python",
"url": "https://github.com/Microsoft/vscode-python/issues/2168",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
306130021 | Change the command used to install pip on AppVeyor to avoid installation errors
Fixes #1107
This pull request:
[x] Has a title summarizes what is changing
[x] Includes a news entry file
[x] Has unit tests & code coverage is not adversely affected (within reason)
[x] Works on all actively maintained versions of Python (e.g. Python 2.7 & the latest Python 3 release)
[x] Works on Windows 10, macOS, and Linux (e.g. considered file system case-sensitivity)
FYI you should always just use python -m pip and not the pip command. It not only is xplat but it also disambiguates which Python interpreter you're about to affect.
| gharchive/pull-request | 2018-03-17T05:30:29 | 2025-04-01T06:37:13.636308 | {
"authors": [
"DonJayamanne",
"brettcannon"
],
"repo": "Microsoft/vscode-python",
"url": "https://github.com/Microsoft/vscode-python/pull/1108",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
173123323 | Provide clipboard access through the VSCode API
VSCode Version: 1.4.0
OS Version: Windows 7
Feature Request:
It would be great to get access to the clipboard in the VSCode API. My use case is very simple, I want to be able to put specific text into the clipboard when a button is pressed (relative path to a file). I'm sure other extensions would benefit from clipboard access access as well.
@jrieken fyi
closing as duplicate of https://github.com/Microsoft/vscode/issues/217. Note that we don't have plan to expose the clipboard in the API but recommend the use of npm modules that do this, like: https://github.com/xavi-/node-copy-paste
| gharchive/issue | 2016-08-25T06:25:07 | 2025-04-01T06:37:13.638812 | {
"authors": [
"RoscoP",
"bpasero",
"jrieken"
],
"repo": "Microsoft/vscode",
"url": "https://github.com/Microsoft/vscode/issues/10933",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
125684496 | VSCode sometimes thinks tsconfig.json is missing.
Possibly related to:
https://github.com/Microsoft/TypeScript/issues/6337
https://github.com/Microsoft/vscode/issues/1032
Summary
tsconfig.json sometimes does not show up in VSCode's file list, nor does Typescript intellisense acknowledge it, despite its existence on disk. This behavior persists between VSCode restarts.
Behavior details
Begin on branchA with no changes. branchB has similar files to branchA including tsconfig.json, which is identical between both branches.
git checkout branchB // Checkout successful
git status // shows no files changed.
VSCode gives me arcane errors about needing to specify the module flag for Typescript. tsconfig.json does not show up in the list of files in VSCode.
ls // shows tsconfig.json
Opening a file explorer also shows tsconfig.json.
I restart VSCode, tsconfig.json is still not listed in VSCode despite existing on the file system.
I create a new tsconfig.json in VSCode. It lets me do it with no errors. tsconfig.json now finally shows up in VSCode. It's a blank file. Then:
git status
On branch branchB
Your branch is up-to-date with 'origin/branchB'.
Changes not staged for commit:
(use "git add <file>..." to update what will be committed)
(use "git checkout -- <file>..." to discard changes in working directory)
modified: tsconfig.json
no changes added to commit (use "git add" and/or "git commit -a")
Weird, because tsconfig.json was unchanged before. Let me reset to the head again.
git reset --hard HEAD
HEAD is now at 765b7b9 Some commit message
VSCode now updates tsconfig.json from disk properly and it's in the file list again.
git status
On branch branchB
Your branch is up-to-date with 'origin/branchB'.
nothing to commit, working directory clean
OK, this is where I wanted to be in the first place.
Just for good measure, I type some changes to let VSCode's Typescript intellisense do its thing. I delete the changes after a few seconds, and reassure myself with the following again:
git status
On branch branchB
Your branch is up-to-date with 'origin/branchB'.
nothing to commit, working directory clean
tsconfig.json is still there. Good.
git checkout branchA
Switched to branch 'branchA'
Your branch is up-to-date with 'origin/branchA'.
tsconfig.json is there. Looks fine.
git checkout branchB
Switched to branch 'branchB'
Your branch is up-to-date with 'origin/branchB'.
tsconfig.json is again perfectly fine, as it should have always been. But I definitely witnessed the behavior detailed above.
Final thoughts
I can understand tsserver.js locking files causing problems, but not tsconfig.json completely disappearing from VSCode's file list even after restarts when it is clearly there on the disk itself. It's like VSCode is completely ignoring tsconfig.json sometimes.
@bpasero I had a hard time fully understanding the issue here. Gave it to you since it seems that the problem is that tsconfig.json is not shown in the explorer.
@Penryn in those cases where tsconfig.json is not showing up, does a refresh of the explorer help?
@bpasero: I actually never tried refreshing VSCode's file list, but I did restart VSCode which still did not list tsconfig.json. To be clear, tsconfig.json was present on the file system this entire time.
If this happens again I'll try refreshing the file list instead.
@Penryn no a restart is fine too. This is weird, even if the file is locked by some other process, asking its parent folder for its children should return it. Any ideas what could go on? Something in the console (from Help menu) maybe?
| gharchive/issue | 2016-01-08T20:12:10 | 2025-04-01T06:37:13.648235 | {
"authors": [
"Penryn",
"bpasero",
"joaomoreno"
],
"repo": "Microsoft/vscode",
"url": "https://github.com/Microsoft/vscode/issues/1885",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
221750716 | Emacs-Style "Electric Indent"
VSCode Version:
OS Version: Windows 10 Creators Update
Steps to Reproduce:
Have a block of text
Press "Tab"
Expected behaviour:
The content jumps to the auto-level of indentation.
In Emacs, this occurs wherever you press "tab", so,
<div>
<p>Foo</p>
</div>
a tab on the line with the <p> will properly indent the line, no matter where the mark is.
Since "tab" has other behaviour in VSCode, I propose that it would do this only when no other behaviour would be executed.
the emacs function executed when you press TAB is indent-for-tab-command. The behavior is described here: https://www.gnu.org/software/emacs/manual/html_node/elisp/Mode_002dSpecific-Indent.html
Right now if you are working on an empty line, pressing tab will auto indent. While if the indentation is already changed, pressing tab will not try to fix it. But in this case, if you set autoIndent as true and type/enter, the indent of this line will get adjusted.
That's precisely the suggestion/request - pressing tab on a line (of course, with auto indent on, and possibly another flag too like editor.electricIndent) would re indent the line :-)
Sent from my Huawei Nexus 6P using FastHub
I'd also love this feature, this is the main thing that I'm missing from VSCode. Most text editors have support for it. This could technically be implemented by just having a new "Reindent line" shortcut that works like "Reindent lines" but would re-indent the current line instead of the selection/whole file right?
@rFlex I belive that "Reindent Selected Lines" will indent the current line, when there is no region selected.
So you may have quite decent indent-on-tab system if you add this to keybindings.json:
{
"key": "tab",
"command": "editor.action.reindentselectedlines",
"when": "editorTextFocus"
}
It works well in JavaScript and CSS files. In HTML it fails to indent attributes broken into multiple lines, so in that case you may need to shift+alt+f (this command is indenting the code properly for some reason ...). Consider this example,
<span>example</span> is being properly indented with tab, while class="some-class"> fails to indent:
(I know that it's over a year after there was anything posted here, but maybe someone will find this useful anyway)
What is sometimes overlooked is that Emacs is really awesome at indenting when it comes to languages where indent changes the syntax.
For languages where the ident is not part of the syntax this task is easier to accomplish than for those where ident matters. If you take Python or Coffeescript, the ident changes the syntax.
Emacs feature is smart enough to know that. Pressing repeatedly jumps to all valid positions of that line. This is a very cool feature and besides Emacs' awesome Killring, the most missed feature in Vscode for me.
So this feature request gets my +1 and another +1 for proper Killring support (all extensions I tried are not as good as Emacs).
| gharchive/issue | 2017-04-14T06:26:24 | 2025-04-01T06:37:13.658071 | {
"authors": [
"martinrode",
"nahuel",
"phardyn",
"rFlex",
"rebornix",
"tigerhawkvok"
],
"repo": "Microsoft/vscode",
"url": "https://github.com/Microsoft/vscode/issues/24744",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
230570020 | Holding Alt-(key) on Sub-Menu item doesn't trigger that Command.
VSCode Version: 1.12.2
OS Version: Windows 10
Steps to Reproduce:
Open Visual Studio Code
Hold Down the 'Alt' key
Press the 'F' (while still holding down the 'Alt' key)
The File menu will appear at the top right of your screen
Press 'N' (while still holding down the 'Alt' key)
Nothing happens.
Expected Behaviour: A new file is added to your solution.
The first item in the File menu is 'New File' with the letter N underlined. The underline indicates that it can be activated by pressing the key. The standard approach is to also allow the 'Alt' key to be pressed at the same time.
This should work if the 'Alt' key is still being pressed as this is the behaviour of most editors in windows e.g. Notepad, Word, Visual Studio...
I hope to see this fixed. 20yrs of muscle memory in VS has me quickly hitting Alt+F, S. Unfortunately in VSCode if I don't release Alt before S is pressed it doesn't accept the sequence. VS (on windows) supports S with or without ALT key still being pressed.
I am having a similar issue though it seems intermittent. Alt+key not working for any of the menus. Didn't work one day. Worked for me yesterday. Not working today.
The same issue as in the comment (as a closed issue) above. Alt+key randomly stopped working for me today. Restarting VSCode doesn't help. Haven't tried restarting the program yet.
Seems to be related to https://github.com/Microsoft/vscode/issues/10525 (which is also closed for some reason).
@WereDev, I figured it out: it was caused by Scroll Lock: https://github.com/Microsoft/vscode/issues/19011
@monk-time , confirmed! I tested it out and it replicated my issue beautifully! I guess I just had a couple fat finger days...
Same here.
While it still works in other applications in VS Code it is not working.
Restart of Windows helped :)
At least Alt+F opens the file menu for you. Doesn't for me on Linux.
Apparently they decided this just wasn't an issue: #10525
Duplicate of #52935
| gharchive/issue | 2017-05-23T02:27:39 | 2025-04-01T06:37:13.664890 | {
"authors": [
"JoelViney",
"WereDev",
"bpasero",
"et304383",
"fider",
"lucasshawn",
"monk-time"
],
"repo": "Microsoft/vscode",
"url": "https://github.com/Microsoft/vscode/issues/27106",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
233274589 | Find/Replace Usability: Enter and Ctrl+Enter shortcuts
VSCode Version: 1.12.2
OS Version: Windows 10
In addition to editor.action.replaceAll and editor.action.replaceOne, there exists other shortcuts that work in the Find and Replace dialog:
CTRL+ENTER is a shortcut for editor.action.replaceAll
ENTER is similar to editor.action.replaceOne
These are amazingly helpful. A lot easier than using the defaults (CTRL+ALT+ENTER and CTRL+SHIFT+1 respectively)
Unfortunately, these shortcuts only work when the Replace text box has focus, not the Find text box.
Is it possible to wire these up to the Find text box so that no matter which input you fill in last, you can execute the a replaceAll or replaceOne as quickly as possible?
FYI: I can override the editor.action.replaceAll shortcut with CTRL+ENTER, and that gets me halfway there, but attempting to override editor.action.replaceOne with ENTER obviously isn't going to work. Interestingly enough, it does let me do it, and it does work (with major side effects obviously), but it executes replaceOne twice when the Replace text box has focus, I assume because of the configurable shortcut and the internal shortcut.
Ah, perhaps it's not a good idea to modify what ENTER does when Focus has selected, but at least wire up CTRL+ENTER inside of _onFindInputKeyDown.
https://github.com/Microsoft/vscode/blob/0346c023f6bb057df1a9ff793cebb5cc5810ac56/src/vs/editor/contrib/find/browser/findWidget.ts#L563
I'd be happy to submit a PR for this.
@langdonx thanks for your help on this! We actually support to replace all when the Find Widget is focused.
Below is current state of key shortcuts of Find Widget.
InputBox
Action
macOS
Windows
Linux
Find
Next Match Result
Enter
Enter
Enter
Find
Previous Match Result
Shift Enter
Shift Enter
Shift Enter
Find
Select All Matches
Option Enter
Alt Enter
Alt Enter
Replace
Replace
Enter
Enter
Enter
Replace
Replace All
Cmd Enter
Ctrl Enter
Ctrl Enter
Find/Replace
Replace All
Cmd Option Enter
Ctrl Alt Enter
Ctrl AltEnter
@rebornix This is a bug fix.
Ctrl Enter only works when the replace text box has focus. If your cursor is on the find text box, the shortcuts do nothing.
@rebonix Sorry, misread the table. This is more of a UX improvement.
Having shortcuts be input-specific, seems like bad UX. Is there a reason Ctrl Enter.
@langdonx one more thing to add is cmd+option+enter works if the focus is in the editor or find widget. Which means is a generic command, it has nothing to do with Input boxes.
Having shortcuts be input-specific, seems like bad UX
What do you mean by this? Enter is input-specific as it means different in different input boxes.
It appears that VS Code's intention with findWidget is for it to be an intrinsic part of the editor (not exclusive to any single file), and less of a traditional dialog. I believe not understanding this is the source of my frustration.
In Notepad, WebStorm, Visual Studio, and probably other editors, for example, I've always used Alt A to replace all from either text box. Without that shortcut, I was looking for something quick to rely on to cover all situations.
With the findWidget being an intrinsic part of the editor, it has to play nicely with the editor shortcut keys, of which, there are many, so Ctrl Alt Enter certainly makes sense as the default.
It seems like I can achieve what I'm looking for by remapping editor.action.replaceAll to Ctrl Enter. With this configuration, the Insert Line Below command (which I don't often use) will still work when the findWidget isn't open.
| gharchive/issue | 2017-06-02T19:16:58 | 2025-04-01T06:37:13.682921 | {
"authors": [
"langdonx",
"rebornix"
],
"repo": "Microsoft/vscode",
"url": "https://github.com/Microsoft/vscode/issues/27945",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
249446992 | Feature Request: Open next file with problems
The problems panel is incredibly helpful, especially when used with linters on large projects during refactors. However, it would be really helpful if there was a command that would allow you to open the next unopened file that has problems. This would allow users to quickly track down and remove errors during large code shifts.
Not sure how much helpful it is to have it in core. Probably an extension can contribute this command.
It's not actually possible to do this in an extension, I tried. The MarkerService is not accessible through any API.
@pzuraq Yes you are right, API does not support getting all diagnostics / markers. Reopening it.
Duplicate of #32159
| gharchive/issue | 2017-08-10T19:11:03 | 2025-04-01T06:37:13.685166 | {
"authors": [
"pzuraq",
"sandy081"
],
"repo": "Microsoft/vscode",
"url": "https://github.com/Microsoft/vscode/issues/32249",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
252715113 | cmd+shift+f (Find in Files) should not focus the editor
The current behavior is that pressing cmd+shift+f toggles between the editor and the side panel, but this drives me crazy as I need to check my focus before pressing (or not pressing) that combo when doing a lot of hunting around.
The number of times I've accidentally typed my search into a random file instead of the search field is painful. Please make cmd+shift+f always focus the search field, or at least allow for such an option.
This issue was originally closed, here: https://github.com/Microsoft/vscode/issues/22814#issuecomment-324732891
I'm not certain why this isn't default behavior as it is common in other editors.
Thanks!
It should ideally also select-all in the search field in addition to focusing it when there is already text there.
A related issue: It doesn't select-all in the search field when pressing cmd+shift+f if the field is already focused. It can be quite disorienting.
It is possible to always focus the search field with a binding like this:
{
"key": "shift+cmd+f",
"command": "workbench.action.findInFiles"
}
I guess it's not the default just because I don't want to change it this late. But
It doesn't select-all in the search field when pressing cmd+shift+f if the field is already focused
that's a bug
I'll take the little victory! Thanks Rob.
Thank you! Cheers to no more git diff surprises and no extra cmd+a keystrokes :)
| gharchive/issue | 2017-08-24T20:02:50 | 2025-04-01T06:37:13.689471 | {
"authors": [
"jimmy-guo",
"mherodev",
"roblourens"
],
"repo": "Microsoft/vscode",
"url": "https://github.com/Microsoft/vscode/issues/33108",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
259689762 | workspace.openTextDocument works incorectly when passing { content?: string; } as argument
calling API method workspace.openTextDocument({content: 'some content'}) or workspace.openTextDocument({}) throws exception, but should work correctly.
The problem lines are in file src\vs\workbench\api\node\extHost.api.impl.ts:
let options = uriOrFileNameOrOptions as { language?: string; content?: string; };
if (!options || typeof options.language === 'string') {
uriPromise = extHostDocuments.createDocumentData(options);
}
should be something like this:
let options = uriOrFileNameOrOptions as { language?: string; content?: string; };
if (options) {
uriPromise = extHostDocuments.createDocumentData(options);
}
Good catch!
| gharchive/issue | 2017-09-22T03:02:01 | 2025-04-01T06:37:13.691381 | {
"authors": [
"bpasero",
"fi11er"
],
"repo": "Microsoft/vscode",
"url": "https://github.com/Microsoft/vscode/issues/34796",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
264465300 | title doesn't change when switching open files via tabs
VSCode Version: VSCode 1.18.0-Insider
OS Version: Win7 64bit
Steps to Reproduce:
1.Open VSCode, and open a few files.
2.switch open files via tab, the title doesn't change to the right file path.
3.if I change the size of the VSCode window, or minimize then maximum, the title is refreshed to the currently open file path.
my settings:
{
"workbench.activityBar.visible": false,
"files.autoGuessEncoding": true,
"editor.minimap.enabled": false,
"workbench.colorTheme": "Visual Studio Light",
"editor.rulers": [
120
],
"editor.renderLineHighlight": "all",
"editor.renderWhitespace": "all",
"window.title": "${dirty}${activeEditorLong}${separator}${rootName}${separator}${appName}",
"workbench.colorCustomizations": {
"tab.border": "#00F",
"tab.activeBackground": "#F99BC5",
"tab.inactiveBackground": "#E7E7EF"
}
}
Reproduces without extensions: Yes
This issue has already been reported. Please use the search function next time before reporting an issue and/or this time if you are curious about the progress of the issue.
| gharchive/issue | 2017-10-11T05:55:49 | 2025-04-01T06:37:13.694716 | {
"authors": [
"LiPengfei19820619",
"bpasero"
],
"repo": "Microsoft/vscode",
"url": "https://github.com/Microsoft/vscode/issues/36018",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
118088838 | Allow detachable multiple windows on the same workspace
Allow opening additional windows (BrowserWIndow) to be able to work on multi-monitor setups.
See also:
Detachable split-view
Ability to have a "folder" open over multiple windows (like Visual studio)
Multiple windows, same folder
+1.
It is quite limiting, and very annoying, that when opening the same workspace (folder, or file) in a new window, VSCode instead jumps to the existing window.
https://github.com/Microsoft/vscode/issues/2686
@bpasero I consider this different from #2686. This is about having multiple OS windows with one VSCode instance. One sidebar, one command bar and being able to drag editors out from one OS window into another OS window.
Do you think the work done for #2686 will also allow this?
+1
+1
+1. Even only a detachable PROBLEMS OUTPUT DEBUG CONSOLE TERMINAL section would make a huge difference. Test results on a different screen would greatly increase productivity.
This is the single issue standing in the way of finally migrating away from Eclipse for me.
The progress on this is underwhelming since 2015.
Opening a new window on a work space that is already open seems to be impossible, no matter how I go about it. This goes very much against my intuition and the "principle of least surprise" - why should the application go out its way to prevent a new window from opening if another one already exists?
VS Code Devs : we need this so bad.
Please !
Here is a work around:
Wrap your project's folder in another, empty folder, like this:
├─ New empty folder
└─── Project root
Now you can open the same project in separate instances of VSCode - just open each folder in VSCode!
This, for me, is also the last reason not to switch to VSCode completely yet.
This looks like a duplicate of https://github.com/Microsoft/vscode/issues/10121. Could y'all go vote there if you get a chance?
This looks like a duplicate of https://github.com/Microsoft/vscode/issues/10121 (and that one has a few more more votes)... could y'all go vote there? :-)
@RoyTinker yeah agree, please cast your vote at https://github.com/Microsoft/vscode/issues/10121
| gharchive/issue | 2015-11-20T17:36:05 | 2025-04-01T06:37:13.703424 | {
"authors": [
"BentOnCoding",
"Robinfr",
"RoyTinker",
"bit2pixel",
"bpasero",
"d4hines",
"dacrouch",
"etoah",
"jeremyd",
"pavlos256",
"stereokai",
"wupaz"
],
"repo": "Microsoft/vscode",
"url": "https://github.com/Microsoft/vscode/issues/367",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
282596627 | main window is blank and no further responses after 1.19
I think my problem is just as the 'main blank is blank' problem listed on FAQ page of official website.
I tried code --disable-gpu code --disable-extensions, but doesn't work.
And I tried vscode insider, well it works fine before I sync the settings using 'sync settings' extension, then the same issue happend. Even if I deteled .vscode-insiders located in C:\Users\username.
Computer: HP 340 G4 | AMD Radeon(TM) R5 M430 | Intel(R) HD Graphics 620
VSCode Version: 1.19
OS Version: windows 10 64bit
Steps to Reproduce:
updated to 1.19
open vscode
blank main page, all the operations get no responses.
Reproduces without extensions: Yes
code --verbose
[15888:1216/181628.909:ERROR:browser_gpu_channel_host_factory.cc(125)] Failed to create channel.
[15888:1216/181628.909:ERROR:browser_gpu_channel_host_factory.cc(125)] Failed to create channel.
[15888:1216/181628.909:ERROR:browser_gpu_channel_host_factory.cc(125)] Failed to create channel.
[15888:1216/181628.909:ERROR:browser_gpu_channel_host_factory.cc(125)] Failed to create channel.
[15888:1216/181628.909:ERROR:browser_gpu_channel_host_factory.cc(125)] Failed to create channel.
[15888:1216/181628.909:ERROR:browser_gpu_channel_host_factory.cc(125)] Failed to create channel.
[15888:1216/181629.807:ERROR:browser_gpu_channel_host_factory.cc(125)] Failed to create channel.
[15888:1216/181629.808:ERROR:browser_gpu_channel_host_factory.cc(125)] Failed to create channel.
[15888:1216/181629.808:ERROR:browser_gpu_channel_host_factory.cc(125)] Failed to create channel.
[15888:1216/181629.808:ERROR:browser_gpu_channel_host_factory.cc(125)] Failed to create channel.
[15888:1216/181629.808:ERROR:browser_gpu_channel_host_factory.cc(125)] Failed to create channel.
[15888:1216/181629.808:ERROR:browser_gpu_channel_host_factory.cc(125)] Failed to create channel.
developer tools errors
Do you still see the issue when running code --user-data-dir <path to some empty folder>?
Keeping the blank window open, run code --status and share the results with us
Hello, I have the same issue here.
After I update my vscode around 15:30, 2017/12/16(GMT+8), my vscode open with blank plane and nothing I can select.
I have tried reinstall and it is still the same.
I also using command code --disable-gpu code --disable-extensions, and it is also give me blank page.😢
VSCode version: 1.19
OS version: macOS High Sierra(10.13.2)
my vscode looks like:
But I can open the debugger, it give me the message:
Thanks for that call stack @stvreumi
@pennwang23 Can you check if you see the same?
setCustomTokenColors is coming when trying to set the theme. cc @aeschli
@stvreumi Can you answer the questions in https://github.com/Microsoft/vscode/issues/40329#issuecomment-352167047?
@ramya-rao-a
Hello, I just reinstall the older version and it's fine for me.
I update to the latest version and the bug still happen.
However, using the command you give ('code --user-data-dir `), I can open vscode with no blank!
But the bug still show up after I quit vscode and restart again.
@stvreumi I moved your issue to https://github.com/Microsoft/vscode/issues/40339 because I do not think it is the same as @pennwang23
@pennwang23 does it reproduce if you run code --disable-gpu ?
still doesn't work, code --user-data-dir works for once, but didn't work after restart. @ramya-rao-a
yes, it reproduces @bpasero
thank you guys, I found the problem, there are some errors in my user settings of vscode. very simple syntax errors. @ramya-rao-a @bpasero
@pennwang23 Can you give us more on information on what errors you found in your user settings?
@stvreumi
However, using the command you give (code --user-data-dir ), I can open vscode with no blank!
But the bug still show up after I quit vscode and restart again.
Looks like there is an issue with your user settings file just like @pennwang23
the link in my last post is my settings, could you open it ? unfortunately I can't see where the issue is @ramya-rao-a
@pennwang23 Yes, I can open that link. I tried all the settings from there and the culprit is editor.tokenColorCustomizations.
The default value for this setting is {}
But in your settings, the value for this is null which is causing the issue.
I can repro the blank window issue by setting editor.tokenColorCustomizations to null in my settings.
@stvreumi Can you check if you have editor.tokenColorCustomizations set to null in your settings file (you can find it at ~/Library/Application Support/Code/User/settings.json). If yes, then set it to {} instead and then try opening VS Code.
Hello @ramya-rao-a
I really set editor.tokenColorCustomizations to null, in the line 13 of my setting.json
Here is my setting.json:
https://gist.github.com/stvreumi/3d7a3ee4bb474b66f3db9130638929bb
After I change it to {} and update to the latest version, I can see my workspace normally.
I think I add this setting because of the extension: vscode-icons.
BTW I check the console in the developer tools, there is no error appear.
Thank you @pennwang23
Thanks for confirming @stvreumi!
cc @bpasero @aeschli
Had the same issues (also using vscode-icons) and confirming that editor.tokenColorCustomizations was indeed set to null after the latest update
duplicate of #40339
you're welcome. @stvreumi
thanks for your work. @ramya-rao-a
| gharchive/issue | 2017-12-16T03:32:53 | 2025-04-01T06:37:13.725950 | {
"authors": [
"aeschli",
"bpasero",
"pennwang23",
"pynklu",
"ramya-rao-a",
"stvreumi"
],
"repo": "Microsoft/vscode",
"url": "https://github.com/Microsoft/vscode/issues/40329",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
284619722 | Clicking on "Install" button of an already installed extension doesn't open extension tab
VSCode Version: Code 1.19.1 (0759f77bb8d86658bc935a10a64f6182c5a1eeba, 2017-12-19T09:35:14.671Z)
OS Version: Darwin x64 17.2.0
Extensions:
Extension
Author (truncated)
Version
vscode-toggle-quotes
Bri
0.2.0
bracket-pair-colorizer
Coe
0.10.14
html-css-class-completion
Zig
1.14.0
npm-intellisense
chr
1.3.0
path-intellisense
chr
1.4.2
vscode-eslint
dba
1.4.3
auto-rename-tag
for
0.0.15
sublime-keybindings
ms-
3.0.3
debugger-for-chrome
msj
4.0.0
vscode-versionlens
pfl
0.19.1
vscode-import-cost
wix
2.6.0
Steps to Reproduce:
Install any VSCode extension. (Example: Markdown PDF)
Go to the installed extension's marketplace url: https://marketplace.visualstudio.com/items?itemName=yzane.markdown-pdf
Click on "Install" button
Expected:
VSCode opens "Extension" tab and loads the extension page.
Actual:
VSCode app gains focus, but nothing else happens (no tabs open).
This is confusing, especially if you have many extensions and are unsure whether you already have an extension installed.
Also there's another poor UI: If I don't have extension installed, and I install it, but not by clicking on the "Instal"/"Cancel" new top bar "notification", but from the button inside the newly open "Extension" tab, the top bar notification doesn't get updated/dismissed. The state of extension (installed/not installed) is not propagated/synced globally)
Reproduces without extensions: Yes
Fixed by https://github.com/Microsoft/vscode/pull/42071
| gharchive/issue | 2017-12-26T22:55:06 | 2025-04-01T06:37:13.735429 | {
"authors": [
"joaomoreno",
"vfonic"
],
"repo": "Microsoft/vscode",
"url": "https://github.com/Microsoft/vscode/issues/40848",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
295304131 | Image zooming: pinch to zoom gestures are inverted
On a Surface Book with Windows 10, the pinch to zoom using the trackpad is inverted:
expanding fingers will zoom out
doing the opposite gesture will zoom in
I expect the other way around.
This happens both with stable 1.20 and latest insider:
Version 1.20.0-insider
Commit f9398645d1d624538708eeebf96e6fdeb546f5d7
Date 2018-02-06T15:20:43.871Z
Shell 1.7.9
Renderer 58.0.3029.110
Node 7.9.0
Architecture x64
Windows 10 x64
update Same problem when using ctrl + two fingers move.
Yes pinching also seems inverted on mac os. The zooming when holding ctrl though matches mac os: up to zoom in, down to zoom out. Perhaps this is because of Apple's "natural" scroll direction. I'll take a look at some other windows apps and try to match their behavior
There is also an option for inverting scrolling in Windows 10: I tried changing it but the bug is still there.
Please try this out in the next insiders build. I'm not 100% sure my fix handles all possible OS and input device combinations correctly so please provide feedback if something still doesn't feel right
@mjbvz
for win7x64, it's fixed now. Thanks!
| gharchive/issue | 2018-02-07T21:45:20 | 2025-04-01T06:37:13.739137 | {
"authors": [
"mjbvz",
"user3323",
"warpdesign"
],
"repo": "Microsoft/vscode",
"url": "https://github.com/Microsoft/vscode/issues/43138",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
305630980 | Highlight matching parentheses/brackets not working in latex math mode
Pretty much said it in the title : matching parentheses/brackets are not highlighted when editing a .tex document when inside one of Latex's math modes : $ $, \begin{equation}, etc....
Is there any setting that I missed ?
VSCode has no built-in support for latex, but I verified that when opening a latex in Plain Text mode the bracket matching works.
I guess you are using an extension for latex , and likly that extension has changed regarding the bracket matching.
Can you file the issue against the latex extension that you are using? Also, I'd suggest adding a full sample code to illustrate the problem.
The Latex extension is "LaTeX-Workshop" by James Yu; it seems that this issue has already been mentioned several time on the extension dedicated github (see here and here) but to quote the author (see second link) it seems to him that :
this is more a textmate/vscode upstream problem which disables brackets pairing in latex.math environments. Sorry it is not a problem that can be fixed from my side. I would recommend you submit an issue on vscode's repo.
Any comments ?
Ok, I see. The problem is that the Latex grammar from 'LaTeX-Workshop' uses the scope string.other.math.block.environment.latex inside the \begin blocks.
Surrounding pairs are disabled inside strings.
It's a bit strange to use that scope for a whole block, so I'd suggest to file an issue against 'LaTeX-Workshop' to improve that.
The request to allow or configure bracket matching inside strings is https://github.com/Microsoft/vscode/issues/12388
| gharchive/issue | 2018-03-15T16:40:41 | 2025-04-01T06:37:13.743709 | {
"authors": [
"aeschli",
"s-ilic"
],
"repo": "Microsoft/vscode",
"url": "https://github.com/Microsoft/vscode/issues/45890",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
311644222 | suggestion: built-in php instellisense
nice to have a built-in php instellisense rather than using an extension.
There are very basic suggestions provided out of the box, but I recommend using an extension.
atleast i tried ^_^ thanks
| gharchive/issue | 2018-04-05T14:53:00 | 2025-04-01T06:37:13.745085 | {
"authors": [
"leetskil",
"roblourens"
],
"repo": "Microsoft/vscode",
"url": "https://github.com/Microsoft/vscode/issues/47243",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
314690166 | not able to create a file by right clicking on folder.
Issue Type: Bug
right click on the folder
select create new file
VS Code version: Code - Insiders 1.23.0-insider (dfeffb55b62cc2c1dde5cc1e25e98632eeb7744b, 2018-04-16T05:17:45.591Z)
OS version: Windows_NT x64 10.0.15063
System Info
Item
Value
CPUs
Intel(R) Core(TM) i5-4590 CPU @ 3.30GHz (4 x 3292)
GPU Status
2d_canvas: enabledchecker_imaging: disabled_offflash_3d: enabledflash_stage3d: enabledflash_stage3d_baseline: enabledgpu_compositing: enabledmultiple_raster_threads: enabled_onnative_gpu_memory_buffers: disabled_softwarerasterization: enabledvideo_decode: enabledvideo_encode: enabledwebgl: enabledwebgl2: enabled
Memory (System)
7.96GB (2.46GB free)
Process Argv
C:\Program Files\Microsoft VS Code Insiders\Code - Insiders.exe
Screen Reader
no
VM
0%
Extensions (10)
Extension
Author (truncated)
Version
ng-template
Ang
0.1.9
jshint
dba
0.10.18
vscode-eslint
dba
1.4.8
gitlens
eam
8.2.1
EditorConfig
Edi
0.12.1
tslint
eg2
1.0.28
python
ms-
2018.3.1
PowerShell
ms-
1.6.0
typescript-javascript-grammar
ms-
0.0.39
vscode-react-native
vsm
0.6.6
The same bug also happens when using the create new file in the project bar under the project explorer.
Same for New Folder
Issue still exists, Win10 latest insiders version.
Duplicate of #47937.
| gharchive/issue | 2018-04-16T14:48:23 | 2025-04-01T06:37:13.755414 | {
"authors": [
"LouieK22",
"carboneater",
"eamodio",
"perlun",
"sanket-bhalerao"
],
"repo": "Microsoft/vscode",
"url": "https://github.com/Microsoft/vscode/issues/47972",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
332283880 | [Accessibility]A11y_VSCode_VSCAdditionalTasks_Screenreader: Jaws does not read the control type of check box.
Environment Details:
VSCode Version : 1.24.1
OS Version : Win10
Additional Details:
JAWS 18.0.5038 ,MAS1.3.1
Repro Steps:
1.On Visual Studio Code Start Menu Folder screen,click on "Next" button.
2.Navigate to check box present on the screen.
3.Check using tab key whether Jaws reads the control types of check box or not
Actual:
Jaws does not read the control type of check box. Its reads only name & checked & unchecked.
Expected:
Jaws should read the control types of checkbox.
Recommendations:
Refer below link which is repository of bug fixes code snippets:
https://microsoft.sharepoint.com/teams/msenable/mas/pages/browse-fixes.aspx
User Impact:
The user will not able to get information about control.
MAS Reference:
https://microsoft.sharepoint.com/:w:/r/teams/msenable/_layouts/15/WopiFrame.aspx?sourcedoc={8492c4eb-c179-40ae-8777-cd044ed725a2}
Does this issue occur when all extensions are disabled?: Yes
Attachment for Reference:
[Accessibility]A11y_VSCode_VSCAdditionalTasks_Screenreader.pptx
@AccessibilityTestingTeam-TCS This needs to go into https://github.com/jrsoftware/issrc
| gharchive/issue | 2018-06-14T07:34:14 | 2025-04-01T06:37:13.761711 | {
"authors": [
"AccessibilityTestingTeam-TCS",
"joaomoreno"
],
"repo": "Microsoft/vscode",
"url": "https://github.com/Microsoft/vscode/issues/51855",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
337170354 | Configure Task Runner Missing VS Code
I can't seem to find the Configure Task Runner option in the search bar in VS Code. Can anyone please help me be able to see this option? Thanks!
The command is named Task: Configure Task. I have updated the doc.
| gharchive/issue | 2018-06-30T00:24:03 | 2025-04-01T06:37:13.763433 | {
"authors": [
"ashucha",
"dbaeumer"
],
"repo": "Microsoft/vscode",
"url": "https://github.com/Microsoft/vscode/issues/53363",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
353600457 | [Extensibility] TextEditor instance got disposed if another document is opened on top of it
Issue Type: Bug
I have an extension that creates two Untitiled-* documents then shows them side by side at column one and column two. I saved the two TextEditor instances so later the extension can update the contents of the two documents, instead of creating two new documents again.
when the two documents are showing, if some other document window is opened in one of the column that has the focus
Now the TextEditor instance in that column does not work any more. Even the extension still can call the editor method to replace content. The change isn't reflected in VS Code UI. The "Untitled-1" tab was expected to show code for the namespace, but previous code get stuck there.
Here's the debugger showing _disposed: true for that instance of TextEditor
The relevant code is at https://github.com/icsharpcode/ilspy-vscode/blob/master/vscode-extension/src/extension.ts#L121-L158
VS Code version: Code 1.26.1 (493869ee8e8a846b0855873886fc79d480d342de, 2018-08-16T18:38:57.434Z)
OS version: Windows_NT x64 10.0.17741
System Info
Item
Value
CPUs
Intel(R) Core(TM) i7-4790 CPU @ 3.60GHz (8 x 3592)
GPU Status
2d_canvas: enabledchecker_imaging: disabled_offflash_3d: enabledflash_stage3d: enabledflash_stage3d_baseline: enabledgpu_compositing: enabledmultiple_raster_threads: enabled_onnative_gpu_memory_buffers: disabled_softwarerasterization: unavailable_softwarevideo_decode: enabledvideo_encode: enabledwebgl: enabledwebgl2: enabled
Memory (System)
31.76GB (15.96GB free)
Process Argv
C:\Program Files\Microsoft VS Code\Code.exe
Screen Reader
no
VM
0%
Extensions (13)
Extension
Author (truncated)
Version
vscode-svgviewer
css
1.4.4
EditorConfig
Edi
0.12.4
tslint
eg2
1.0.36
vscode-firefox-debug
hbe
1.5.0
ilspy-vscode
ics
0.7.0
azurecli
ms-
0.4.2
csharp
ms-
1.15.2
Go
ms-
0.6.88
PowerShell
ms-
1.8.3
debugger-for-chrome
msj
4.8.2
debugger-for-edge
msj
1.0.2
vscode-docker
Pet
0.1.0
cordova-tools
vsm
1.4.0
That is the design of editors in the API. A TextEditor is text document that is currently visible which in return means the editor is being disposed when not visible anymore. You can hold on to a document and open it again as (new) text editor
| gharchive/issue | 2018-08-24T00:24:25 | 2025-04-01T06:37:13.777904 | {
"authors": [
"jeremymeng",
"jrieken"
],
"repo": "Microsoft/vscode",
"url": "https://github.com/Microsoft/vscode/issues/57119",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
353914668 | Git repo in %appdata%..\Roaming\code\user appear in project SCM
Issue Type: Bug
If one is creating a git repo in %appdata%\..\Roaming\code\user which adds VSCode settings to version control and subsequent also creates some updates on that repo, like changing settings, this is reflected as project SCM changes.
This is not expected since the project SCM should only care about project related repo (if exists).
VS Code version: Code 1.26.1 (493869ee8e8a846b0855873886fc79d480d342de, 2018-08-16T18:38:57.434Z)
OS version: Windows_NT x64 10.0.18214
System Info
Item
Value
CPUs
Intel(R) Core(TM) i3 CPU M 370 @ 2.40GHz (4 x 2394)
GPU Status
2d_canvas: enabledchecker_imaging: disabled_offflash_3d: enabledflash_stage3d: enabledflash_stage3d_baseline: enabledgpu_compositing: enabledmultiple_raster_threads: enabled_onnative_gpu_memory_buffers: disabled_softwarerasterization: unavailable_softwarevideo_decode: enabledvideo_encode: enabledwebgl: enabledwebgl2: unavailable_off
Memory (System)
2.74GB (0.32GB free)
Process Argv
C:\Users\andrei\AppData\Local\Programs\Microsoft VS Code\Code.exe
Screen Reader
no
VM
0%
Extensions (1)
Extension
Author (truncated)
Version
vscode-yaml
red
0.0.14
The repo opens if someone is trying to edit the VSCode settings (and that's of course a file).
So in that sense, you could say that it opens if the settings file get's open.
You can configure Code to ignore certain Git repositories using the git.ignoredRepositories setting.
| gharchive/issue | 2018-08-24T20:19:15 | 2025-04-01T06:37:13.785998 | {
"authors": [
"joaomoreno",
"n3rd4i"
],
"repo": "Microsoft/vscode",
"url": "https://github.com/Microsoft/vscode/issues/57170",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
378563339 | The minimum size of a terminal should be an option in Settings
[(https://github.com/Microsoft/vscode/issues/45678)]
The related issue was closed, but I think it need more work!
Some terminal can be resized to a small size if they are not important. In some case, showing multiple terminals is just to notify the user whether they are running. Currently the minimum size of terminal is fixed and too large.
I suggest that the minimum size of a terminal should be an option in Settings.
@Tumiz do you have a better suggestion for the minimum size (currently 250px)? #45678 was purely to prevent a broken layout in the terminal space and having terminals get hidden.
Some terminal is OK as small like this. In this case, I only need a few lines to show a program's status. So partly hidden is also acceptable.
It will be better if the minimum size can be configured in Settings.
| gharchive/issue | 2018-11-08T03:24:55 | 2025-04-01T06:37:13.789729 | {
"authors": [
"Tumiz",
"Tyriar"
],
"repo": "Microsoft/vscode",
"url": "https://github.com/Microsoft/vscode/issues/62759",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
154913772 | Menu item 'Check For Updates' should be disabled when there's no frontmost window
VSCode Version: 1.1.0
OS Version: Mac OS X 10.11.4
Steps to Reproduce:
Launch VSCode
Close all windows
Go to menu 'Code' and select 'Check For Updates'
Nothing, no alert, no message. Like the menu item was never selected.
This is OK. The application will still be updated if there is an update in this situation.
| gharchive/issue | 2016-05-15T15:19:50 | 2025-04-01T06:37:13.792393 | {
"authors": [
"joaomoreno",
"peter-lau"
],
"repo": "Microsoft/vscode",
"url": "https://github.com/Microsoft/vscode/issues/6383",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
156529882 | Closing HTML tags with hyphens are not deindented properly
VSCode Version: 1.1.1
OS Version: Windows 7 x64
Steps to Reproduce:
Open an HTML file
Type "someTag" and hit Tab, Enter
Closing tag is placed on newline at same indentation level as opening tag
On a new line, type "some-tag" and hit Tab, Enter
Closing tag is indented from opening tag
duplicate of #6371
| gharchive/issue | 2016-05-24T14:56:48 | 2025-04-01T06:37:13.794986 | {
"authors": [
"aeschli",
"mdickin"
],
"repo": "Microsoft/vscode",
"url": "https://github.com/Microsoft/vscode/issues/6811",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
425719240 | terminal current project path does not display
Issue Type: Bug
ctrl+` open terminal , current project path does not display
VS Code version: Code 1.32.3 (a3db5be9b5c6ba46bb7555ec5d60178ecc2eaae4, 2019-03-14T23:43:35.476Z)
OS version: Windows_NT x64 10.0.17763
System Info
Item
Value
CPUs
Intel(R) Core(TM) i7-7700 CPU @ 3.60GHz (8 x 3600)
GPU Status
2d_canvas: enabledchecker_imaging: disabled_offflash_3d: enabledflash_stage3d: enabledflash_stage3d_baseline: enabledgpu_compositing: enabledmultiple_raster_threads: enabled_onnative_gpu_memory_buffers: disabled_softwarerasterization: enabledsurface_synchronization: enabled_onvideo_decode: enabledwebgl: enabledwebgl2: enabled
Memory (System)
7.96GB (4.00GB free)
Process Argv
Screen Reader
no
VM
0%
Extensions (13)
Extension
Author (truncated)
Version
path-intellisense
chr
1.4.2
vscode-eslint
dba
1.8.2
gitlens
eam
9.5.1
prettier-vscode
esb
1.8.1
auto-rename-tag
for
0.0.15
vusion-peek
lin
0.1.2
vscode-language-pack-zh-hans
MS-
1.32.4
vetur
oct
0.17.1
LiveServer
rit
5.5.1
code-settings-sync
Sha
3.2.7
vscodeintellicode
Vis
1.1.5
vscode-icons
vsc
8.4.0
local-history
xyz
1.7.0
(1 theme extensions excluded)
I have the same issue. I noticed that the path is not empty but there are a lot of newline characters and because your window height it appears to be empty but its up in the window (check scrollbar).
/duplicate of #57803
| gharchive/issue | 2019-03-27T01:24:42 | 2025-04-01T06:37:13.807680 | {
"authors": [
"TheAndroidGuy",
"Yinyiran",
"usernamehw"
],
"repo": "Microsoft/vscode",
"url": "https://github.com/Microsoft/vscode/issues/71252",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
161311881 | Why do we differentiate preview open files with italic font?
VSCode Version:1.3.0-insiders
OS Version:Windows 7
I'm not sure there is benefit differentiating preview open file (with italic font) from fully opened files. What information does this give the user and how would they act on it?
Italics in the heading doesn't look good on some lower resolution monitors.
@gregvanl see https://github.com/Microsoft/vscode-docs/blob/vnext/release-notes/June_2016.md#preview-editors
We do this so that users aren't surprised when the content of a file they are looking at is replaced with the content of a new file that they single click on. We think that we need to distinguish these files. Doing so is a common pattern - Sublime Text italicizes these files and Visual Studio positions them to the far right of the document well. Knowing that a file is open in a preview tab would then let the user know that if they want to keep it open they have to take another action (the keep open action).
One question is whether or not italics is the best way to communicate this to users. A challenge we have though is that we don't want this to require significant UI management or overhead. We do not want to distract the user from focusing on their content.
One last update: Greg and I verified that the bad italic screenshot above was due to ClearType not being enabled by default on Windows Server 2008 (the OS he was using). On this particular OS, you must enable it manually, probably because it's not typical desktop machine but is meant to be tuned to be a server.
I don't think this will be a common scenario. We should be fine with continuing with italic usage, and since it's a pattern used elsewhere, I think it's appropriate.
@stevencl are there a way to never open "italic" file window (I open file by cmd+p shortcut, so no way to double click)? I find the "italic" file window annoying because in most of the time I would like the window I opened be there. And I close what I want to close manually. Thanks!
"workbench.editor.enablePreview": false
@bpasero THANKKK YOUUUUU. I loathe this feature in Visual Studio, and equally as much as VSCode. If I had to choose a single setting to put in my user settings, this would be it. Nothing more infuriating than code files closing when I click somewhere else. I don't mind inline preview windows (the idea of previewing is definitely powerful), but having "preview" tabs that disappear is one of the worst/most unintuitive UI decisions ever made by the VS team (or whatever other application they borrowed the idea from).
That said, making every setting easily overridable via one big, searchable text file is easily one of the best decisions ever made because it totally negates all of my hostility toward settings I don't like. 👍
Note to others seeing this: the keybind workbench.action.keepEditor allows you to convert a preview editor to a main editor. If your workflow requires you to dig deep into codebases (e.g. following chains of definitions) and you want to save your current point as a tab before moving into the next file WITHOUT having a tab for every file you touch in the chain, keeping previews on but committing that keybind to memory is an excellent sanity saver.
| gharchive/issue | 2016-06-20T22:36:19 | 2025-04-01T06:37:13.814510 | {
"authors": [
"bgashler1",
"bpartridge",
"bpasero",
"dbechrd",
"gregvanl",
"helinwang",
"stevencl"
],
"repo": "Microsoft/vscode",
"url": "https://github.com/Microsoft/vscode/issues/7904",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
161615407 | Scrollbar out of style for extension descriptions
Open an extension description
Notice the native scrollbar which looks out of style since we do not use it anywhere else in VSCode
I would expect our regular scroll bar which we use everywhere else (also for markdown preview)
You are like an eagle 🐦
Also looks horrible on linux + the other scroll bar should get hidden
| gharchive/issue | 2016-06-22T07:52:39 | 2025-04-01T06:37:13.817418 | {
"authors": [
"isidorn",
"joaomoreno"
],
"repo": "Microsoft/vscode",
"url": "https://github.com/Microsoft/vscode/issues/7970",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
250220939 | Latest Unity File From Releases Missing Scripts??
Hello, it seems like the unity package you all have for release is missing scripts. None of the scripts inside of unity after importing it has a way to update stats to xbox live or leaderboards. Looking at some of the scripts here in the git repository, it seems like one of the main ones is statmanager, which is missing after you import to Unity. Also the documentation for updating stats for leaderboards isn't really detailed here or here https://docs.microsoft.com/en-us/windows/uwp/xbox-live/get-started-with-creators/add-stats-and-leaderboards-in-unity. Or did I miss something somewhere? Thank you so much for all the hard work you all have put in thus far though.
I think we discussed this in https://github.com/Microsoft/xbox-live-unity-plugin/issues/105
If there's more, let me know and I'll reopen
Actually @jasonsandlin, I played around with it last night and I feel like I got it all setup correctly, however when I run it on my PC through visual studio, nothing shows up in the leaderboard. Is that because I'm in a test environment? Should my score at least show up in there? How can I make sure it is actually working properly. Or do I have to run it on my dev Xbox One?
You don't have to dev on Xbox One. It'll work on PC too, and it'll work in a test environment. The easiest way to debug is to send over a Fiddler trace so we can see what network calls are going out. Here's how to set that up: https://docs.microsoft.com/en-us/windows/uwp/xbox-live/using-xbox-live/troubleshooting/how-to-set-up-fiddler-for-debugging. Just send over a .saz file and we'll take a look
Ok, thank you. I will send one over as soon as I get a chance.
Quick question, what kind of data should I be seeing inside of it? I did put one other account as a test user as well.
Well, you should see a call to statswrite.xboxlive.com with the stats data you set, and then you should see a leaderboard call with the leaderboard data coming back down and the leaderboard should be for the stat you sent up and have previously configured on the windows dev center portal website in the xbox live section.
That's some great info. What should I see in the leaderboard? In the Unity editor it logs in the fake user and leaderboard shows all 10 fake users with scores... however on Visual Studio test, it logs me in with my gamerscore name and number, no picture, but the leaderboard is completely blank. Should I at least be seeing the other account that's a test user?
Another thing you might want to do is run:
https://github.com/Microsoft/xbox-live-samples/tree/master/Samples/CreatorsSDK/Leaderboards/UWP/Cpp
This sample does the stats write & leaderboard read. Run fiddler while its running and you'll see the calls you should see in your game.
Ok great. I'm not sure what time I'll be home to mess with it tonight but this is great info on where I can look and test in case you all are away tomorrow and thi a weekend. I'll update the post with any findings and send in a .saz file asap.
Yeah in the editor, fake data is expected. But when you build via Visual Studio its real data and you should see your own leaderboard. You should see data for all users who have posted data to that stat assuming you've configured the stat to be a leaderboard at https://developer.microsoft.com/en-US/dashboard
I do have a stat setup there and I'm using the same ID on my IntegerStat. I'll send you that .raz file as soon as I can. Probably not much else you can do without it. Thanks a lot for all the input this far though.
Yeah, we can narrow down what the issue is anyway once we see that
Let me know when you have this info and we'll reopen and investigate
Hey @jasonsandlin , sorry for not getting back with you sooner. I ran the game on my xbox and it actually posted a score. Not sure what the problem was but it seems to be working now.
| gharchive/issue | 2017-08-15T04:47:54 | 2025-04-01T06:37:13.832160 | {
"authors": [
"DutyStudios",
"jasonsandlin"
],
"repo": "Microsoft/xbox-live-unity-plugin",
"url": "https://github.com/Microsoft/xbox-live-unity-plugin/issues/104",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
851262681 | Update user-access.md
https://github.com/MicrosoftDocs/OfficeDocs-SkypeForBusiness/issues/6799
Added important note.
@JohanFreelancer9 Copy editing is needed for this PR. Thanks.
@cichur PR has been copyedited and is ready for final review, could you please check and merge? Thanks!
| gharchive/pull-request | 2021-04-06T09:59:47 | 2025-04-01T06:37:13.850763 | {
"authors": [
"lindspea",
"scanum"
],
"repo": "MicrosoftDocs/OfficeDocs-SkypeForBusiness",
"url": "https://github.com/MicrosoftDocs/OfficeDocs-SkypeForBusiness/pull/6981",
"license": "CC-BY-4.0",
"license_type": "permissive",
"license_source": "github-api"
} |
640163150 | Incorrect WPA url
In section https://docs.microsoft.com/en-us/Workplace-Analytics/tutorials/power-bi-bc#odata-request-failed
Url is listed as https://workplace.analytics.office.com , there is an extra dot between workplace & analytics.
It should be https://workplaceanalytics.office.com
Document Details
⚠ Do not edit this section. It is required for docs.microsoft.com ➟ GitHub issue linking.
ID: 208c8848-3e5a-2c94-d89f-fba11efdd208
Version Independent ID: 054b67b2-f71f-4f71-f055-1b47362dc6e3
Content: Power BI Business continuity - Workplace Intelligence
Content Source: WorkplaceAnalytics/Tutorials/power-bi-bc.md
Product: wpa
GitHub Login: @madehmer
Microsoft Alias: madehmer
Hi rpanday,
Thanks much for this correction! We've made the fix in the documentation:
OData request failed
-- paul9955
| gharchive/issue | 2020-06-17T05:55:07 | 2025-04-01T06:37:13.856210 | {
"authors": [
"paul9955",
"rpanday"
],
"repo": "MicrosoftDocs/OfficeDocs-WorkplaceAnalytics",
"url": "https://github.com/MicrosoftDocs/OfficeDocs-WorkplaceAnalytics/issues/133",
"license": "CC-BY-4.0",
"license_type": "permissive",
"license_source": "github-api"
} |
429551924 | _SMSTSClientCache variable
Does "_SMSTSClientCache" still exist in 1806 onwards? As it's not referenced over on the main Variables page - https://docs.microsoft.com/en-us/sccm/osd/understand/task-sequence-variables
Document Details
⚠ Do not edit this section. It is required for docs.microsoft.com ➟ GitHub issue linking.
ID: 3453e235-3b67-aee4-dd8c-e6c4c60bdaa4
Version Independent ID: 5d1d00c7-f6db-3d0d-74bd-66ccbfd76fa7
Content: Create a task sequence to install an operating system - Configuration Manager
Content Source: sccm/osd/deploy-use/create-a-task-sequence-to-install-an-operating-system.md
Product: configuration-manager
Technology: configmgr-osd
GitHub Login: @aczechowski
Microsoft Alias: aaroncz
Thanks @mattsmacblog for the feedback. Yeah, we don't generally remove variables. There are a number of variables that aren't included in the doc, and on my to do list to work with engineering to review and revise (see #551).
I'm going to adjust this feedback to link to the variables article, as that's where the change needs to happen.
| gharchive/issue | 2019-04-05T01:55:28 | 2025-04-01T06:37:13.892629 | {
"authors": [
"aczechowski",
"mattsmacblog"
],
"repo": "MicrosoftDocs/SCCMdocs",
"url": "https://github.com/MicrosoftDocs/SCCMdocs/issues/1400",
"license": "CC-BY-4.0",
"license_type": "permissive",
"license_source": "github-api"
} |
529155395 | Customized image
Hi,
the sentence: "You have custom requirements including using a custom base image, using third-party disk encryption, or require WinPE offline operations." I am having an issue where I want to clear out built in apps, I do this with dism, but still after the TS is finished, I get the "original" OS with all the built in apps installed. Why is this happening?
Thanks
Document Details
⚠ Do not edit this section. It is required for docs.microsoft.com ➟ GitHub issue linking.
ID: 05088ead-21fe-a787-b5b6-d0a90c219f5b
Version Independent ID: a801446b-a068-f305-3fcc-a84420698bb9
Content: Upgrade to Windows 10 - Configuration Manager
Content Source: sccm/osd/deploy-use/upgrade-windows-to-the-latest-version.md
Product: configuration-manager
Technology: configmgr-osd
GitHub Login: @aczechowski
Microsoft Alias: aaroncz
Thanks @TonitoDux for taking the time to share this with the Configuration Manager documentation team. I’m sorry that you’re having problems, and I want to make sure it gets to the right people that can help. There are a couple of options available to provide the best level of support for this:
Post on Configuration Manager forums. These are moderated by our community MVPs who are very experienced, knowledgeable, and helpful.
Use Microsoft Support to search for the issue, or open a case with professional support.
Contact your Microsoft Premier team, such as a Premier Field Engineer or Technical Account Manager.
I'm closing this issue as there's nothing actionable for docs at this time.
| gharchive/issue | 2019-11-27T06:57:23 | 2025-04-01T06:37:13.899138 | {
"authors": [
"TonitoDux",
"aczechowski"
],
"repo": "MicrosoftDocs/SCCMdocs",
"url": "https://github.com/MicrosoftDocs/SCCMdocs/issues/2038",
"license": "CC-BY-4.0",
"license_type": "permissive",
"license_source": "github-api"
} |
533500137 | Update optimize-windows-10-update-delivery.md
The reference "all windows updates" is causing confusion for IT Pros as they assume this includes Office which is not the case. Adding line of clarity on how to unwind Office integration with SCCM to use DO.
@davguent : Thanks for your contribution! The author(s) have been notified to review your proposed change.
| gharchive/pull-request | 2019-12-05T17:46:19 | 2025-04-01T06:37:13.900657 | {
"authors": [
"PRMerger9",
"davguent"
],
"repo": "MicrosoftDocs/SCCMdocs",
"url": "https://github.com/MicrosoftDocs/SCCMdocs/pull/2057",
"license": "CC-BY-4.0",
"license_type": "permissive",
"license_source": "github-api"
} |
616511037 | Problemas com a aplicação Glovo
Este é o link da aplicação Glovo
https://slack-redir.net/link?url=https%3A%2F%2Finstall.appcenter.ms%2Forgs%2Fglovoapp%2Fapps%2Fglover-1%2Fdistribution_groups%2Fpublic
Thanks for filing this report & sorry that you haven't heard back from us. It sounds like this issue for technical support, but this repository isn't monitored by our support team.
If you need technical help, you can contact support by:
Going to https://appcenter.ms/apps
Selecting ? > Contact Support (in the upper right corner).
Reference this issue for context if needed.
If you've already resolved your issue/still want us to document this, give us an update of your current observations. We're catching up on our backlog here.
| gharchive/issue | 2020-05-12T09:37:55 | 2025-04-01T06:37:13.906222 | {
"authors": [
"Claytonglovo",
"King-of-Spades"
],
"repo": "MicrosoftDocs/appcenter-docs",
"url": "https://github.com/MicrosoftDocs/appcenter-docs/issues/1027",
"license": "CC-BY-4.0",
"license_type": "permissive",
"license_source": "github-api"
} |
485033385 | Comprehensive
I found this document to be comprehensive, beautifully written (with examples to clarify each point) and very useful. Great job.
Document Details
⚠ Do not edit this section. It is required for docs.microsoft.com ➟ GitHub issue linking.
ID: ad651d6a-f8d8-2050-e7a7-adbae7326112
Version Independent ID: 4fddd724-cabd-8623-7692-857088c1dccb
Content: API design guidance - Best practices for cloud applications
Content Source: docs/best-practices/api-design.md
Service: architecture-center
Sub-service: cloud-fundamentals
GitHub Login: @dragon119
Microsoft Alias: pnp
@ashishjain87 Thanks for the kudos and sharing your feedback and thoughts on this matter. Surely, it'll benefit others visiting thread.
We will now proceed to close this thread for now since this channel is more for fixing issues in the scope of the documentation and driving improvements towards it. If there are further questions regarding this matter, please reopen it and we will gladly continue the discussion.
| gharchive/issue | 2019-08-26T04:13:22 | 2025-04-01T06:37:13.910876 | {
"authors": [
"KetanChawda-MSFT",
"ashishjain87"
],
"repo": "MicrosoftDocs/architecture-center",
"url": "https://github.com/MicrosoftDocs/architecture-center/issues/1738",
"license": "CC-BY-4.0",
"license_type": "permissive",
"license_source": "github-api"
} |
735133199 | Business Central on Azure Implementation Details
Good day Team Microsoft
I would love to get Dynamics Business Central on Azure implementation details with a good reference architecture too.
There seem to be a big demand for this solution here in Africa.
Will really appreciate your feedback.
Regards,
Thanks for reporting - this issue is under review. This is a Microsoft Internal DevOps Bug ID AB#331229
Thank you for your feedback. We have created internal ticket 24406 to track and investigate this issue.
#please-close
| gharchive/issue | 2020-11-03T08:49:16 | 2025-04-01T06:37:13.912851 | {
"authors": [
"absheik",
"gchatira",
"skabou"
],
"repo": "MicrosoftDocs/architecture-center",
"url": "https://github.com/MicrosoftDocs/architecture-center/issues/2474",
"license": "CC-BY-4.0",
"license_type": "permissive",
"license_source": "github-api"
} |
1563043719 | Unlist Azure Artifact packages link in DevOps not showing documentation for unlisting
When you want to unlist a package in Azure Artifacts, you get prompted with a sidebar which shows you the details of what you're unlisting.
At the top of the sidebar, you get this notification:
If you click through to read more, you are brought through to the page this issue is associated with. This document is around deleting and recovering packages, and doesn't cover the details of unlisting packages. Exploring the rest of the documentation, there doesn't appear to be information around how the unlist function works.
Document Details
⚠ Do not edit this section. It is required for learn.microsoft.com ➟ GitHub issue linking.
ID: 31a8d31b-229b-7383-1cc0-f3d7c65eca84
Version Independent ID: 70ba7acf-adae-e1b7-a42d-3f1344c5293c
Content: Delete and recover packages - Azure Artifacts
Content Source: docs/artifacts/how-to/delete-and-recover-packages.md
Service: azure-devops-artifacts
GitHub Login: @ramiMSFT
Microsoft Alias: rabououn
@jamesmcroft -- James, thank you for your thorough feedback. Please consider these resources:
Azure DevOps Services
Azure DevOps on Stack Overflow
@ramiMSFT -- Rami, please look into this issue.
Thanks for the feedback! Unlisting applies to NuGet packages. We have that information under the NuGet tab, but I will update the article to make it more clear.
| gharchive/issue | 2023-01-30T19:10:50 | 2025-04-01T06:37:13.918981 | {
"authors": [
"WilliamAntonRohm",
"jamesmcroft",
"ramiMSFT"
],
"repo": "MicrosoftDocs/azure-devops-docs",
"url": "https://github.com/MicrosoftDocs/azure-devops-docs/issues/13076",
"license": "CC-BY-4.0",
"license_type": "permissive",
"license_source": "github-api"
} |
704462110 | Stage can be set to run based on a variable output in a previous stage.
[Enter feedback here]
You cannot currently specify that a stage run based on the value of an output variable set in a previous stage.
It is not true.
You can specify that a stage run based on the value of an output variable set in a previous stage.
- stage: FirstStage
jobs:
- job: decide_what_to_run
steps:
- task: PowerShell@2
name: trigger_stages
inputs:
targetType: 'inline'
script: |
Write-Output "##vso[task.setvariable variable=SecondStage_should_run;isOutput=true]true"
- stage: SecondStage
condition: eq(dependencies.FirstStage.outputs['decide_what_to_run.trigger_stages.SecondStage_should_run'], 'true')
jobs:
- job: do_something
steps:
- powershell: |
Write-Output "I have been run"
Please also check #8486 and #8366 and Sprint-168-Update which states it is possible now
Document Details
⚠ Do not edit this section. It is required for docs.microsoft.com ➟ GitHub issue linking.
ID: d322215c-8025-4f21-0700-7dfa7dc5c46e
Version Independent ID: 141fcdbb-8394-525b-bb29-eff9a693a9c4
Content: Stages in Azure Pipelines - Azure Pipelines
Content Source: docs/pipelines/process/stages.md
Product: devops
Technology: devops-cicd-process
GitHub Login: @juliakm
Microsoft Alias: jukullam
@ivanakcheurov Thank you for the feedback.
The documentation has been updated and is pending review.
This is not correct. The sprint documentation linked to states that job level conditions can use output variables set in a previous stage, not stage level conditions.
| gharchive/issue | 2020-09-18T15:19:37 | 2025-04-01T06:37:13.925282 | {
"authors": [
"chadwickalex",
"ivanakcheurov",
"msebolt"
],
"repo": "MicrosoftDocs/azure-devops-docs",
"url": "https://github.com/MicrosoftDocs/azure-devops-docs/issues/9306",
"license": "CC-BY-4.0",
"license_type": "permissive",
"license_source": "github-api"
} |
707472196 | Installed Software on Ubuntu 18.04 managed agent is not matching the documentation
Hello,
The Docs has a link to all software installed on the Ubuntu 18.04. In this list Az Modules (Powershell) are present.
But when I use this image on my Release pipeline, I get an error saying that the Az module is not installed.
To verify, I have also added a step on my pipeline that firstly install the Az Module. Then it all works fine.
According to this doc page, the Ubuntu 18.04 managed agent list of pre installed software is here: https://github.com/actions/virtual-environments/blob/main/images/linux/Ubuntu1804-README.md
and there Az Modules (from version 1.0.0 to 4.6.0) are listed.
Looks like there is a mismatch on the image that the Managed Agent is actually from a gallery image to build the runtime agent VM..
Thanks
Alex
Document Details
⚠ Do not edit this section. It is required for docs.microsoft.com ➟ GitHub issue linking.
ID: 61db3371-6395-88a2-fb78-09f98790ab1b
Version Independent ID: ef6f1bf7-02b8-9093-b84e-c9bde7d98e57
Content: Microsoft-hosted agents for Azure Pipelines - Azure Pipelines
Content Source: docs/pipelines/agents/hosted.md
Product: devops
Technology: devops-cicd-agents
GitHub Login: @steved0x
Microsoft Alias: sdanie
Issue was on my side. I had to use the Azure PowerShell task
| gharchive/issue | 2020-09-23T15:35:07 | 2025-04-01T06:37:13.931397 | {
"authors": [
"alexmpmenezes"
],
"repo": "MicrosoftDocs/azure-devops-docs",
"url": "https://github.com/MicrosoftDocs/azure-devops-docs/issues/9335",
"license": "CC-BY-4.0",
"license_type": "permissive",
"license_source": "github-api"
} |
742937378 | Update storage-blob-readme.md
I created a sample project using azure blob SDK and I getting an error on this part, does the url + sas seem wrong or I am the one who is wrong? I added, "/?" on my sample project, and its works successfully.
Docs Build status updates of commit 43b7506:
:clock10: Pending: waiting for processors (20 builds ahead of you)
:warning: Docs Build is busy, currently there are 20 builds ahead of this one, for more information you can view the Build queue graph on the Docs Portal.
Docs Build status updates of commit 43b7506:
:x: Validation status: errors
Please follow instructions here which may help to resolve issue.
File
Status
Preview URL
Details
:x:Error
Details
[Error-FailedToRestoreDependentRepository]
Failed to restore dependent repository from URL 'https://github.com/Microsoft/templates.docs.msft', branch 'master' into folder '_themes'. Please open a ticket in https://SiteHelp and include exception details: System.TimeoutException: Git operation '-c http.extraheader="Authorization: *****" pull origin --no-edit' timed out. Details: git.exe (argument: -c http.extraheader="Authorization: *****" pull origin --no-edit, working directory: W:\olty-s\_themes) didn't exit in 180000 ms. Forced to exit..
at Microsoft.OpenPublishing.Build.Common.ProcessUtility.ExecuteExeCore(ProcessExecutionInfo executionInfo, Action`1 outputAction, StreamWriter stderrStreamWriter, Int32 timeoutInMilliseconds, CancellationToken cancellationToken)
at Microsoft.OpenPublishing.Build.Common.ProcessUtility.ExecuteExe(ProcessExecutionInfo executionInfo, StreamWriter stdoutStreamWriter, StreamWriter stderrStreamWriter, Int32 timeoutInMilliseconds, CancellationToken cancellationToken)
at Microsoft.OpenPublishing.Build.DataAccessor.Git.ExternalGitAccessor.RunGitExeAndProcessOutputStream(String argument, String workingDirectory, Uri repositoryUrl, String branchName, Action`1 outPutStreamAction, Boolean silentException, Int32 timeout, Boolean retryOnTimeout)
at Microsoft.OpenPublishing.Build.DataAccessor.Git.ExternalGitAccessor.PullCore()
at Microsoft.OpenPublishing.Build.DataAccessor.Git.BaseGitAccessor.Pull()
at Microsoft.OpenPublishing.Build.Applications.InitializeCRRTool.Program.InitializeDependentRepository(String dependentRepoPath, String cloneAccessToken, Boolean isCopiedFromCachedRepo, Boolean isPreservedTemplateFolder, Options options, DependentRepository dependentRepository)
at Microsoft.OpenPublishing.Build.Applications.InitializeCRRTool.Program.InitializeCRRCore(String dependentRepoPath, String cloneAccessToken, Boolean isPreservedTemplateFolder, Boolean isCopiedFromCachedRepo, String branch, Options options, DependentRepository dependentRepository, List`1 lastDependencyStatus)
at Microsoft.OpenPublishing.Build.Applications.InitializeCRRTool.Program.<InitializeCRRCoreAsync>d__8.MoveNext()
--- End of stack trace from previous location where exception was thrown ---
at System.Runtime.ExceptionServices.ExceptionDispatchInfo.Throw()
at System.Runtime.CompilerServices.TaskAwaiter.HandleNonSuccessAndDebuggerNotification(Task task)
at Microsoft.OpenPublishing.Build.Applications.InitializeCRRTool.Program.<InitializeCRRAsync>d__5.MoveNext()
For more details, please refer to the build report.
If you see build warnings/errors with permission issues, it might be due to single sign-on (SSO) enabled on Microsoft's GitHub organizations. Please follow instructions here to re-authorize your GitHub account to Docs Build.
Note: Broken links written as relative paths are included in the above build report. For broken links written as absolute paths or external URLs, see the broken link report.
Note: Your PR may contain errors or warnings unrelated to the files you changed. This happens when external dependencies like GitHub alias, Microsoft alias, cross repo links are updated. Please use these instructions to resolve them.
For any questions, please:Try searching in the Docs contributor and Admin GuideSee the frequently asked questionsPost your question in the Docs support channel
@jbmarvin24 Thanks for the contribution! This definitely looks like a reasonable adjustment. I mirrored your update over to the source of this repo.
Hey @jbmarvin24 it looks like the sample is just written to work with the SAS token that is generated form the azure portal. When you grab the SAS from azure it'll have the ?sv= already appended.
I think given that, the sample is probably ok.
| gharchive/pull-request | 2020-11-14T06:07:33 | 2025-04-01T06:37:13.944333 | {
"authors": [
"jbmarvin24",
"opbld30",
"opbld32",
"scbedd"
],
"repo": "MicrosoftDocs/azure-docs-sdk-node",
"url": "https://github.com/MicrosoftDocs/azure-docs-sdk-node/pull/955",
"license": "CC-BY-4.0",
"license_type": "permissive",
"license_source": "github-api"
} |
1425352321 | How to deploy the custom connector to logic app from azure devops
[Enter feedback here]
Document Details
⚠ Do not edit this section. It is required for learn.microsoft.com ➟ GitHub issue linking.
ID: 01046995-1aea-643d-cbfd-c2ef469e28a8
Version Independent ID: 1a9bc699-b04a-c763-34e1-e412bbdcde0b
Content: Create built-in connectors for Standard logic apps - Azure Logic Apps
Content Source: articles/logic-apps/create-custom-built-in-connector-standard.md
Service: logic-apps
GitHub Login: @ecfan
Microsoft Alias: estfan
@AnandKatta Thanks for reaching out and posting the question. We will review and update you.
@AnandKatta You can leverage the ARM template resource type Microsoft.Web/customApis to define your custom connector and integrate ARM template with your DevOps pipeline.
If you need any further assistance on configuring/creating ARM template, you can always create a thread on the forums - Microsoft Q&A or Stack Overflow
Once you post your issue on forums, it will have visibility across the community which is a better suited audience for such types of issues.
I will go ahead and close this issue now.
| gharchive/issue | 2022-10-27T09:48:31 | 2025-04-01T06:37:13.975587 | {
"authors": [
"AnandKatta",
"MayankBargali-MSFT",
"mumurug-MSFT"
],
"repo": "MicrosoftDocs/azure-docs",
"url": "https://github.com/MicrosoftDocs/azure-docs/issues/100558",
"license": "CC-BY-4.0",
"license_type": "permissive",
"license_source": "github-api"
} |
1426172231 | Role for managing plan
Please add information about what role is needed to change or cancel a purchased plan
Document Details
⚠ Do not edit this section. It is required for learn.microsoft.com ➟ GitHub issue linking.
ID: d7f5fdf9-4250-6b2e-8cf4-d270e8e6c94d
Version Independent ID: 3284cb1b-315f-76dd-2d4b-5022e8729d97
Content: Get started with Microsoft Defender for IoT - Microsoft Defender for IoT
Content Source: articles/defender-for-iot/organizations/getting-started.md
Service: defender-for-iot
GitHub Login: @batamig
Microsoft Alias: bagol
@DeanGross
Thanks for your feedback! We will investigate and update as appropriate.
Hi @DeanGross, thanks for writing into Microsoft Learn!
Changes on the Plans and prices page require a Security Admin, Contributor, or Owner role.
You can find more information here: Azure user roles and permissions for Defender for IoT.
I'm going to close this issue for now, but please feel free to continue commenting if you have additional questions.
We appreciate your feedback!
#please-close
| gharchive/issue | 2022-10-27T19:33:46 | 2025-04-01T06:37:13.981392 | {
"authors": [
"AjayBathini-MSFT",
"DeanGross",
"batamig"
],
"repo": "MicrosoftDocs/azure-docs",
"url": "https://github.com/MicrosoftDocs/azure-docs/issues/100596",
"license": "CC-BY-4.0",
"license_type": "permissive",
"license_source": "github-api"
} |
1453322445 | Readable Dates (International)
Hi - is it possible not to use US date format on this page but to use something that is clearer for an international audience. For example at the top at the moment you have "11/9/2022: Released for download", which is 11th September 2022 as I see it (UK based), but I think you mean 9th November 2022. There are many examples of this in the document and its only by scrolling down to the heading 2.1.1.0 where it reads "3/24/2022: Released for download only, not available for auto upgrade" that the date format becomes obvious.
Thanks
Document Details
⚠ Do not edit this section. It is required for learn.microsoft.com ➟ GitHub issue linking.
ID: 57f5478b-3132-21d0-a420-a1140db235de
Version Independent ID: 266bbd16-2117-f680-6493-e183eb09fdbc
Content: Azure AD Connect: Version release history - Microsoft Entra
Content Source: articles/active-directory/hybrid/reference-connect-version-history.md
Service: active-directory
Sub-service: hybrid
GitHub Login: @billmath
Microsoft Alias: rodejo
@brianreidc7
Thanks for your feedback! We will investigate and update as appropriate.
@brianreidc7
Thanks for bringing this to our attention.
I'm going to assign this to the document author so they can take a look at it accordingly.
@billmath
Could you please add your comments on the above query
Thank you for submitting this issue. Unfortunately, due to high volumes I was unable to respond to this in a timely manner. At this point, I am going to close this due to it's age and the possibility that it may have been addressed. If this is still a problem, please consider submitting a new issue. Thank you.
#please-close
| gharchive/issue | 2022-11-17T13:11:10 | 2025-04-01T06:37:13.988904 | {
"authors": [
"AjayBathini-MSFT",
"ManoharLakkoju-MSFT",
"billmath",
"brianreidc7"
],
"repo": "MicrosoftDocs/azure-docs",
"url": "https://github.com/MicrosoftDocs/azure-docs/issues/101583",
"license": "CC-BY-4.0",
"license_type": "permissive",
"license_source": "github-api"
} |
1473032776 | Avoid Issues With Example Code
I suggest to add the attributes of the VotesController class
[Produces("application/json")]
[Route("api/Votes")]
to the final VotesController in the Update the VotesController.cs file section.
Pro: the attributes belong to the class.
Con: the previous implementation had the attributes already
Document Details
⚠ Do not edit this section. It is required for learn.microsoft.com ➟ GitHub issue linking.
ID: c0d03b4f-4157-0d04-0f48-0a3e2fb966cb
Version Independent ID: 6b7bed89-eb53-105a-7971-476ddd4cf3b7
Content: Create a .NET app on Service Fabric in Azure - Azure Service Fabric
Content Source: articles/service-fabric/service-fabric-tutorial-create-dotnet-app.md
Service: service-fabric
GitHub Login: @tomvcassidy
Microsoft Alias: tomcassidy
@mstroppel
Thanks for your feedback! I've assigned this issue to the author who will investigate and update as appropriate.
@tomvcassidy
Can you please check and add your comments on this doc update request as applicable.
Thanks for your dedication to our documentation. Unfortunately, at this time we have been unable to review your issue in a timely manner and we sincerely apologize for the delayed response. Though the scope of our feedback channel here on GitHub covers specific documentation fixes, we can help redirect you to the right support channel to get an answer to your question: [reference to alternative product/service or community endpoint]. Because this issue does not relate to documentation, we are closing this issue. #please-close
| gharchive/issue | 2022-12-02T15:18:14 | 2025-04-01T06:37:13.995351 | {
"authors": [
"SaibabaBalapur-MSFT",
"mstroppel",
"tomvcassidy"
],
"repo": "MicrosoftDocs/azure-docs",
"url": "https://github.com/MicrosoftDocs/azure-docs/issues/102192",
"license": "CC-BY-4.0",
"license_type": "permissive",
"license_source": "github-api"
} |
1657882177 | Please add more details about a reset for Active-Active VPN Gateways
Hi Team,
Today I had a case in which the client is concerned about why his Active-Active BGP gateway Instance 1 doesn`t reset when they do a reset over the portal as their understanding is that the 2ND INSTANCE would be rebooted after the first reset is done and they issue a new reset, which is not the case based on this wiki article below, can you please have this doc updated?:
https://supportability.visualstudio.com/AzureNetworking/_wiki/wikis/Wiki/330301/About-ResetGatewayOperation
Is there a difference in ResetGatewayOperation on an Active – Active Gateway?
In Active – Active setup both the GatewayTenantWorker_IN_0 & GatewayTenantWorker_IN_1 are active. Both being Active Instances, we still mark GatewayTenantWorker_IN_0 as Primary 99% of the times and GatewayTenantWorker_IN_1 as Secondary.
In short - GatewayTenantWorker_IN_1 will never be reset.
How to reset an Active – Active Gateway to ensure that GatewayTenantWorker_IN_1 Is also reset?
List the Gateway VIP –
Primary VIP is associated with GatewayTenantWorker_IN_0
Secondary VIP is associated with GatewayTenantWorker_IN_1
$Gateway = Get-AzVirtualNetworkGateway -Name "ContosoVirtualGateway"
Reset-AzVirtualNetworkGateway -VirtualNetworkGateway $Gateway -GatewayVip "
Document Details
⚠ Do not edit this section. It is required for learn.microsoft.com ➟ GitHub issue linking.
ID: 98d7fd30-80a8-f690-6191-d97ae4cac174
Version Independent ID: 3e15f99b-2f2d-7adc-a2d9-0b2184989db7
Content: Azure VPN Gateway FAQ
Content Source: articles/vpn-gateway/vpn-gateway-vpn-faq.md
Service: vpn-gateway
GitHub Login: @cherylmc
Microsoft Alias: cherylmc
@PabloG-MSsupport
Thanks for your feedback! We will investigate and update as appropriate.
@cherylmc It appears RESET behavior is different for a Active-Active Gateway.
It would be good if we could document this in the FAQ section as well is in Reset a VPN gateway or a connection
@PabloG-MSsupport I have assigned this thread to the author, they shall review this and update the docs as required
#reassign:aybatra
#assign:deeharris-23
#reassign:radwiv
Hello @radwiv any update with this request?
Hi @KapilAnanth-MSFT @cherylmc and @deeharris-23 any update regarding this document update ?
We've updated this article: https://learn.microsoft.com/azure/vpn-gateway/reset-gateway
#please-close
| gharchive/issue | 2023-04-06T18:39:44 | 2025-04-01T06:37:14.004961 | {
"authors": [
"KapilAnanth-MSFT",
"PabloG-MSsupport",
"YashikaTyagii",
"cherylmc",
"deeharris-23"
],
"repo": "MicrosoftDocs/azure-docs",
"url": "https://github.com/MicrosoftDocs/azure-docs/issues/107817",
"license": "CC-BY-4.0",
"license_type": "permissive",
"license_source": "github-api"
} |
1664616895 | unclear text regarding the "path parameter"
Are you sure this text makes sense?
You can try passing different values to the function by changing the path parameter.
What path is this?
Document Details
⚠ Do not edit this section. It is required for learn.microsoft.com ➟ GitHub issue linking.
ID: e62f248c-b170-d9fd-96e2-830ed163b6fc
Version Independent ID: 3e7b26b1-3464-2062-ec31-9a0c279cd87e
Content: Tutorial: Managed identity to invoke Azure Functions
Content Source: articles/spring-apps/tutorial-managed-identities-functions.md
Service: spring-apps
GitHub Login: @KarlErickson
Microsoft Alias: margard
@ciufudean
Thanks for your feedback! We will investigate and update as appropriate.
Thanks, @ciufudean! I've submitted a pull request to fix this. I'm not sure what that line refers to, so I'm removing it to avoid confusion.
| gharchive/issue | 2023-04-12T13:53:07 | 2025-04-01T06:37:14.010436 | {
"authors": [
"AjayBathini-MSFT",
"KarlErickson",
"ciufudean"
],
"repo": "MicrosoftDocs/azure-docs",
"url": "https://github.com/MicrosoftDocs/azure-docs/issues/108101",
"license": "CC-BY-4.0",
"license_type": "permissive",
"license_source": "github-api"
} |
1833377458 | Broken Link in document
[Enter feedback here]
The document refers to another document to support multiple data sources. The link is broken:
https://learn.microsoft.com/en-us/samples/azure-samples/azure-search-dotnet-samples/multiple-data-sources/
Document Details
⚠ Do not edit this section. It is required for learn.microsoft.com ➟ GitHub issue linking.
ID: e990abaf-1e21-6de6-c6e8-fecf49e6cf23
Version Independent ID: 55fdf842-baf2-7c13-7332-76907a1cc0ca
Content: Index large data sets for full text search - Azure Cognitive Search
Content Source: articles/search/search-howto-large-index.md
Service: cognitive-search
GitHub Login: @HeidiSteen
Microsoft Alias: heidist
#please-close
| gharchive/issue | 2023-08-02T15:10:02 | 2025-04-01T06:37:14.014899 | {
"authors": [
"HeidiSteen",
"rmillman"
],
"repo": "MicrosoftDocs/azure-docs",
"url": "https://github.com/MicrosoftDocs/azure-docs/issues/113086",
"license": "CC-BY-4.0",
"license_type": "permissive",
"license_source": "github-api"
} |
2071845904 | Target based scaling with Service Bus queues and topics using batching and sessions
I'm not sure if this is the right place to ask but there is nothing in the documentation about target based scaling with service bus using both batching and sessions. Does it work? Or does it fall back to incremental scaling? Would be nice to add this to the documentation.
Thanks in advance!
Document Details
⚠ Do not edit this section. It is required for learn.microsoft.com ➟ GitHub issue linking.
ID: e4b9daf3-aa85-9dd9-6056-e9e07b2a8285
Version Independent ID: 5f8d6b0f-c46c-158e-3309-bcc645b0a046
Content: Target-based scaling in Azure Functions
Content Source: articles/azure-functions/functions-target-based-scaling.md
Service: azure-functions
GitHub Login: @ggailey777
Microsoft Alias: glenga
@smolleman
Thanks for your feedback! We will investigate and update as appropriate.
| gharchive/issue | 2024-01-09T08:12:36 | 2025-04-01T06:37:14.019314 | {
"authors": [
"Naveenommi-MSFT",
"smolleman"
],
"repo": "MicrosoftDocs/azure-docs",
"url": "https://github.com/MicrosoftDocs/azure-docs/issues/118658",
"license": "CC-BY-4.0",
"license_type": "permissive",
"license_source": "github-api"
} |
2139261704 | Manual on GStreamer installation on macOs
It is not clear how to properly install and configure gstreamer on macOs:
I can either install latest version using 'brew', but as I understand latest version it is not supported by Speech SDK. Or I can download older versions directly from gstreamer website. Which option to prefer? Which version to use?
From the manual it is not clear which version of the gstream to download (after digging the internet I guess that currently Speech SDK supports only gstreamer version 1.18.3)
it is not clear which env variables should be set up on macOs. I definitely have gstreamer working and installed, but I still get SPXERR_GSTREAMER_NOT_FOUND_ERROR. My guess here that Azure package is looking at either GST_PLUGIN_PATH or GSTREAMER_ROOT_X86_64 or PATH. However setting either of them worked for me. And again it is not clear how to read the Speech SDK doc, because it is telling that you need to point to ".dlyb" files. Usually you don't add ".dlyb" files to PATH on mac.
So overall I think that current version of the doc is clearly describing how to setup gstreamer on Windows, but lack of crucial steps for macOs.
Thank you.
Document Details
⚠ Do not edit this section. It is required for learn.microsoft.com ➟ GitHub issue linking.
ID: 095c74d0-66b3-9cd7-4588-291bf4f56a47
Version Independent ID: 8125527b-407f-c34f-57e8-fea89db144e8
Content: How to use compressed input audio - Speech service - Azure AI services
Content Source: articles/ai-services/speech-service/how-to-use-codec-compressed-audio-input-streams.md
Service: azure-ai-speech
GitHub Login: @eric-urban
Microsoft Alias: eur
@Sirorezka
Thanks for your feedback! We will investigate and update as appropriate.
@Sirorezka
I've delegated this to @eric-urban, a content author, to review and share their valuable insights.
@Sirorezka - Right we haven't documented the details about macOS. Please see the GStreamer docs for more information about that. Thanks!
#please-close
Why has this ticket being closed? There is no resolution, apart from "check the GStreamer docs", which is really not very helpful.
@Sirorezka were you able to figure this out? I'm stuck as well.
Same here
No, I didn't find a solution. And at the end decided to convert audio to .wav as a best alternative. However, speaking about the issue, I didn't test the link posted by eric, maybe it did have a solution.
Because the issue was closed maybe you need to create a new issue if you experience the same problem.
| gharchive/issue | 2024-02-16T19:11:31 | 2025-04-01T06:37:14.028366 | {
"authors": [
"Naveenommi-MSFT",
"RamanathanChinnappan-MSFT",
"Sirorezka",
"WaterKnight1998",
"adinapoli",
"eric-urban",
"neeagl"
],
"repo": "MicrosoftDocs/azure-docs",
"url": "https://github.com/MicrosoftDocs/azure-docs/issues/119937",
"license": "CC-BY-4.0",
"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.