Vscode: Improve VS Code when working with network drives

Created on 13 Oct 2016  ·  71Comments  ·  Source: microsoft/vscode

Some issues impact users that try to run VS Code against slow network drives. Specifically:

  • [ ] some folders fail to open or display (#26931, #20207, #21397, #21520, #25885)
  • [ ] we remove folders from the recent folders list when they do not seem to exist even though they are (13699)
  • [x] when opening an already opened file, we load its metadata to find out if it changed (#5353)
  • [x] when opening an already opened file, we open it to detect the encoding (https://github.com/Microsoft/vscode/issues/21604)
  • [x] we close an editor when we detect a file delete event (this can happen from faulty file events) (#11642)
  • [x] we prevent saving when we think the contents on disk are newer (which seems to happen sporadically on network drives) (#18037)
debt file-io

Most helpful comment

I was trying to use VS Code with my git repository on a samba share. I am using mac and the samba share is on a remote Ubuntu 18.04 machine hosted on AWS. VS code and git would not work properly when I had my repository on a samba share but I had to get work done so I went ahead and started working without git. I was editing a file that was 2 days old, hours of work, and VS code hanged. When I reopened VS Code I was shocked to see my file on the server was blank. In fact, every file I had open at the time of the crash was blank. Now if git would have been workign properly I would not have lost data.

I guess it is back to Dreamweaver? Before I do somethign crazy like that, I am trying also a NFS. Git hub is still having issues. but I got 2 commits to go through. I am hacing trouble with more commits. It keeps telling me I can not write and index.lock file. Prob because VS code is unable to write to the network drive

All 71 comments

we remove folders from the recent folders list when they do not seem to exist even though they are

Just had this happen, I did notice once the drive is reconnected and I open the folder manually all the open files I had were re-opened.

Edit: Closer inspection, these are files I had open at some point, but not what I had open last. Also did not maintain the order of the files opened.

Another problem while using network drives is that the go to file ctrl+p is very slow, pretty much unusable.

This happens a lot while working on my computer (not a network hard drive).
To make things worse, it flushes any "undo history" I may have. So even if I go through the compare and accept changes, then I cannot undo any of the stuff.

It is starting to get a bit frustrating. You should at least add a setting to disable it.

Update: With the new update I can't even compare and save anymore. Now after I press the compare button I get a notification pop up that says Could not resolve any model with provided uri..
So now I have a lot of unsaved changes in my file and no way to save them.

This is happening very frequently to me as well, working on a network drive as well, every other save I get an error

I got the same frequent error on a samba share via virtualbox host only network.

I pushed a fix for the annoying save error, happy for feedback in todays insider build.

Great ! For me however, no errors since 1.8.1

2017-01-04 8:11 GMT+01:00 Benjamin Pasero notifications@github.com:

I pushed a fix for the annoying save error, happy for feedback in todays
insider build.


You are receiving this because you commented.
Reply to this email directly, view it on GitHub
https://github.com/Microsoft/vscode/issues/13665#issuecomment-270307761,
or mute the thread
https://github.com/notifications/unsubscribe-auth/AEay-GUZdCViO5OfpV2OURwkvU_-jBP2ks5rO0YhgaJpZM4KVtCp
.

Hi all - I've been using VS Code for a week. I'm getting this error 4 out 10 times I save the file. This happens specifically with files in a SharePoint.

When I click on compare it does show me the differences, but I can't push my version over the SP one.

@fegumi2003
Can you try to reproduce with our nightly insider builds? You can give our preview releases a try from: http://code.visualstudio.com/Download#insiders

@bpasero I've been able to reproduce this using the insider version. What troubles me is that when I'm comparing the disk vs VS Code versions; discard or accept the changes does exactly the same (changes wiped out).

Ok thanks for trying, I think this needs another look then. What I fixed specifically was this error happens when you have a slow connection and manage to hit Ctrl+S multiple times without the first save actually returning. There might still be a bug that the mtime we get back after the save is not the same as the one on the server. Though still hard for me to understand how that can happen.

Regarding the issue described as "we prevent saving when we think the contents on disk are newer":

It happens to me on a LOCAL drive too, when for instance a js beautifier is triggered in the background via monitoring script.

We fixed that one for upcoming 1.9 release.

We fixed that one for upcoming 1.9 release.

Just local files?

after upgrade to 1.9 "contents on disk are newer" is still happening :(

@meuwka when you then compare with the contents on disk, are there actually changes?

@bpasero Yes, probably I found: our project use ESLint with fix in config (for example add missing spaces). Is there a possibility to disable VS Code compare mode after save file?

Currently not and not sure this is the right fix. When we find the version on disk to be newer compared to the buffer in the editor we need to deal with it one way or the other to prevent dirty writes.

understand, but now it's very time consuming process (many extra clicks) and doesn't work undoing changes after comparing :(

Yeah, I still have to find out how it is possible that the file is modified after it has been saved from the editor.

May have to dig into this myself as it is driving me nuts. Normally use local external but recently had to migrate some projects to our NAS and now I get bad file descriptor errors like every other save. About to lose my mind!! Yes this is my worthless contribution to this thread as my post has no value whatsoever other than I feel better now for whining. :)

Can you share the full error message?

People here that suffer from the sporadic error when saving and we think the file is newer ( @meuwka @rbtbar @DanZeuss): I have created a new insiders build with some additional logging to find out what is going on:

Please start the instance from the command line passing in the --verbose flag. On macOS, you can directly run the executable like this: ./VS Code.app/Contents/MacOS/electron --verbose. Once you hit the save error, please open the console (Help > Toggle Developer Tools > Console) and copy/paste the output to some location where I can access it (secret Gist is fine, the output contains the contents of the file!). Ideally you would try to reproduce with a small file to have less verbose output.

Here is a screenshot how the error would be printed like:

image

Thanks!

Hi There,
I tried the the insider version on Ubuntu 16.04. Sadly for gvfs, the error persist:

Version 1.11.0-insider
Commit 0bc62d1679fbb59a4bd8f62203743408893ba4ee
Date 2017-03-02T13:49:45.971Z
Shell 1.4.6
Renderer 53.0.2785.143
Node 6.5.0

ipc.cp.ts:152 [IPC Library: Watcher] Error: watch /run/user/1001/gvfs/sftp:host=app.test,user=vagrant/srv/www/app/htdocs/venv/lib/python2.7/site-packages/pip/_vendor/requests/packages/urllib3/packages/ssl_match_hostname/__init__.py ENOSPC(anonymous function) @ ipc.cp.ts:152e.invoke @ callbackList.ts:63e.fire @ event.ts:118r @ event.ts:12emitTwo @ events.js:106emit @ events.js:191process.nextTick @ internal/child_process.js:744_combinedTickCallback @ internal/process/next_tick.js:67_tickCallback @ internal/process/next_tick.js:98 ipc.cp.ts:152 [IPC Library: Watcher] Error: watch /run/user/1001/gvfs/sftp:host=app.test,user=vagrant/srv/www/app/htdocs/venv/lib/python2.7/site-packages/pip/_vendor/requests/packages/urllib3/packages/ssl_match_hostname/_implementation.py ENOSPC(anonymous function) @ ipc.cp.ts:152e.invoke @ callbackList.ts:63e.fire @ event.ts:118r @ event.ts:12emitTwo @ events.js:106emit @ events.js:191process.nextTick @ internal/child_process.js:744_combinedTickCallback @ internal/process/next_tick.js:67_tickCallback @ internal/process/next_tick.js:98 ipc.cp.ts:152 [IPC Library: Watcher] Error: watch /run/user/1001/gvfs/sftp:host=app.test,user=vagrant/srv/www/app/htdocs/venv/lib/python2.7/site-packages/pip/_vendor/requests/packages/urllib3/packages/ssl_match_hostname/_implementation.pyc ENOSPC extensionHost.ts:282[Extension Host] Error: This socket is closed at Socket._writeGeneric (net.js:673:19) at Socket._write (net.js:725:8) at doWrite (_stream_writable.js:307:12) at writeOrBuffer (_stream_writable.js:293:5) at Socket.Writable.write (_stream_writable.js:220:11) at Socket.write (net.js:651:40) at sendCommand (/home/user/.vscode-insiders/extensions/donjayamanne.python-0.5.9/out/client/providers/jediProxy.js:419:20) at JediProxy.sendCommand (/home/user/.vscode-insiders/extensions/donjayamanne.python-0.5.9/out/client/providers/jediProxy.js:128:16) at JediProxyHandler.sendCommand (/home/user/.vscode-insiders/extensions/donjayamanne.python-0.5.9/out/client/providers/jediProxy.js:548:24) at PythonHoverProvider.<anonymous> (/home/user/.vscode-insiders/extensions/donjayamanne.python-0.5.9/out/client/providers/hoverProvider.js:103:54)d.logExtensionHostMessage @ extensionHost.ts:282 8ipc.cp.ts:152[IPC Library: Watcher] Error: watch /run/user/1001/gvfs/sftp:host=app.test,user=vagrant/srv/www/app/htdocs/.git/index.lock ENOSPC(anonymous function) @ ipc.cp.ts:152 ipc.cp.ts:152[IPC Library: Watcher] Error: watch /run/user/1001/gvfs/sftp:host=app.test,user=vagrant/srv/www/app/htdocs/.git/FETCH_HEAD ENOSPC(anonymous function) @ ipc.cp.ts:152 2ipc.cp.ts:152[IPC Library: Watcher] Error: watch /run/user/1001/gvfs/sftp:host=app.test,user=vagrant/srv/www/app/htdocs/.git/index.lock ENOSPC(anonymous function) @ ipc.cp.ts:152 ipc.cp.ts:152[IPC Library: Watcher] Error: watch /run/user/1001/gvfs/sftp:host=app.test,user=vagrant/srv/www/app/htdocs/.git/FETCH_HEAD ENOSPC(anonymous function) @ ipc.cp.ts:152 2ipc.cp.ts:152[IPC Library: Watcher] Error: watch /run/user/1001/gvfs/sftp:host=app.test,user=vagrant/srv/www/app/htdocs/.git/index.lock ENOSPC

@pouyana great that you can still reproduce, but I am missing the output logging I added. Are you sure you got this message showing up that the contents on disk are newer than what you have? Can you show a screenshot when this happens?

Btw you no longer have to download the special build, insiders are now updating again to latest and it includes my changes for debugging.

@bpasero so i tried to run the vscode with --verbose, here is the output:
ipc.cp.ts:152[IPC Library: Watcher] Error: watch /run/user/1001/gvfs/sftp:host=app.test,user=vagrant/srv/www/app/htdocs//venv/lib/python2.7/site-packages/pip/_vendor/requests/packages/urllib3/packages/ssl_match_hostname/__init__.pyc ENOSPC(anonymous function) @ ipc.cp.ts:152 ipc.cp.ts:152[IPC Library: Watcher] Error: watch /run/user/1001/gvfs/sftp:host=app.test,user=vagrant/srv/www/app/htdocs//venv/lib/python2.7/site-packages/pip/_vendor/requests/packages/urllib3/packages/ssl_match_hostname/_implementation.pyc ENOSPC(anonymous function) @ ipc.cp.ts:152 ipc.cp.ts:152[IPC Library: Watcher] Error: watch /run/user/1001/gvfs/sftp:host=app.test,user=vagrant/srv/www/app/htdocs//venv/lib/python2.7/site-packages/pip/_vendor/requests/packages/urllib3/packages/ssl_match_hostname/__init__.py ENOSPC(anonymous function) @ ipc.cp.ts:152 ipc.cp.ts:152[IPC Library: Watcher] Error: watch /run/user/1001/gvfs/sftp:host=app.test,user=vagrant/srv/www/app/htdocs//.git/index.lock ENOSPC(anonymous function) @ ipc.cp.ts:152 ipc.cp.ts:152[IPC Library: Watcher] [ADDED] /run/user/1001/gvfs/sftp:host=app.test,user=vagrant/srv/www/app/htdocs//.git/index.lock ipc.cp.ts:152[IPC Library: Watcher] >> normalized [ADDED] /run/user/1001/gvfs/sftp:host=app.test,user=vagrant/srv/www/app/htdocs//.git/index.lock ipc.cp.ts:152[IPC Library: Watcher] [DELETED] /run/user/1001/gvfs/sftp:host=app.test,user=vagrant/srv/www/app/htdocs//.git/index.lock ipc.cp.ts:152[IPC Library: Watcher] >> normalized [DELETED] /run/user/1001/gvfs/sftp:host=app.test,user=vagrant/srv/www/app/htdocs/.git/index.lock ipc.cp.ts:152[IPC Library: Watcher] [CHANGED] /run/user/1001/gvfs/sftp:host=app.test,user=vagrant/srv/www/app/htdocs/package.py ipc.cp.ts:152[IPC Library: Watcher] >> normalized [CHANGED] /run/user/1001/gvfs/sftp:host=app.test,user=vagrant/srv/www/app/htdocs/package.py ipc.cp.ts:152[IPC Library: Watcher] [DELETED] /run/user/1001/gvfs/sftp:host=app.test,user=vagrant/srv/www/app/htdocs/package.py ipc.cp.ts:152[IPC Library: Watcher] >> normalized [DELETED] /run/user/1001/gvfs/sftp:host=app.test,user=vagrant/srv/www/app/htdocs/package.py ipc.cp.ts:152[IPC Library: Watcher] [ADDED] /run/user/1001/gvfs/sftp:host=app.test,user=vagrant/srv/www/app/htdocs/.git/index.lock ipc.cp.ts:152[IPC Library: Watcher] >> normalized [ADDED] /run/user/1001/gvfs/sftp:host=app.test,user=vagrant/srv/www/app/htdocs/.git/index.lock ipc.cp.ts:152[IPC Library: Watcher] [CHANGED] /run/user/1001/gvfs/sftp:host=app.test,user=vagrant/srv/www/app/htdocs/.git/index.lock ipc.cp.ts:152[IPC Library: Watcher] [CHANGED] /run/user/1001/gvfs/sftp:host=app.test,user=vagrant/srv/www/app/htdocs/.git/index ipc.cp.ts:152[IPC Library: Watcher] >> normalized [CHANGED] /run/user/1001/gvfs/sftp:host=app.test,user=vagrant/srv/www/app/htdocs/.git/index.lock ipc.cp.ts:152[IPC Library: Watcher] >> normalized [CHANGED] /run/user/1001/gvfs/sftp:host=app.test,user=vagrant/srv/www/app/htdocs/.git/index ipc.cp.ts:152[IPC Library: Watcher] [DELETED] /run/user/1001/gvfs/sftp:host=app.test,user=vagrant/srv/www/app/htdocs/.git/index.lock ipc.cp.ts:152[IPC Library: Watcher] >> normalized [DELETED] /run/user/1001/gvfs/sftp:host=app.test,user=vagrant/srv/www/app/htdocs/.git/index.lock

Hopefully it includes your changes!

@pouyana to find out you need to come to the point where the following error shows up and only then check for the console:

image

@bpasero Sadly, I don't see the error, but still the tab closes after clickingcrtl+s more than 2 or 3 times randomly. I uploaded the whole log, if it helps.

vscode_log.txt (console log)
terminal.txt (the log written in shell)

@pouyana but you see the error from my screenshot in https://github.com/Microsoft/vscode/issues/13665#issuecomment-283699648 ?

@bpasero no. No popups and no other errors. The tab vanishes without silently. Do I have the right insider version?

Version 1.11.0-insider
Commit 0bc62d1679fbb59a4bd8f62203743408893ba4ee
Date 2017-03-02T13:49:45.971Z
Shell 1.4.6
Renderer 53.0.2785.143
Node 6.5.0

@pouyana sorry I was not clear with my question, I wanted to hunt down the issue related to not being able to save. As for the issue of tabs randomly closing I actually have an idea for a fix or setting.

@pouyana for your case of tabs suddenly closing, I have a custom build ready that I would like to try on you that gives some more hints:

Please start the instance from the command line passing in the --verbose flag. On macOS, you can directly run the executable like this: ./VS Code.app/Contents/MacOS/electron --verbose. and when you see a tab closing out of nowhere, please send me the full output of the console.

Also let me know if you can no longer reproduce. Thanks!

@bpasero So I tested your build on my Ubuntu 16.04. The randomly closing tab problem is gone 👍 ! Sadly autocomplete and intellisense are not working. But I think it is another bug probably.

extensionHost.ts:282 [Extension Host] ---------------------------d.logExtensionHostMessage @ extensionHost.ts:282(anonymous function) @ extensionHost.ts:137emitTwo @ events.js:106emit @ events.js:191process.nextTick @ internal/child_process.js:744_combinedTickCallback @ internal/process/next_tick.js:67_tickCallback @ internal/process/next_tick.js:98
extensionHost.ts:282 [Extension Host] sendCommand jediProxyd.logExtensionHostMessage @ extensionHost.ts:282(anonymous function) @ extensionHost.ts:137emitTwo @ events.js:106emit @ events.js:191process.nextTick @ internal/child_process.js:744_combinedTickCallback @ internal/process/next_tick.js:67_tickCallback @ internal/process/next_tick.js:98
extensionHost.ts:282 [Extension Host] Error (sendCommand) This socket has been ended by the other partyd.logExtensionHostMessage @ extensionHost.ts:282(anonymous function) @ extensionHost.ts:137emitTwo @ events.js:106emit @ events.js:191process.nextTick @ internal/child_process.js:744_combinedTickCallback @ internal/process/next_tick.js:67_tickCallback @ internal/process/next_tick.js:98
extensionHost.ts:282 [Extension Host] Error: This socket has been ended by the other party
    at Socket.writeAfterFIN [as write] (net.js:286:12)
    at sendCommand (/home/user/.vscode-insiders/extensions/donjayamanne.python-0.5.9/out/client/providers/jediProxy.js:419:20)
    at JediProxy.sendCommand (/home/user/.vscode-insiders/extensions/donjayamanne.python-0.5.9/out/client/providers/jediProxy.js:128:16)
    at JediProxyHandler.sendCommand (/home/user/.vscode-insiders/extensions/donjayamanne.python-0.5.9/out/client/providers/jediProxy.js:548:24)
    at PythonDefinitionProvider.provideDefinition (/home/user/.vscode-insiders/extensions/donjayamanne.python-0.5.9/out/client/providers/definitionProvider.js:41:38)
    at /home/user/Downloads/VSCode-linux-x64/resources/app/out/vs/workbench/node/extensionHostProcess.js:12:31393
    at /home/user/Downloads/VSCode-linux-x64/resources/app/out/vs/workbench/node/extensionHostProcess.js:6:17637
    at new n.Class.derive._oncancel (/home/user/Downloads/VSCode-linux-x64/resources/app/out/vs/workbench/node/extensionHostProcess.js:6:13740)
    at Object.f [as asWinJsPromise] (/home/user/Downloads/VSCode-linux-x64/resources/app/out/vs/workbench/node/extensionHostProcess.js:6:17600)
    at e.provideDefinition (/home/user/Downloads/VSCode-linux-x64/resources/app/out/vs/workbench/node/extensionHostProcess.js:12:31347)

@pouyana that is good news, please keep me updated how it goes over the next couple of days and then I can turn it into a real fix.

Some updates to this issue:

  • before closing an editor we now do an additional check if the file is really deleted and only then close the editor (this helps for randomly closing editors after save)
  • on top of that, I introduced a new setting workbench.editor.closeOnFileDelete to disable closing of deleted files entirely (see https://github.com/Microsoft/vscode/issues/11642#issuecomment-284359219 for more info)
  • finally, I believe the sporadic "contents on disk are newer" errors when saving have also been resolved with my recent changes to the file model synchronisation, otherwise please let me know

As usual, use insider builds to benefit from these changes (http://code.visualstudio.com/Download#insiders)

Working with remote drive with VS code is actually not possible (too much bugs), I think it would be better to work on a remote explorer directly integrated with VS code.

Another problem while using network drives is that the go to file ctrl+p is very slow, pretty much unusable.

Yeah, this is really bad. At work we're using network drives with fairly complex file structures and the go to file function (ctrl + p) takes over 5 minutes to find a file.

To me, it was a problem just when working on VMWare on drives shared in
Linux to Windows, however, I'm using Docker with Virtual Box and I'm no
longer getting this error.

On Wed, Apr 5, 2017 at 3:36 AM, Patrik Thorsson notifications@github.com
wrote:

Another problem while using network drives is that the go to file ctrl+p
is very slow, pretty much unusable.

Yeah, this is really bad. At work we're using network drives with fairly
complex file structures and the go to file function (ctrl + p) takes
over 5 minutes to find a file.


You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
https://github.com/Microsoft/vscode/issues/13665#issuecomment-291767899,
or mute the thread
https://github.com/notifications/unsubscribe-auth/AXGAI3qfw4Tw3yWRQlcvShVk_wr1upq3ks5rszZagaJpZM4KVtCp
.

A month or two ago working over my network on a Samba share was _flawless_. Using Windows 10 and a Mounted Network Drive to a physical machine with Ubuntu 16.04 and Samba on the same network, everything is wired.

Now:

  • Ctrl + p is unusable (as previously mentioned)
  • Find in Folder is unusable
  • Since latest update (today or yesterday) I'm getting (deleted from disk) in tabs _constantly_

image

As @wesruv mentioned, I too am seeing "(deleted from disk)" in the tabs constantly, yet clicking on the tab causes it to go back to normal.

Sorry to hear that. We mark a file as deleted when we receive a file-delete event from our file watcher. I will check again if there are other cases where we could report this wrongly. However it seems that something is causing delete-events even though the files are not being deleted.

@wesruv can you please open a separate issue on the search issues and quick open? We recently changed our search mechanism to use RipGrep (which btw you can disable via "search.useRipgrep": false).

I have a fix in mind to get rid of the ugly "deleted from disk" issues people are seeing, stay tuned.

@bpasero I tried "search.useRipgrep": false, but there wasn't any notable change i believe.

@DanZeuss I just found out that we're using VMWare too. I haven't tried any other shared drive though.

Todays insider build contains a fix for the "deleted on disk" annotation showing up sporadically. Let me know how it goes: http://code.visualstudio.com/Download#insiders

turns out I was running into https://github.com/Microsoft/vscode/issues/20876

Seeing the content on disk is newer message too sometimes running 1.12.1 after I've edited and tried to save, mostly when working with files that have been changed due to git - seems to mostly happen with merge conflicts but couldn't reproduce in a plain repo. It would be helpful if the file was always refreshed to show its state on the disk automatically rather than waiting until after I start edit and try to save. These files are shared with a docker process and built with webpack, tho not the in-memory files of webpack-dev-server...

With the latest update, I no longer see the "deleted on disk" message.

another problem that I'm getting that may be related to when working with
network drives is that, the file search is not working. I don't know if
it's because permission

On Mon, May 8, 2017 at 7:12 PM, Paul E. Jones notifications@github.com
wrote:

With the latest update, I no longer see the "deleted on disk" message.


You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
https://github.com/Microsoft/vscode/issues/13665#issuecomment-300005619,
or mute the thread
https://github.com/notifications/unsubscribe-auth/AXGAI2lQy3mWK0Na_VblzpiHXNBBWZtRks5r35M5gaJpZM4KVtCp
.

@DanZeuss please report search issues separately, thanks.

I apologize. I thought it were some issue related only when working with
network drives.

On Mon, May 22, 2017 at 2:37 AM, Benjamin Pasero notifications@github.com
wrote:

@DanZeuss https://github.com/danzeuss please report search issues
separately, thanks.


You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
https://github.com/Microsoft/vscode/issues/13665#issuecomment-303003605,
or mute the thread
https://github.com/notifications/unsubscribe-auth/AXGAIyFR4TBVtmlbv5JQVKvE8eKsI5iuks5r8R83gaJpZM4KVtCp
.

@bpasero I'm still getting the sporadic saving error using version 1.13.1. Is the saving problem still an ongoing issue?

@AdamBJ what is your setup? and when you then compare files, is there a change to the remote file?

I'm running Ubuntu 14.04 LTS as a virtual machine (via VMWare Workstation Player). The problem occurs when I've SSHed into a machine at my school and am working in one of the folders stored on the remote machine. As for the compare, the file on disk that is "newer" is the previous version of the file (i.e. the file minus any changes I've made since opening it).

One odd thing is that the problem doesn't seem to happen unless I'm on campus -- when I remote in from home I don't run into the error.

Let me know if you want more specific details on anything.

Hi.
I prefer to mount network shares to folders.
e.g. mklink /d c:\proj \\server\share\foldername\
I was faced with a problem that if you open such folder the editor does not read its settings from .vscode at all. But with same share folder mounted as disk vscode works fine.

Is there a way to force vscode work with such folders as with a network drive? (maybe list of folders in user settings)

VsCode: 1.13.1
Windows 7

Intellisense & autocomplete don't work on the latest version with an SMB share.

@bpasero The network drives save issue only seems to happen when I'm using my laptop which is running a 32-bit version of VS Code on Ubuntu 14.04 LTS.

UPDATE: When I run into this error, it seems to happen every other save. I love VS Code, but this save issue really is killing my productivity.

Any news?

I can confirm @samirotiv's statement regarding autocomplete/intellisense not working with an SSHFS-mounted drive (on Mac). Similar to @nguyenbathanh, I'm curious what may be going on here. FWIW, the Path Intellisense extension seems to have no issue auto-completing paths with the same mounted drive. Thanks for any and all support!

@bpasero is this still the proper place to bring up related issues or would a new issue make more sense? Happy to accommodate in whatever way makes the most sense. Thanks!

Error: Failed to save 'syncSummary.txt': The content on disk is newer. Click on Compare to compare your version with the one on disk. (Settings Sync 2.8.2 on Dec 18, 2017)

@ChadRidings Is that a Visual Studio Code issue or a Settings Sync issue (https://marketplace.visualstudio.com/items?itemName=Shan.code-settings-sync). That repo is https://github.com/shanalikhan/code-settings-sync/

@paulej that is a good question. Yea I was looking at that repo yesterday as well. I've been taken down a rabbit hole where it's hard for me to determine if this is a problem with the application or a problem with the extension. I'll report back when I have a better idea.

hi:I also have similar problem, I work with samba, open a folder on remote machine, (mine is Win, remote‘s is linux), and I wanna use shell on remote machine, can VSCode integrated teminal do it?

@THCloud yes, you can, but I strongly recommend you to use the integrated terminal by SSH, I mean, connect to the server using SSH session.

Is there any update for this? As I m using it for Linux network drive and it's taking so much time compared to sublime text. Sublime text it's blazing fast.

I have problems with sshfs mounted directories in Visual Studio Code on Ubuntu, too. Autocomplete and IntelliSense don't work properly.

So I've created a small workaround that I want to share: It worked for me to copy the sshfs directory locally, watch changes in this directory with inotifywait and copy them into the sshfs directory with a small shell script. This doesn't reflect changes on the remote filesystem but this is not important for me.

Another Samba user here. I have two workflows - If I'm remote, I work with a Debian VM on Hyper-V (on my laptop). If I am at my desk, I work with a Debian VM on a VMware server that has better hardware.

In both cases, the Git features in code have very poor performance. Search is fairly performant though.

I was trying to use VS Code with my git repository on a samba share. I am using mac and the samba share is on a remote Ubuntu 18.04 machine hosted on AWS. VS code and git would not work properly when I had my repository on a samba share but I had to get work done so I went ahead and started working without git. I was editing a file that was 2 days old, hours of work, and VS code hanged. When I reopened VS Code I was shocked to see my file on the server was blank. In fact, every file I had open at the time of the crash was blank. Now if git would have been workign properly I would not have lost data.

I guess it is back to Dreamweaver? Before I do somethign crazy like that, I am trying also a NFS. Git hub is still having issues. but I got 2 commits to go through. I am hacing trouble with more commits. It keeps telling me I can not write and index.lock file. Prob because VS code is unable to write to the network drive

I am closing this issue and ask people to file individual issues with reproducible steps after searching for duplicates. Having one large issue turns out to be not very helpful to manage.

Was this page helpful?
0 / 5 - 0 ratings