Github failed to sync branch - github

I'm using W7 64 bi , and just got an error from the github client app. It says:
failed to sync branch. you might need to open a shell and debug the state of this repo.
What do I do now ?

I know this will sound crazy, but try restarting your computer.
This happend to me yesterday; I was getting this error, and upon checking: \AppData\Local\GitHub\TheLog.txt
I found messages like:
AppData\Local\GitHub\PortableGit_ca477551eeb4aea0e4ae9fcd3358bd96720bb5c8\bin\sh.exe:
*** Couldn't reserve space for cygwin's heap, Win32 error 0
The problem comes around to GitHub for Windows updating itself (in particular the cygwin-ized PortableGit) in the background while I was using it. Ultimately, some cygwin dlls from the previous PortableGit dlls were still loaded in memory causing errors when trying to execute the new (updated) PortableGit commands.
Restarting cleared out all the previously loaded cygwin dlls.

You might have a more complex problem than my answer can solve, but -- for anyone else who comes across this question -- some basic solutions can be found in this video. In short, status.github.com, git status, and gitstatus are your friends. See what they tell you and then continue your sleuthing with that new information in mind. You can use these tools using Git Shell that comes with the Github Windows client.
I'll note that my own problem stemmed from trying to sync a file that was too large: I only found this by using the Git Shell, which gave me the error when I tried git sync. I'm currently looking for ways to remove the file in question in previous commits so I can sync my repo appropriately. The guide I am currently following can be found here, and if it seems to be taking me in the right direction (it was actually recommended in the Git Shell error message!)

I had this error before, and I fixed it by reinserting the credentials for my github account. Turns out that github logged me out for some reason

Related

A fatal error was encountered. The library 'hostpolicy.dll' required to execute the application was not found

I tried internet search on this issue and it is very common with many suggestions. The problem started with an update to VS2019 in which I paused because my projects in VSCode were breaking.I did a system restore and when I tried to start VSCode I got a message saying my installation was corrupt. So I reinstalled to the latest version and now get this error.
A fatal error was encountered. The library 'hostpolicy.dll' required to execute the application was not found...
I found 3 extensions that needed updating so I do so but did not fix this problem. Do not know if I need to update the .Net Core since it looks like a version mismatch. It seems these updates are wreckless since they always cause thing to go wrong.s
If you are facing this problem, it means you are using Windows.
This official microsoft live share github page explains how to fix the issue: https://github.com/MicrosoftDocs/live-share/issues/102
Especially, check out this comment: https://github.com/MicrosoftDocs/live-share/issues/102#issuecomment-366509735
make vscode is up to date
uninstall live share from vscode
close vscode completely
restart vscode as administrator
reinstall live share
I should add:
close completely vscode and restart it

Asset Catalog Compile Error with On-Demand Resources: has no output specification

I've been trying to get On-Demand Resource to work but I keep getting this compile error:
/* com.apple.actool.errors */
: error: The tag combination "tagName" for "xxx.imageset/xxx#3x.png" has no output specification.
I had a look at actool man page and there's an option:
--asset-pack-output-specifications filename
Which says:
Tells actool where to write the information about ODR resources found
in the asset catalog. The emitted file will be a plist.
But I'm not really sure what to put as an argument/where this plist is used or even if this option is in the right track of fixing the error.
My coworkers and I struggled with this error for over a day and were only able to fix it by wiping our existing local repos and installing a fresh clone from our remote repos with the code that contains the on-demand resources.
In our case, I was the one that created the on-demand resources functionality and did the tagging for the assets. I built and ran all of that code, and everything worked fine locally on my machine. I pushed those commits to our remote, and when my coworkers pulled they received the asset catalog compile error that you reported when they tried to build.
I compared my build logs with those of my coworkers and found that I had the --asset-pack-output-specifications flag along with a filename whereas they did not, even though all of our production code was the same. I never set that flag manually myself during development, it was automatically generated at some point in the process but I have no idea where -- I didn't even know it existed until this build failure occurred. After struggling for many hours we noticed that if my coworkers deleted their local projects entirely and basically started fresh by installing a new project and repo again from the remote, they suddenly were able to build. They had already tried to clean and nuke their derived data, but that didn't work. Only totally deleting the repos and the projects entirely did the trick. Not sure why, but something about wiping the project and all associated directories and building themselves totally fresh from their own local machines triggered something that enabled the --asset-pack-output-specifications flag.
I just faced this issue and was totally against deleting my repo and cloning again.
I noticed that alongside this error, I also got a warning stating that I had assets under the same name (thus being duplicated).
Deleting the duplicated asset in order to get rid of the warning, fixed the compilation error.
Hope this helps someone, as deleting the repo and cloning again shouldn't be an option.
I just faced the issue.
I could solve it just by deleting on-demand resource tags and tagging them back again.
I just solve it by rebooting Xcode then run successfully. The error seems to only appear once...I don't know why.

Eclipse EGit pushing latest commit hangs plink.exe

Yesterday I setup EGit and PuTTY so I can share my latest changes to the code to my RPi which then packs my binary into a executable jar, and with quiet alot of extra work, I got it working to some extent, I could successfully push my local resipotory to github, and my RPi could fetch the latest commit if needed. But I still had some minor problems with my RPi executing my code and compiling etc.
However, today everything suddenly stopped working, I could no more push my local resipotory to github for whatever reason and I have just not been able to figure it out (the funny thing about ALL this is that I thought it would be very beneficial if my RPi could stay up to date with my latest code so I wouldn't have to manually export my code to the RPi by remote desktop and dropbox, but it turned out to be even more complicated than I thought it was).
The most annoying thing is that it was working just fine yesterday, I did change a few things in my project within eclipse but other than that, everything should've been the same. Basically EGit first told me that the server host key was not cached. And I resolved that by calling plink git#github.com from the command line. After that I was left with another problem: PuTTY Fatal Error: Disconnected: No supported authentication methods available, and after Googleing the problem for a good hour or so, I found out that Pageant had apparently no private key in the Pageant key list. So I had to convert the id_dsa private key generated by eclipse to a PuTTY Private Key File, which resolved that issue, after launching EGit again, I first thought it would work, but oh no! I've never been so wrong in all my life. I tried to push my local resipotory to github, but it got stuck at getting remote branches information, so I tried to cancel it. But it didn't work, I couldn't close the wizard due the process had to be cancelled first. So I had to manually TERMINATE the plink.exe process to close the wizard.
I would be so thankful if anyone could help me out here. Thanks in advance!

Eclipse 'Unhandled loop exception' on Team > Commit

I recently started using git with the egit plugin in eclipse, and I thought I had everything working. I cloned a remote repository to my machine, then made a change to a file, and went to commit the change via selecting the file in the project, and going to Team > Commit.... As soon as I do, I get the following error:
Unhandled event loop exception
org.eclipse.swt.SWTError: No more handles [Unknown Mozilla path (MOZILLA_FIVE_HOME not set)]
at org.eclipse.swt.SWT.error(SWT.java:4308)
at org.eclipse.swt.browser.Mozilla.initMozilla(Mozilla.java:1826)
at org.eclipse.swt.browser.Mozilla.create(Mozilla.java:687)
at org.eclipse.swt.browser.Browser.<init>(Browser.java:96)
...
(output abbreviated)
I am fully willing to accept that this is a bug of sorts (even though I thought this was working earlier today), but I am hoping someone may have encountered this problem once before.
Any insights? Even if it is only who I should file the bug against? I can provide any additional details. I should also point out that things work fine if I commit through the git repository view.
The problem is that the particular git workflow is creating a SWT Browser component that is having trouble loading on your particular OS + Windowing system + mozilla library version combination. There is a bit of black magic that has to go on to make sure you get the right configuration to allow the SWT browser to work with your embedded mozilla renderer. You can get detailed information on the SWT FAQ page here.
Also there are many other posts on this site that talk about how to resolve the dreaded "NO MORE HANDLES" problem here.

Eclipse egit: "Packfile corruption detected: Unknown zlib error." How to circumvent?

We have a git repository managed by gitosis under Ubuntu, which has worked well all up to the disk ran full. After reading up a bit on the issue, I found that git gc and git gc --aggresive got me quite a bit of diskspace back. Very nice.
Unfortunately this appears to have broken something in egit, as I get this message when trying to clone our repository (during the checkout phase at around 10%)
Packfile corruption detected: Unknown
zlib error.
Interestingly enough the git in msysgit works just fine as before.
I tried upgrading egit to the nightly build of 0.12 as there was some mailing list messages hinting this had been fixed within the last week, but to no avail.
My question now is, what can I do to my repository to get to a state where egit works again? I have full control over the Ubuntu instance running gitosis.
EDIT: I got a stack trace from the Eclipse event log
org.eclipse.jgit.errors.TransportException: Packfile corruption detected: Unknown zlib error.
at org.eclipse.jgit.transport.BasePackFetchConnection.doFetch(BasePackFetchConnection.java:287)
at org.eclipse.jgit.transport.BasePackFetchConnection.fetch(BasePackFetchConnection.java:225)
at org.eclipse.jgit.transport.FetchProcess.fetchObjects(FetchProcess.java:214)
at org.eclipse.jgit.transport.FetchProcess.executeImp(FetchProcess.java:149)
at org.eclipse.jgit.transport.FetchProcess.execute(FetchProcess.java:111)
at org.eclipse.jgit.transport.Transport.fetch(Transport.java:903)
at org.eclipse.egit.core.op.CloneOperation.doFetch(CloneOperation.java:228)
at org.eclipse.egit.core.op.CloneOperation.run(CloneOperation.java:135)
at org.eclipse.egit.ui.internal.clone.GitCloneWizard.executeCloneOperation(GitCloneWizard.java:259)
at org.eclipse.egit.ui.internal.clone.GitCloneWizard.access$3(GitCloneWizard.java:252)
at org.eclipse.egit.ui.internal.clone.GitCloneWizard$4.run(GitCloneWizard.java:233)
at org.eclipse.core.internal.jobs.Worker.run(Worker.java:54)
Caused by: org.eclipse.jgit.errors.CorruptObjectException: Packfile corruption detected: Unknown zlib error.
at org.eclipse.jgit.transport.PackParser$InflaterStream.read(PackParser.java:1530)
at org.eclipse.jgit.transport.PackParser$InflaterStream.skip(PackParser.java:1500)
at org.eclipse.jgit.util.IO.skipFully(IO.java:203)
at org.eclipse.jgit.transport.PackParser.inflateAndSkip(PackParser.java:1352)
at org.eclipse.jgit.transport.PackParser.indexOneObject(PackParser.java:834)
at org.eclipse.jgit.transport.PackParser.parse(PackParser.java:448)
at org.eclipse.jgit.storage.file.ObjectDirectoryPackParser.parse(ObjectDirectoryPackParser.java:178)
at org.eclipse.jgit.transport.PackParser.parse(PackParser.java:410)
at org.eclipse.jgit.transport.BasePackFetchConnection.receivePack(BasePackFetchConnection.java:649)
at org.eclipse.jgit.transport.BasePackFetchConnection.doFetch(BasePackFetchConnection.java:280)
... 11 more
Edit: Opened https://bugs.eclipse.org/bugs/show_bug.cgi?id=340305
One way to get a clone without EGit is to:
make a bundle on the server
copy the unique file representing that bundle on the client side
try to clone it in command-line.
Since you mention you can process the bundle with msysgit, that suggests a bug on the Egit or JGit side, as illustrated by bug 330758.
The usual course of action is to update to the nightly latest fo EGit, using this p2 update site, and see if the problem is still there.
If the issue persists, you can then file a bug report or complete the existing one (330758).
Please check the versions of Git you are running and make sure they match.
The latest egit 0.12 does not show this behaviour.
Hopefully the bug has been fixed for real, and not just an accidental side effect.