Every time I do a git pull from SourceTree, I receive the following message:
git -c diff.mnemonicprefix=false -c core.quotepath=false -c credential.helper=sourcetree submodule update --init --recursive
No submodule mapping found in .gitmodules for path 'server/simple_server'
Completed with errors, see above
However, preceding the above message, the pull works as expected and everything is fine (code pulls, auto-merge, etc works as expected). This happens for every repo I've cloned using SourceTree. If I use the CLI to do a git pull, no submodule checks are performed and so I never see the above ('error') message. Why is SourceTree performing this extra step in the git pull (which does not happen from the CLI). I've gone through every possible option in SourceTree preferences.
I'm running SourceTree on a Mac now, but I had a (Windows) PC last week (and prior) running SourceTree and received the same errors – this appears to be a (default) feature of git pull in SourceTree, regardless of platform. And, once again, only happens via the GUI, not CLI – on all of my cloned repos.
Any advice would be appreciated. I hate seeing a false negative in my git pulls while using SourceTree.
Related
I just created a new repository on GitHub and trying to initially push my local Java project.
I added the ssh key in my GitHub account, and it's referenced correctly by Eclipse, I assume.
I only get this error message when I'm trying to push:
Can't connect to any repository:
git#github.com:MaximStein/MyProject.git
(git#github.com:MaximStein/MyProject.git: Missing unknown
366362as232d670123a2267b4879bbd01d142426)
Any ideas?
Your Git repository seems to be broken or at least in an invalid state.
The root cause seems to be Missing unknown 366362as232d670123a2267b4879bbd01d142426 which means something (probably the HEAD) points to the object with the hash 366362as232d670123a2267b4879bbd01d142426 (which will be stored in the file .git/objects/36/6362as232d670123a2267b4879bbd01d142426) that does not exist.
In the Git History you might right-click the last commit and choose Reset > Soft (HEAD Only). If commits are missing, you might be able to recover them via the Git Reflog view.
Make also sure your Eclipse and EGit/JGit is up to date (for instance, there was an issue causing a Missing unknown error that has been fixed more than four years ago) and that you do a refresh (F5) in Eclipse after you have executed Git operations on the command line.
Try first if the push would succeed from command line (which means you need in to install Git first):
check you are correctly identified with ssh -Tv git#github.com
check your remote origin URL: git remote -v
check your local status (git status: to make sure you are on a branch, here I assume main, and that you have made at least one commit)
check your initial push: git push -u origin main
Check also if you have declared any submodules (.gitmodules file), or look for any nested Git repository (.git subfolder anywhere beside your repository root folder)
It is helpful to rule out (or not) a git-core issue, before investigation an Egit/JGit one.
And the OP Maxim confirms in the comments:
It's a good idea to try it in the command line first.
It says "fatal: bad object HEAD" when I do git status.
I think I've only made one local commit for that project.
Indeed, making local commits, in a branch, is a prerequisite to pushing anything.
In your case, do a git branch -avv to check if there are any.
If you already pushed that missing commit, assuming your main branch is main (not master):
git fetch origin
git reset --soft origin/main
I am trying to update submodules in a repo. I have successfully cloned the repo and used git checkout to successfully switch to the correct branch I need to be on. When I run git submodule update --init -–recursive in that branch I get the following:
error: pathspec '–-recursive' did not match any file(s) known to git.
I have previously followed this process on a Macbook pro and did not get an error, but am currently getting the error on Windows. I have the latest version of Git installed (2.32.0). I first considered that this could be a Github SSH key issue, but I have confirmed that it asks for my SSH passphrase and proceeds with the cloning to my local machine, so probably not security related. The Git documentation states that this submodule update is mostly foolproof, so I'm unsure where to go from here. I've also tried running git submodule foreach git pull origin master before the submodule update
and same error. Any assistance is appreciated.
pathspec '–-recursive' means it interprets what should be an option as a filename/filepath instead.
Double-check your '-' (in case they are 'minus' instead of the regular 'hyphen-minus')
Make sure you are using the latest Git for Windows (2.33.0 from today)
Yesterday when I was at work, I was working on my project that i have on my home computer. The problem was when I tried cloning the repo it eclipse would crash. So instead I downloaded it manually and initialized the repo and then force pushed to my home computer. This deleted all the history and files in on the repo. I did some googling and tried reverting the commit but it the force push erased the commit history so I can only go back to the force push commit.
I have a computer at home which hasn't pulled the updates since this mistake and I was able to get its commit SHA. I used that on the web and found the last version its still there but I can't pull that commit because it got "deleted" and won't show in the repos history. Is there a way I can use the SHA from the last good commit and make that the master branch restoring my history?
When changing the HEAD (the working position in the git repository), it tracks your history in the reflog. Executing git reflog in the command line will show your complete history. You can read more about it on git-scm.com
When you have the desired commit SHA, you can do a git checkout with it. Executing git checkout [SHA1] will take you to the desired position in the git history. From there, you can make a new branch from it by running git checkout -b "new_branchname".
If you want to set your current branch to the commit, you can run git reset --hard [SHA1]. This would reset the current branch, but the previous state can however be found in the reflog.
I do NOT recommend doing this as general practice BUT...You should be able to force push from the other computer, thus resetting the repo to its former state. Providing you don't want to keep anything you did while making a giant mess of your git repo yesterday.
Since you know the right version you want to recover, it will make things easier.
In your local repo, use below commands:
git reset --hard <commit sha value>
git push -f
Note: if you execute the commands in the local not initialized (not cloned from remote), you should fetch first by git fetch --all, and then execute the two commands .
Now the version will go back to the latest as you need both locally and remotely.
I uploaded my whole project on my repository of github yesterday.
I'd like to update the online version today, and when I use git push -u origin masterorder, the bash window says:
! [rejected] master -> master (fetch first)
So how do I "fetch"?
If I use git pull first, would my local files be overwritten by the online version?
So how do I "fetch"?
By using the command:
git fetch
Basically, your local repo is out of sync with the remote (Github) repository. If you have multiple remote repos, you can specify which one, and you can also specify which branch. Try
git help fetch
for a complete description of the command and the various parameters you can pass in.
If I use git pull first, would my local files be overwritten by the online version?
git pull is like doing a git fetch followed by git merge -- that is, it gets the updates from the remote and attempts to merge those into your local files. That may be fine, or not, depending on what you intend. Read the help for both commands so that you can make an informed decision.
This is the online reference I am using. However when I enter the remote repository and hit publish I get an error
You cannot publish local branch master to the remote repository origin
because a branch with the same name already exists there. You might
want to rename your local branch and try again.
I dont want to rename anything as yet. I want the first version of the code to go into the master branch and at the same time create an upstream tracking branch locally.
Then I tried using the command line only. After creating a fresh repository, I ran the following commands
$ cd (project-directory)
$ git init
$ (add some files)
$ git add .
$ git commit -m 'Initial commit'
$ git remote add origin <url>
$ git push --force
This worked but now I have all my obj,bin,packages directories along with it. I don't want all those directories to go. Editing the git ignore did not help either. What I am doing wrong? Its taken more twice the amount of time to get the code into the repository than it has taken to write the code!
I only want my sources in the repository and have a tracking local branch to begin with.