I am developing on a Linux machine. It happened once that I added a file 'sin(x):2.tex' or something similar. This caused some trouble for a friend who uses a Windows system and wanted to clone my repository.
I know there are much more invalid file names on Windows than on Linux. Is there a possibility to make git warn me if I try to add such an invalid file name to my repository?
You can see a good example of pre-commit hook in this gist, in reference to this answer about valid Windows names
# A hook script to check that the to-be-commited files are valid
# filenames on a windows platform.
# Sources:
# - https://stackoverflow.com/a/62888
# - http://msdn.microsoft.com/en-us/library/aa365247.aspx
#
# To enable this hook, rename this file to "pre-commit", move it to ".git/hook" and make it executable
It uses git diff:
git diff --cached --name-only --diff-filter=A -z $against
Note that a pre-commit hook is a client-side hook, meaning it has to be deployed in all repos for all users.
And it can be bypassed (with git commit --no-verify).
Another approach is to set a pre-receive hook (server-side hook) which will block any push including invalid filenames.
Related
I have a VPS running a Node.js / React app.
I want to update the files in the vps each time I push data to the git(hub).
I found out, using this answer, that I can add some hooks in git, executing commands on "post-receive".
What I didn't quite understand :
Why did he init another git ? Couldn't he have done this in the .git directory and create the hooks/post-receive file?
Why git checkout -f ? If the goal is to update local files, so nodemon / create-react-app restarts the server / app, why not execute a git pull instead ?
Is there a better way of doing this ?
In the recommended answer there, nobody is using GitHub and there is no other Git repository yet. So the answer to your question:
Couldn't he have done this in the .git directory and create the hooks/post-receive file?
is: No, there was no .git directory in the first place. The target machine had nothing at all, no Git repository, no working tree, etc. The git init --bare created the Git repository (the ".git directory").
The git checkout -f is a poor-man's / low-quality implementation of push to deploy. A receiving repository is normally "bare", and this one is no exception.
why not execute a git pull instead ?
That would require creating a third Git repository. That would have been an option.
"Better" is in the eye of the beholder. There are many ways of doing this, each with its own pluses and minuses. See also Deploy a project using Git push, which notes that since Git 2.3, receive.denyCurrentBranch = updateInstead is available; it was not available prior to 2015 (and in 2015, many people had older versions of Git installed).
Note further that if you're using GitHub as a hosting system, this changes a number of variables. The questions and answers you and I have linked here are aimed at those not using GitHub.
It is possible to track one file, but discard changes to it:
git update-index --assume-unchanged <file>
But this is local setting.
How can I make this visible to my team members, without forcing them to locally do the git update-index --assume-unchanged dance?
The answer is that you can't. Git doesn't support doing this and using git update-index --assume-unchanged for this purpose is not a supported use of that feature.
If you have a config file or other file that needs to change depending on the user or system, then check in a template and generate it as part of a local build step. For example, GitHub uses templating and a set of scripts to set up local configuration in a project-independent way.
I am on a windows machine and my .java files are stored on my disk (C:\Users\myname\workspace..etc)
I have a git repository set up in a cloud and I SSH to it using putty.
My question is, how do I add the files in my workspace to the repository?
When I do git add filename.java the linux terminal says it cannot find specified file.
So do I some how bring the the copies of the java files into this terminal directory?
Btw I'm using an Atlassian Stash on Amazon Web Services.
Go to current working directory to your local project and run following commands:
Step 1 : Initializes a new Git repository. Until you run this command inside a repository or directory, it’s just a regular folder. Only after you input this does it accept further Git commands.
git init
Step 2 : This does not add new files to your repository. Instead, it brings new files to Git’s attention. After you add files, they’re included in Git’s “snapshots” of the repository.
git add .
Step 3: Check the status of your repository. See which files are inside it, which changes still need to be committed, and which branch of the repository you’re currently working on.
git status
Step 4: Git’s most important command. After you make any sort of change, you input this in order to take a “snapshot” of the repository. Usually it goes git commit -m “Message here.” The -m indicates that the following section of the command should be read as a message.
git commit -m "Committing Files - Message"
Step 5: Add origin to the .gitconfig file.
git remote add origin <url to your git repo.git>
Step 6: Push the changes from your local repository.
git push -u origin master
I have an old Synology DS-106j server where I tried to install git using ipkg command. The installation went smoothly, but git failed to work correctly. I am currently learning how to use git, so I don't know if it is a bug from git with the version I am using or something else is wrong.
What I did was create a new local repository with a specified name, add a new file, commit it, and got an error:
NAS_SERVER> git init Test
Initialized empty Git repository in /root/Test/.git/
NAS_SERVER> ls
Packages.gz git_1.8.4.2-1_powerpc.ipk
Test
NAS_SERVER> cd Test
NAS_SERVER> git status
# On branch master
#
# Initial commit
#
nothing to commit (create/copy files and use "git add" to track)
NAS_SERVER> touch Test.cs
NAS_SERVER> ls
Test.cs
NAS_SERVER> git add *
NAS_SERVER> git status
# On branch master
#
# Initial commit
#
# Changes to be committed:
# (use "git rm --cached <file>..." to unstage)
#
# new file: Test.cs
#
NAS_SERVER> git commit -m "Test"
fatal: 57e2b8c52efba71d84c56bf6f37581686b9061a3 is not a valid object
I thought...maybe I did something wrong, so I used git on Windows OS and try a push. Still an error. Transfer the whole repository to the server and check the status. It seems fine. Try a commit. Still the same result. What worse is that I can't update git version without having to compile it, which I don't even know how to do so. Any suggestion to what might be wrong?
If your goal is to push into a git repo located on the synology disk(s) for backup purposes I'd recommend a different approach which would avoid having to install a rather old git version on the synology box itself (which could lead to problems if/when using a newer git version on the windows machine).
Export a samba share from synology, mount it on windows and use the windows git to create the backup repo (maybe even a bare repo, eventually group shared if you plan to share work with other people). Then push from your working repo into this backup repo - all on the windows box. In this scenario the synology box doesn't need git installed, it just serves files (i.e. its original job).
I'm using such setup but with a linux machine instead of a windows one and with the bare repo on the synology disks exported via NFS instead of Samba.
Recently I migrated from opensuse to centos and after that GIT has started to ignore my custom commit-msg hook. It simply doesn't execute it. (I checked it by add small piece of code to "add_ChangeId" function )
Hook generates Change-Id hash for every commit
GIT version: 1.8.1.2
File is located in following location: .git/hooks/
For debugging purposes I even have set 0777 permissions to the whole .git directory
Here is the full text of commit-msg file - http://pastebin.com/zmYNi0ED
timoras you are gold. Then I tried to execute script using sh .git/hooks/scriptname it worked, but when tried to call it using .git/hooks/scriptname and shell returned that I haven't permissions to execute it.
After that I looked at fstab, and found out that have forgot to add exec flag to the partition where this file was located.
Now everything works.
One more time thanks timoras!