#Git now considers a symlink to "git~1" something git fsck should warn about, on Linux, because on Windows, that happens to be a way to refer to the .git directory.
rotfl
joey@darkstar:/tmp/foo>ln -s git1 foo
joey@darkstar:/tmp/foo>git add foo
...
joey@darkstar:/tmp/foo>git fsck
Checking object directories: 100% (256/256), done.
warning in blob 086654741cca64b19ba88286fce998d933897870: symlinkPointsToGitDir: symlink target points to git dir
Note it says, "As always, it is best to avoid cloning repositories from untrusted sources."
Now consider how many languages will clone a repo of a transitive dependency, or direct dependency at time of dep resolution -- often before any dep analysis/presentation tools could give you a means to evaluate the transitive git deps.
Like #clojure, one of my favorites. I have not review dep resolution/fetch code in other languages, but it seems that they should all heed the advice in that advisory.
GitNex 5.5.0 is now OUT with a new comment UI, the ability to update your profile avatar, the option to insert notes to issues/PRs/releases, and many other improvements.
We have also reached 400 stars on #codeberg. Thanks to all for supporting GitNex all the way.
With the upcoming 6.0.0 release, GitNex will have different version numbers, such as 6.0.0, 7.0.0, and so on.
I hope you enjoy the new release and find it even more productive.
This is detailed, step-by-step instructions of getting this set up, with screenshots to help with commands and outcomes. An "idiot's guide" if you like, because I was that idiot.
Right now my "default" is to go to GitHub, and then maybe GitLab and Codeberg...but that leaves out all the other hosts out there. Once #ForgeFed happens we'll need it even more :fedi:
(By the way, I say "git" because that's where most code is, but if other kinds of repos are supported, even better! 🔍 )
I wish that when I reverted a revert, #git was smart enough to attribute things to the original author.
That is, person A does a commit; git blame foo.c shows the lines as being changed by person A. Person B reverts that commit (temporarily), and then later on person B reverts the revert (thereby putting person A's changes back in). But now when you git blame foo.c, the line change is from person B.
[I know you can manually set --author or whatever, but I feel like it should be automatic.]
A clean Git history is the key to successful teamwork and quick bug fixes. Errors can only be successfully tracked down if it is always possible to trace when and where code was changed by whom and for what reason.
🥴 However, in the rush of the battle, the changes that are packaged in a commit are sometimes not taken very seriously. Who has never experienced this? A change that is actually unrelated to the current work package has made it into the commit because the file has already been saved temporarily.
💡The solution: With an "interactive add" (git add -i), you can pack partial changes ("hunks") into a commit and specify line by line what should be included in the next commit.
I know, Git is a mess. But, since we're stuck with it, we may as well try to learn how it works with resources like this, which aims to lead to some form of Git enlightenment.
Highlights: Haskell instead of Scheme, JSON instead of sexps, SSH instead of OpenPGP, additional features such as per-file authorizations and unsigned merge commits.
Dear #logseq and #obsodian users, #git is not a backup system. Store your notes in a reposotory, if you must, but please configure a backup for your notes.