MayaMayaMayaFeb 27
Brooke

I've just published a reference for anyone using git while trans, explaining how to hide or remove your deadname in git repositories

Please share ❤️

https://code.curly.kiwi/2026/02/27/using-git-while-trans/

Using git while trans | Brooke's code blog

A guide for hiding or removing your deadname in git repositories.

Feb 26 at 11:25pmWeb
Show threadmszegedyFeb 27
@nerdnz this would have been really nice to have in like 2019. asynchronously thank you
Show threadEmily_SFeb 27

@mszegedy I was just about to post something similar, including the year :D

Really helpful, thank you @nerdnz

Show threadIrenes (many)Feb 27
@nerdnz thanks for this! it's nice to know the situation has improved since when we had to do it, and your writeup is super readable and will doubtless help people
Show threadJackie JudeFeb 27
@nerdnz the actual reason why my website jackiejude.me isn't open source is because the git history contains my deadname oops
Show threadvfigFeb 27
@nerdnz thats a really good writeup! simple, easy to follow, and clear about the pros and cons of each approach.
Show threadEQFeb 27
@nerdnz Also useful for people that are less than happy with sharing their real name online. Thanks!
Show threadferricoxideFeb 27
@nerdnz Was assuming there was going to be commit-rewriting: would that change all of your commit-history's dates? How well would this work for a repository you didn't own (or otherwise has branch- or similar "protection" enabled)?

While I'm not trans, seeing solutions like this always makes curious on details, N-order effects, etc..
Show threadBrookeFeb 27

@ferricoxide The history rewriting approach I've mentioned doesn't change commit dates - I wanted to make sure I found something that preserved as much metadata as possible

Rewriting the history like that is unlikely to be viable on work projects or open source ones with lots of collaborators, but it is a good option to know about for things like personal projects

Show threadferricoxideFeb 27
@nerdnz

Meant commit IDs. I've seen attempts at removing committer information from projects end up re-IDing everything. Thus my curiosity.

I've been lucky enough that I've generally been able to not use name-based emails for public projects.
Show threadBrookeFeb 27
@ferricoxide Yeah rewriting history in _any_ way means the old commits are gone and replaced with fresh new ones. Commit metadata like author name is used for generating commit hashes so there's no way to modify it without reconstructing a brand new commit. But it is possible to do that in a way that preserves the commit date part of the metadata
Show threadCarlos SolísFeb 27
Makes sense that the only way to retroactively rewrite history in a repository is to, indeed, retroactively rewrite all the relevant commits and force-push them. At least you'll generally have to do it only once in your lifetime, right? ...Right?
Show threadBrookeFeb 27
@csolisr Depends on how many times you change your name I suppose. In my experience, force pushing a change like this in any particular project is either completely unviable or not a problem at all. I'm not sure there are all that many cases where it's somewhere in between
Show threadMalte Lau PetersenFeb 27
@nerdnz I'm a collaborator (and cis ally) on a big open source project with a lot of non-technical committees who would not be able to handle it if we rewrote history on them. We also have a good amount of people who contributed both before and after transition. Do you think it would still be helpful to inform/suggest the mailmap thing?
Show threadBrookeFeb 27

@maltelau Yeah history rewriting is usually only really viable on personal projects and the like

Being aware of mailmap as an option is certainly better than not being aware. It definitely does have its own downsides, which I hope I laid out clearly, but it's something individual contributors could make a call on based on their own priorities

Show threadCassandrich5d ago
@nerdnz @maltelau As a maintainer of a project that can't have history rewritten and that might have deadname-credited commits in history, it's important to make sure anyone who asks for mailmap treatment understands the part about how it does make them stand out, and does make an immutable record mapping their deadname to their current name, before just adding them to mailmap and potentially making things worse.
Show threadRyan J. YoderFeb 27
@nerdnz
Oof, this sucks. Sorry to anyone who has to rewrite git history.
Show threadsarah tonin Feb 28
@nerdnz this is one of the biggest issues in git imo, identity is considered immutable, which is just not realistic
Show threadBrookeFeb 28
@SRAZKVT Yeah, it's a safety issue and it's really not treated like one. Hard to go back in time and change that early design decision though, can't rewrite that history unfortunately
Show threadsarah tonin Feb 28

@nerdnz and when the history is shared with other people it leads to problems

there is really just no win

Show threadFrost, wolf of winter 🐺🎄Feb 28

@nerdnz Cool!...oh you recommend filter-repo

Careful with filter-repo

It tries to trash your repository and then remove all the safety backups.

That's why it's so adamant about "run it on a fresh clone!!!!!!!!!!!!". What if there IS no fresh clone? What if there is no remote? Then it falls over horribly, and leaves you with a broken project.

filter-branch is better.

Show threadFrost, wolf of winter 🐺🎄Feb 28
@nerdnz At least, it used to do that back when we tried it a couple years ago. I immediately swore it off and started using FILTER_BRANCH_SQUELCH_WARNING=1 instead.