git/list[1] front-page[2] threads[3] people[4] search[5] about
 

Re: [BUG] git checkout <branch> allowed with uncommitted changes

From
Andreas Ericsson <ae@op5.se>
Date
Oct 13, 2011, 12:22 UTC
Message-ID
<4E96D819.20905@op5.se>
In-Reply-To
<loom.20111013T130924-792@post.gmane.org>
On 10/13/2011 01:51 PM, arQon wrote:
Show 29 quoted lines
> Snipping the bug and focusing on one of the after-effects of the bug is,
> unfortunately, not helpful to me unless I'm missing your point (which is
> certainly possible).
> 
> git switched branches while there were uncommitted files. It's not supposed to
> do this, ever, unless given -f or -m, and it broke the tree as a result. Even
> *with* -f or -m, the behavior I described is incorrect.
> The git docs seem to agree with me, which is why there's git stash. If the docs
> are wrong, fine, though it seems pretty strange to have a change on BranchA
> appear by magic "in" BranchB without any merging.
> 
> What I'm after is an understanding / explanation of how something that isn't
> supposed to happen, does. I don't care if it's "Because I'm an idiot", "Because
> git is broken", or even "Make sure your config has 'git.makebranchesworkproperly
> = true' in it, the default is false". If there is no explanation for why git
> switches branches when there are still uncommitted files, and there doesn't seem
> to be, then it's a pretty catastrophic bug and fixing it would be a Good Thing.
> 
> *AFAICT*, committing *a* file is what triggers it.
> If you commit -a, which is what all the commits prior to this were, it works
> properly. You change branches, and the files on the disk become what they should
> be.
> If you commit nothing, you correctly get the "uncommitted files" error.
> If you do a partial commit though, your tree breaks.
> 
> Like I say, if the man page, quote:
> "If you have local modifications to one or more files that are different between
> the current branch and the branch to which you are switching, the command
> refuses to switch branches in order to preserve your modifications in context."

This means that if fileX on branchA is different from fileX on branchB and you *also* have local modifications to fileX, git will refuse to switch branches. If, on the other hand branchA:fileX == branchB:fileX and you have modifications to fileX in your work tree, there's no reason to refuse the branch change. Partly because nothing will be lost and partly because you can just switch branches back if you decide you've switched branches before committing things to the first branch.

Show 7 quoted lines
> is wrong, and this behavior is deliberate, that's fine. Bizarre, but fine in
> the sense that git is doing what it's supposed to (regardless of how
> counterintuitive and destructive it is).
> If the man page is right though, this is a bug. Maybe it's only in msysgit,
> but this is the second time it's happened, so hopefully it's fairly easy to
> reproduce.
> 
It's not a bug. You just read the manpage a bit wrong.

Consider this scenario: $dev works on featureA on branchA, modifying fileX, fileZ and fileY and then does a commit of fileZ and fileY, but realizes that the changes in fileX will be good for developing featureB as well, so he changes to a separate branch to do the update to fileX and be able to merge those changes to both branchA and branchB.

I've done this myself on numerous occasions when re-working small project- local API's, and it's very, very handy indeed. If git would refuse me to change branches without first committing everything I'd have to first commit the change separately, switch branch, cherrypick the change, go back to the first branch and remove the commit I made there, merge the other branch where the commit really belonged and only then I could go on about my business. If, on the other hand, I happen to switch branches before committing fileZ in the above example, I can just switch back and amend my last commit on the first branch.

So yes, this is a feature, and it's a handy one.
-- 
Andreas Ericsson                   andreas.ericsson@op5.se
OP5 AB                             www.op5.se
Tel: +46 8-230225                  Fax: +46 8-230231

Considering the successes of the wars on alcohol, poverty, drugs and
terror, I think we should give some serious thought to declaring war
on peace.
Previous: arQonNext: arQon
Message 5 of 36 in “[BUG] git checkout <branch> allowed with uncommitted changes”
  1. arQonOct 13, 2011
  2. Nguyen Thai Ngoc DuyOct 13, 2011
  3. Alexey ShumkinOct 13, 2011
  4. arQonOct 13, 2011
  5. Andreas EricssonOct 13, 2011
  6. arQonOct 13, 2011
  7. Carlos Martín NietoOct 13, 2011
  8. arQonOct 13, 2011
  9. Alexey ShumkinOct 13, 2011
  10. Jakub NarebskiOct 13, 2011
  11. arQonOct 13, 2011
  12. Carlos Martín NietoOct 13, 2011
  13. arQonOct 13, 2011
  14. Alexey ShumkinOct 13, 2011
  15. Alexey ShumkinOct 14, 2011
  16. Holger HellmuthOct 13, 2011
  17. Carlos Martín NietoOct 13, 2011
  18. arQonOct 13, 2011
  19. Junio C HamanoOct 13, 2011
  20. arQonOct 13, 2011
  21. Jeff KingOct 14, 2011
  22. Holger HellmuthOct 14, 2011
  23. Victor EngmarkOct 14, 2011
  24. arQonOct 16, 2011
  25. Junio C HamanoOct 16, 2011
  26. Holger HellmuthOct 16, 2011
  27. Carlos Martín NietoOct 13, 2011
  28. Sergei OrganovOct 13, 2011
  29. PJ WeisbergOct 13, 2011
  30. Holger HellmuthOct 13, 2011
  31. arQonOct 13, 2011
  32. Holger HellmuthOct 13, 2011
  33. Victor EngmarkOct 13, 2011
  34. arQonOct 13, 2011
  35. Victor EngmarkOct 14, 2011
  36. Michael J GruberOct 13, 2011

Read the whole thread, see it on lore, or plain text.

$ cat FOOTERMessages come from the public archive at lore.kernel.org/git, fetched every hour. The front page is chosen and written each morning by an AI editor and can be wrong; the threads themselves are the record. About and API. For agents: an MCP server at https://gitlist.dev/mcp, and any thread, story or person page as Markdown by adding .md to its URL (or sending Accept: text/markdown). Details in /llms.txt.