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

Re: git-diff-tree rename detection bug

From
Junio C Hamano <junkio@cox.net>
Date
Sep 14, 2005, 18:51 UTC
Message-ID
<7v3bo7jxdn.fsf@assigned-by-dhcp.cox.net>
In-Reply-To
<59a6e583050914094777c4fe96@mail.gmail.com>
Wayne Scott <wsc9tt@gmail.com> writes:
> Look at the diffs between ad6571a78ac74e9fa27e581834709067dba459af and
> it's parent with and without rename detection enabled.  (In linux-2.6
> git tree)
Show 8 quoted lines
> $ git-diff-tree -r  -M $REV^1 $REV | grep termios.h
> :000000 100644 0000000000000000000000000000000000000000
>    237533bb0e9f1a3e640c4906d8b350deafd315b9 A      include/asm-powerpc/termios.h
> :100644 000000 97c6287a6cbaa5903ee1a5934a5553e9e485d8e7
>    0000000000000000000000000000000000000000 D      include/asm-ppc/termios.h
>
> Notice how the the fact that include/asm-ppc64/termios.h is deleted gets lost?
> Looks broken to me.

It looks broken to me, too. I rebuilt from a reasonably ancient source (v0.99) and re-run the test but I could not get it to produce 'A' for include/asm-powerpc/termios.h. So I rewound it further to 4d235c8044a638108b67e22f94b2876657130fc8 commit, which is really ancient version, but it still says it is renamed from asm-ppc64 directory. FWIW, all the v0.99* tagged versions seem to detect that rename correctly and not lose anything in my tests.

Which version of git do you run and on what platform? It might be that something in the diffcore chain is broken in non-i386 and/or non-GNU/Linux and/or non-GCC environment.

Shoot, I thought it would be a good practice-case for me to use 'git bisect' in reverse to find the commit that fixed a bug ;-). My copy of linux-2.6 repository for testing is fully packed so I could not try the commit that introduced diffcore-rename.c, but that is what I really wanted to try.

Previous: Nicolas PitreNext: Junio C Hamano
Message 20 of 23 in “git-diff-tree rename detection bug”
  1. Wayne ScottSep 14, 2005
  2. Junio C HamanoSep 14, 2005
  3. Wayne ScottSep 14, 2005
  4. Linus TorvaldsSep 14, 2005
  5. Fix alloc_filespec() initializationLinus Torvalds, Sep 14, 2005
  6. Paul MackerrasSep 15, 2005
  7. Paul MackerrasSep 15, 2005
  8. Linus TorvaldsSep 15, 2005
  9. Paul MackerrasSep 15, 2005
  10. Junio C HamanoSep 15, 2005
  11. Josef WeidendorferSep 15, 2005
  12. Paul MackerrasSep 20, 2005
  13. Linus TorvaldsSep 15, 2005
  14. Junio C HamanoSep 16, 2005
  15. H. Peter AnvinSep 15, 2005
  16. Junio C HamanoSep 15, 2005
  17. Linus TorvaldsSep 15, 2005
  18. Matthias UrlichsSep 16, 2005
  19. Nicolas PitreSep 15, 2005
  20. Junio C HamanoSep 14, 2005
  21. Junio C HamanoSep 14, 2005
  22. Linus TorvaldsSep 14, 2005
  23. Junio C HamanoSep 14, 2005

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.