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

Re: [PATCH] builtin: replace the_repository parameter in is_bare_repository()

From
Junio C Hamano <gitster@pobox.com>
Date
Aug 27, 2026, 19:51 UTC
Message-ID
<xmqqh5kf8hqc.fsf@gitster.g>
In-Reply-To
<xmqqo6en8jof.fsf@gitster.g>
Junio C Hamano <gitster@pobox.com> writes:
Show 5 quoted lines
>> - Only update sites that do not introduce any functional changes.
>
> How did you make sure that the places touched by this patch do not?
>
> How do you deal with repo==NULL case in these code paths?

I guess this was a bit too short, so let me explain in a bit more detail.

Show 21 quoted lines
>> diff --git a/builtin/blame.c b/builtin/blame.c
>> index 48d5251c6d..dbf4b4ffc7 100644
>> --- a/builtin/blame.c
>> +++ b/builtin/blame.c
>> @@ -957,7 +957,7 @@ static void build_ignorelist(struct blame_scoreboard *sb,
>>  int cmd_blame(int argc,
>>  	      const char **argv,
>>  	      const char *prefix,
>> -	      struct repository *repo UNUSED)
>> +	      struct repository *repo)
>>  {
>>  	struct rev_info revs;
>>  	char *path = NULL;
>> @@ -1187,7 +1187,7 @@ int cmd_blame(int argc,
>>  
>>  	revs.disable_stdin = 1;
>>  	setup_revisions(argc, argv, &revs, NULL);
>> -	if (!revs.pending.nr && is_bare_repository(the_repository)) {
>> +	if (!revs.pending.nr && is_bare_repository(repo)) {
>>  		struct commit *head_commit;
>>  		struct object_id head_oid;

There are a handful of uses of the_repository before the execution reaches here. But you left them unmodified.

The original code used to consistently used the_repository. Here you changed it to use "repo". In practice, they are most likely the same when "repo" is not NULL, so in that sense, this may not be breaking anything, but you must ask yourself what the point is, unless you convert all uses of the_repository with "repo". It does not help libification effort at all.

In general, builtin/foo.c::cmd_foo() are concrete programs that work on specific repository (i.e., the_repository), and there is not much reason to rewrite the use of the_repository to use "repo" given by the caller which is git potty. You'd also need to deal with the case where "repo" is NULL (hint: "cd / && git foo -h").

They are quite different from other parts of the system, things outside builtin/, many of which are general utility/helper routines, many of which should be designed to work with given repository.

Previous: Junio C HamanoNext: Hardik Kumar
Message 3 of 47 in “builtin: replace the_repository parameter in is_bare_repository()”
  1. builtin: replace the_repository parameter in is_bare_repository()Hardik Kumar, Aug 27, 2026
  2. Junio C HamanoAug 27, 2026
  3. Junio C HamanoAug 27, 2026
  4. Hardik KumarAug 27, 2026
  5. Junio C HamanoAug 27, 2026
  6. Ben KnobleAug 27, 2026
  7. Junio C HamanoAug 27, 2026
  8. D. Ben KnobleAug 28, 2026
  9. Junio C HamanoAug 28, 2026
  10. 0/8 More sensible checkout/switch/restore code refactoringJunio C Hamano, Aug 28, 2026
  11. 1/8 checkout: pass cb_option explicitly to branch name parsersJunio C Hamano, Aug 28, 2026
  12. 2/8 checkout: validate new branch name in checkout_branch()Junio C Hamano, Aug 28, 2026
  13. 3/8 checkout: validate stage and merge option compatibility in checkout_paths()Junio C Hamano, Aug 28, 2026
  14. 4/8 checkout: extract option validation and pathspec helpersJunio C Hamano, Aug 28, 2026
  15. 5/8 checkout: extract branch setup and tracking helpersJunio C Hamano, Aug 28, 2026
  16. 6/8 checkout: restructure switch, restore, and checkout entrypointsJunio C Hamano, Aug 28, 2026
  17. 7/8 checkout: wrap overly long linesJunio C Hamano, Aug 28, 2026
  18. Junio C HamanoAug 28, 2026
  19. Junio C HamanoAug 29, 2026
  20. 8/8 checkout: move post_checkout_hook() to checkout.cJunio C Hamano, Aug 28, 2026
  21. Junio C HamanoAug 28, 2026
  22. Junio C HamanoAug 29, 2026
  23. 0/8 More sensible checkout/switch/restore code refactoringJunio C Hamano, Aug 30, 2026
  24. 1/8 checkout: pass cb_option explicitly to branch name parsersJunio C Hamano, Aug 30, 2026
  25. Karthik NayakSep 1, 2026
  26. 2/8 checkout: validate new branch name in checkout_branch()Junio C Hamano, Aug 30, 2026
  27. Karthik NayakSep 1, 2026
  28. 3/8 checkout: validate stage and merge option compatibility in checkout_paths()Junio C Hamano, Aug 30, 2026
  29. Karthik NayakSep 1, 2026
  30. Junio C HamanoSep 1, 2026
  31. Karthik NayakSep 2, 2026
  32. Junio C HamanoSep 2, 2026
  33. Karthik NayakSep 3, 2026
  34. 4/8 checkout: extract option validation and pathspec helpersJunio C Hamano, Aug 30, 2026
  35. 5/8 checkout: extract branch setup and tracking helpersJunio C Hamano, Aug 30, 2026
  36. 6/8 checkout: restructure switch, restore, and checkout entrypointsJunio C Hamano, Aug 30, 2026
  37. Karthik NayakSep 1, 2026
  38. Junio C HamanoSep 1, 2026
  39. Karthik NayakSep 2, 2026
  40. 7/8 checkout: wrap overly long linesJunio C Hamano, Aug 30, 2026
  41. 8/8 checkout: move post_checkout_hook() to checkout.cJunio C Hamano, Aug 30, 2026
  42. D. Ben KnobleAug 29, 2026
  43. do not pass "repo" to builtin commmand implementationsJunio C Hamano, Aug 27, 2026
  44. Hardik KumarAug 28, 2026
  45. Junio C HamanoAug 28, 2026
  46. Hardik KumarAug 28, 2026
  47. Hardik KumarAug 27, 2026

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.