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

Re: [PATCH v2 2/4] commit: write commit-graph with Bloom filters

From
Junio C Hamano <gitster@pobox.com>
Date
Apr 14, 2020, 17:26 UTC
Message-ID
<xmqqa73e9dws.fsf@gitster.c.googlers.com>
In-Reply-To
<74e4e8d6-d401-081d-14cc-c4b6087bdeda@gmail.com>
Derrick Stolee <stolee@gmail.com> writes:
Show 21 quoted lines
> On 4/13/2020 6:21 PM, Junio C Hamano wrote:
>> Taylor Blau <me@ttaylorr.com> writes:
>> 
>>> Hmm. I'm not crazy about a library function looking at 'GIT_TEST_*'
>>> environment variables and potentially ignoring its arguments, but given
>>> the discussion we had in v1, I don't feel strongly enough to recommend
>>> that you change this.
>>>
>>> For what it's worth, I think that 'write_commit_graph' could behave more
>>> purely if callers checked 'GIT_TEST_COMMIT_GRAPH_CHANGED_PATHS' and set
>>> 'flags' appropriately from the outside,...
>> 
>> Yeah, I agree that it would be a lot cleaner if that is easy to
>> arrange.  I have a suspicion that Derrick already tried and the
>> resulting code looked messier and was discarded?
>
> Perhaps I could fix both concerns by
>
> 1. Use a macro instead of a library call.
>
> 2. Check the _CHANGED_PATHS variable in the macro.

I am not sure how use of a macro "fixes" purity, though. And what is the other concern?

How widely would this "if we are testing, write out the graph file" call be sprinkled over the codebase? I am hoping that it won't be "everywhere", but only at strategic places (like "just once before we leave a subcommand that creates one or bunch of commits"). And how often would they be called? I am also hoping that it won't be "inside a tight loop". In short, I am wondering if we can promise our codebase that

 - git_test_write_commit_graph_or_die() calls won't be an eyesore
   (and/or distraction) for developers too much.
 - git_env_bool() call won't have performance impact.
Show 15 quoted lines
> The macro would look like this:
>
>
> #define git_test_write_commit_graph_or_die() \
> 	if (git_env_bool(GIT_TEST_COMMIT_GRAPH, 0)) { \
> 		int flags = 0; \
> 		if (git_env_bool(GIT_TEST_COMMIT_GRAPH_CHANGED_PATHS, 0)) \
> 			flags = COMMIT_GRAPH_WRITE_BLOOM_FILTERS; \
> 		if (write_commit_graph_reachable( \
> 			the_repository->objects->odb, flags, NULL)) \
> 			die("failed to write commit-graph under GIT_TEST_COMMIT_GRAPH"); \
> 	}
>
> Thanks,
> -Stolee
Previous: Derrick StoleeNext: Derrick Stolee
Message 38 of 48 in “Integrate changed-path Bloom filters with 'git blame'”
  1. 0/3 Integrate changed-path Bloom filters with 'git blame'Derrick Stolee via GitGitGadget, Apr 11, 2020
  2. 1/3 revision: complicated pathspecs disable filtersDerrick Stolee via GitGitGadget, Apr 11, 2020
  3. Junio C HamanoApr 11, 2020
  4. Derrick StoleeApr 13, 2020
  5. Junio C HamanoApr 14, 2020
  6. Derrick StoleeApr 15, 2020
  7. Derrick StoleeApr 15, 2020
  8. Junio C HamanoApr 15, 2020
  9. Junio C HamanoApr 15, 2020
  10. Junio C HamanoApr 15, 2020
  11. Taylor BlauApr 16, 2020
  12. Jakub NarębskiApr 15, 2020
  13. Taylor BlauApr 16, 2020
  14. Derrick StoleeApr 16, 2020
  15. Taylor BlauApr 16, 2020
  16. Junio C HamanoApr 16, 2020
  17. Taylor BlauApr 12, 2020
  18. Junio C HamanoApr 12, 2020
  19. Taylor BlauApr 13, 2020
  20. Derrick StoleeApr 13, 2020
  21. 2/3 commit: write commit-graph with bloom filtersDerrick Stolee via GitGitGadget, Apr 11, 2020
  22. Junio C HamanoApr 11, 2020
  23. Taylor BlauApr 12, 2020
  24. Derrick StoleeApr 13, 2020
  25. Junio C HamanoApr 13, 2020
  26. 3/3 blame: use changed-path Bloom filtersDerrick Stolee via GitGitGadget, Apr 11, 2020
  27. Junio C HamanoApr 11, 2020
  28. Eric SunshineApr 12, 2020
  29. Junio C HamanoApr 11, 2020
  30. 0/4 Integrate changed-path Bloom filters with 'git blame'Derrick Stolee via GitGitGadget, Apr 13, 2020
  31. 1/4 revision: complicated pathspecs disable filtersDerrick Stolee via GitGitGadget, Apr 13, 2020
  32. Taylor BlauApr 13, 2020
  33. Junio C HamanoApr 13, 2020
  34. 2/4 commit: write commit-graph with Bloom filtersDerrick Stolee via GitGitGadget, Apr 13, 2020
  35. Taylor BlauApr 13, 2020
  36. Junio C HamanoApr 13, 2020
  37. Derrick StoleeApr 14, 2020
  38. Junio C HamanoApr 14, 2020
  39. Derrick StoleeApr 14, 2020
  40. Taylor BlauApr 15, 2020
  41. 3/4 commit-graph: write commit-graph in more testsDerrick Stolee via GitGitGadget, Apr 13, 2020
  42. 4/4 blame: use changed-path Bloom filtersDerrick Stolee via GitGitGadget, Apr 13, 2020
  43. Taylor BlauApr 13, 2020
  44. 0/3 Integrate changed-path Bloom filters with 'git blame'Derrick Stolee via GitGitGadget, Apr 16, 2020
  45. 1/3 revision: complicated pathspecs disable filtersDerrick Stolee via GitGitGadget, Apr 16, 2020
  46. SZEDER GáborJun 7, 2020
  47. 2/3 tests: write commit-graph with Bloom filtersDerrick Stolee via GitGitGadget, Apr 16, 2020
  48. 3/3 blame: use changed-path Bloom filtersDerrick Stolee via GitGitGadget, Apr 16, 2020

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.