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

Re: [PATCH] Introduce get_octopus_merge_bases() in commit.c

From
Junio C Hamano <gitster@pobox.com>
Date
Jun 11, 2008, 23:51 UTC
Message-ID
<7vprqna5t6.fsf@gitster.siamese.dyndns.org>
In-Reply-To
<1213219039-1211-1-git-send-email-vmiklos@frugalware.org>
Miklos Vajna <vmiklos@frugalware.org> writes:
Show 22 quoted lines
> diff --git a/commit.c b/commit.c
> index bbf9c75..f9410d8 100644
> --- a/commit.c
> +++ b/commit.c
> @@ -600,6 +600,33 @@ static struct commit_list *merge_bases(struct commit *one, struct commit *two)
>  	return result;
>  }
>  
> +struct commit_list *get_octopus_merge_bases(struct commit_list *in, int cleanup)
> +{
> +	struct commit_list *i, *j, *k, *ret = NULL;
> +	struct commit_list **pptr = &ret;
> +
> +	for (i = in; i; i = i->next) {
> +		if (!ret)
> +			pptr = &commit_list_insert(i->item, pptr)->next;
> +		else {
> +			struct commit_list *new = NULL, *end = NULL;
> +
> +			for (j = ret; j; j = j->next) {
> +				struct commit_list *bases;
> +				bases = get_merge_bases(i->item, j->item, cleanup);

Why is passing "cleanup" through to get_merge_bases() safe, even when you are running it more than once?

get_merge_bases() uses PARENT1, PARENT2 and other flags internally, and the information necessary to clean them efficiently (i.e. the set of commits that could be smudged with these flags due to its operation) is known only to it _after it finishes_. You cannot even say "previous run did not clean things up, so please clean-up before starting this round" upfront. That is why we have clean-up loops at the _end_ of the function after it finishes computing what it wants to return.

The only reason clean-up is optional to the function is because there are one-shot callers such as "git merge-base" that do not want to pay penalty for cleaning up (the only time you do not have to tell it to clean up is when you know the invocation is the _last_ invocation of the function, iow there is no later invocation that will be harmed by leftover flags).

Previous: Miklos VajnaNext: Miklos Vajna
Message 13 of 14 in “Build in merge”
  1. 00/10 Build in mergeMiklos Vajna, Jun 11, 2008
  2. 01/10 Move split_cmdline() to alias.cMiklos Vajna, Jun 11, 2008
  3. 02/10 Move commit_list_count() to commit.cMiklos Vajna, Jun 11, 2008
  4. 03/10 Move parse-options's skip_prefix() to git-compat-util.hMiklos Vajna, Jun 11, 2008
  5. 04/10 Add new test to ensure git-merge handles pull.twohead and pull.octopusMiklos Vajna, Jun 11, 2008
  6. 05/10 parseopt: add a new PARSE_OPT_ARGV0_IS_AN_OPTION optionMiklos Vajna, Jun 11, 2008
  7. 06/10 Move read_cache_unmerged() to read-cache.cMiklos Vajna, Jun 11, 2008
  8. 07/10 git-fmt-merge-msg: make it usable from other builtinsMiklos Vajna, Jun 11, 2008
  9. 08/10 Introduce get_octopus_merge_bases() in commit.cMiklos Vajna, Jun 11, 2008
  10. 09/10 Introduce filter_independent() in commit.cMiklos Vajna, Jun 11, 2008
  11. 10/10 Build in mergeMiklos Vajna, Jun 11, 2008
  12. Introduce get_octopus_merge_bases() in commit.cMiklos Vajna, Jun 11, 2008
  13. Junio C HamanoJun 11, 2008
  14. Introduce get_octopus_merge_bases() in commit.cMiklos Vajna, Jun 12, 2008

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.