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

Re: [PATCH 5/6] sequencer: duplicate the result of resolve_ref_unsafe()

From
Derrick Stolee <derrickstolee@github.com>
Date
Sep 30, 2022, 17:23 UTC
Message-ID
<1711157b-8f04-008b-1672-f1fbac3c3e81@github.com>
In-Reply-To
<20220930140948.80367-6-szeder.dev@gmail.com>
On 9/30/2022 10:09 AM, SZEDER Gábor wrote:
Show 21 quoted lines
> When 'git rebase -i --update-refs' generates the todo list for the
> rebased commit range, an 'update-ref' instruction is inserted for each
> ref that points to one of those commits, except for the rebased branch
> (because at the end of the rebase it will be updated anyway) and any
> refs that are checked out in a worktree; for the latter a "Ref <ref>
> checked out at '<worktree>'" comment is added.  One of these comments
> can be missing under some circumstances: if the oldest commit with a
> ref pointing to it has multiple refs pointing to it, and at least one
> of those refs is checked out in a worktree, and one of them (but not
> the first) is checked out in the worktree associated with the last
> directory entry in '.git/worktrees'.
> 
> The culprit is the add_decorations_to_list() function, which calls
> resolve_ref_unsafe() to figure out the refname of the rebased branch.
> However, resolve_ref_unsafe() can (and in this case does) return a
> pointer to a static buffer.  Alas, add_decorations_to_list() holds on
> that static buffer until the end of the function, using its contents
> to compare refnames with the rebased branch, while it also calls a
> function that invokes refs_resolve_ref_unsafe() internally [1], and
> which overwrites the content of that static buffer, and messes up
> subsequent refname comparisons.
Good catch!
Show 6 quoted lines
> -	const char *head_ref = resolve_ref_unsafe("HEAD",
> +	const char *head_ref = xstrdup_or_null(resolve_ref_unsafe("HEAD",
>  						  RESOLVE_REF_READING,
>  						  NULL,
> -						  NULL);
> +						  NULL));

Moving to a 'char *' matches our typical pattern of "I am responsible for freeing this or passing that responsibility to someone else."

> +test_expect_success 'what should I call this?!' '
Perhaps 'with multiple refs, correctly report worktree' ?

Thanks, -Stolee

Previous: Ævar Arnfjörð BjarmasonNext: SZEDER Gábor
Message 13 of 20 in “rebase --update-refs: smooth out some rough edges”
  1. 0/6 rebase --update-refs: smooth out some rough edgesSZEDER Gábor, Sep 30, 2022
  2. 1/6 t2407-worktree-heads.sh: remove outdated loopSZEDER Gábor, Sep 30, 2022
  3. 2/6 t3404-rebase-interactive: mark a test with REFFILES prereqSZEDER Gábor, Sep 30, 2022
  4. Ævar Arnfjörð BjarmasonSep 30, 2022
  5. Junio C HamanoOct 5, 2022
  6. 4/6 sequencer: avoid empty lines after 'update-ref' instructionsSZEDER Gábor, Sep 30, 2022
  7. Derrick StoleeSep 30, 2022
  8. Junio C HamanoOct 5, 2022
  9. 3/6 rebase -i: emphasize that 'update-ref' expects a fully-qualified refSZEDER Gábor, Sep 30, 2022
  10. 5/6 sequencer: duplicate the result of resolve_ref_unsafe()SZEDER Gábor, Sep 30, 2022
  11. Ævar Arnfjörð BjarmasonSep 30, 2022
  12. Ævar Arnfjörð BjarmasonSep 30, 2022
  13. Derrick StoleeSep 30, 2022
  14. SZEDER GáborOct 9, 2022
  15. 6/6 sequencer: fail early if invalid ref is given to 'update-ref' instructionSZEDER Gábor, Sep 30, 2022
  16. Ævar Arnfjörð BjarmasonSep 30, 2022
  17. Derrick StoleeSep 30, 2022
  18. SZEDER GáborOct 1, 2022
  19. Junio C HamanoOct 1, 2022
  20. Junio C HamanoOct 5, 2022

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.