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

Re: [PATCH v2 2/7] fetch: backfill tags before setting upstream

From
Junio C Hamano <gitster@pobox.com>
Date
Feb 17, 2022, 22:07 UTC
Message-ID
<xmqqzgmp6tbo.fsf@gitster.g>
In-Reply-To
<b0a067bbc1e7b9d3ad76acd437d8095c5a48c5d7.1645102965.git.ps@pks.im>
Patrick Steinhardt <ps@pks.im> writes:
Show 17 quoted lines
> The fetch code flow is a bit hard to understand right now:
>
>     1. We optionally prune all references which have vanished on the
>        remote side.
>     2. We fetch and update all other references locally.
>     3. We update the upstream branch in the gitconfig.
>     4. We backfill tags pointing into the history we have just fetched.
>
> It is quite confusing that we fetch objects and update references in
> both (2) and (4), which is further stressed by the point that we use a
> `skip` goto label to jump from (3) to (4) in case we fail to update the
> gitconfig as expected.
>
> Reorder the code to first update all local references, and only after we
> have done so update the upstream branch information. This improves the
> code flow and furthermore makes it easier to refactor the way we update
> references together.

OK, as "setting upsream" is more or less unrelated to the act of actual fetching the refs and objects reachable from them, moving it outside the main code that is about fetching does make sense.

Show 5 quoted lines
>  static int do_fetch(struct transport *transport,
>  		    struct refspec *rs)
>  {
> -	struct ref *ref_map;
> +	struct ref *ref_map = NULL;

This is needed because we will always do free_refs() on the variable after this patch, and one early "goto cleanup" happens even before we touch ref_map, when truncate_fetch_head() fails.

Show 7 quoted lines
> @@ -1620,11 +1620,24 @@ static int do_fetch(struct transport *transport,
>  			retcode = 1;
>  	}
>  	if (fetch_and_consume_refs(transport, ref_map, worktrees)) {
> -		free_refs(ref_map);
>  		retcode = 1;
>  		goto cleanup;

And because we always free_refs(ref_map), we can lose a call here. OK.

Show 13 quoted lines
> +	/*
> +	 * If neither --no-tags nor --tags was specified, do automated tag
> +	 * following.
> +	 */
> +	if (tags == TAGS_DEFAULT && autotags) {
> +		struct ref *tags_ref_map = NULL, **tail = &tags_ref_map;
> +
> +		find_non_local_tags(remote_refs, &tags_ref_map, &tail);
> +		if (tags_ref_map)
> +			backfill_tags(transport, tags_ref_map, worktrees);
> +
> +		free_refs(tags_ref_map);
> +	}

Here, the new code uses a local and separete tags_ref_map variable and free it before we leave, instead of reusing ref_map variable.

OK.
Previous: Patrick SteinhardtNext: Patrick Steinhardt
Message 12 of 35 in “fetch: improve atomicity of `--atomic` flag”
  1. 0/7 fetch: improve atomicity of `--atomic` flagPatrick Steinhardt, Feb 17, 2022
  2. 1/7 fetch: increase test coverage of fetchesPatrick Steinhardt, Feb 17, 2022
  3. Christian CouderFeb 17, 2022
  4. Patrick SteinhardtFeb 21, 2022
  5. Junio C HamanoFeb 17, 2022
  6. Junio C HamanoFeb 17, 2022
  7. Patrick SteinhardtFeb 18, 2022
  8. Junio C HamanoFeb 18, 2022
  9. Junio C HamanoMar 3, 2022
  10. Patrick SteinhardtMar 3, 2022
  11. 2/7 fetch: backfill tags before setting upstreamPatrick Steinhardt, Feb 17, 2022
  12. Junio C HamanoFeb 17, 2022
  13. 3/7 fetch: control lifecycle of FETCH_HEAD in a single placePatrick Steinhardt, Feb 17, 2022
  14. Junio C HamanoFeb 17, 2022
  15. 4/7 fetch: report errors when backfilling tags failsPatrick Steinhardt, Feb 17, 2022
  16. Junio C HamanoFeb 17, 2022
  17. 6/7 fetch: make `--atomic` flag cover backfilling of tagsPatrick Steinhardt, Feb 17, 2022
  18. Junio C HamanoFeb 17, 2022
  19. 5/7 refs: add interface to iterate over queued transactional updatesPatrick Steinhardt, Feb 17, 2022
  20. 7/7 fetch: make `--atomic` flag cover pruning of refsPatrick Steinhardt, Feb 17, 2022
  21. Christian CouderFeb 17, 2022
  22. Junio C HamanoFeb 17, 2022
  23. 0/7 fetch: improve atomicity of `--atomic` flagPatrick Steinhardt, Feb 21, 2022
  24. 1/7 fetch: increase test coverage of fetchesPatrick Steinhardt, Feb 21, 2022
  25. Junio C HamanoMar 3, 2022
  26. Junio C HamanoMar 3, 2022
  27. Patrick SteinhardtMar 3, 2022
  28. Junio C HamanoMar 3, 2022
  29. Patrick SteinhardtMar 3, 2022
  30. 2/7 fetch: backfill tags before setting upstreamPatrick Steinhardt, Feb 21, 2022
  31. 3/7 fetch: control lifecycle of FETCH_HEAD in a single placePatrick Steinhardt, Feb 21, 2022
  32. 4/7 fetch: report errors when backfilling tags failsPatrick Steinhardt, Feb 21, 2022
  33. 5/7 refs: add interface to iterate over queued transactional updatesPatrick Steinhardt, Feb 21, 2022
  34. 6/7 fetch: make `--atomic` flag cover backfilling of tagsPatrick Steinhardt, Feb 21, 2022
  35. 7/7 fetch: make `--atomic` flag cover pruning of refsPatrick Steinhardt, Feb 21, 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.