Git development
 help / color / mirror / Atom feed
From: Phillip Wood <phillip.wood123@gmail.com>
To: Son Luong Ngoc via GitGitGadget <gitgitgadget@gmail.com>,
	git@vger.kernel.org
Cc: Kristoffer Haugsbakk <kristofferhaugsbakk@fastmail.com>,
	Son Luong Ngoc <sluongng@gmail.com>
Subject: Re: [PATCH v3 2/2] rebase: guard non-branch symref targets
Date: Fri, 7 Aug 2026 16:22:33 +0100	[thread overview]
Message-ID: <98682fa4-55d9-4829-97f1-02e244b35266@gmail.com> (raw)
In-Reply-To: <a653f56ea214e74ea71ba31f5378f9cbf8b04dde.1784708107.git.gitgitgadget@gmail.com>

On 22/07/2026 09:15, Son Luong Ngoc via GitGitGadget wrote:
> From: Son Luong Ngoc <sluongng@gmail.com>
> 
> A local branch symbolic ref may point outside refs/heads/. Such an alias
> cannot be skipped like a branch-to-branch alias because its concrete
> target ref is absent from the local branch decoration list.
> 
> However, queuing each alias independently can update the same target ref
> more than once and make the second compare-and-swap fail. A reservation
> from another worktree can also name either an alias or its resolved
> target ref, so checking only one form can miss an in-progress update.
> 
> Fix these cases by checking both the literal alias and its resolved
> target ref against checked-out reservations. Deduplicate updates by
> target ref. Also reserve both forms when loading another worktree's
> update-refs state. This makes different aliases honor the same
> in-progress update.
> 
> This keeps non-branch symrefs supported without allowing duplicate or
> cross-worktree ref updates.

I've left a couple of questions below. If you're short on time I'd 
suggest we just concentrate on getting the first patch merged as that 
fixes the common case of one branch being a symlink to another. We can 
worry about two branches being symlinks to a ref outside "refs/heads/*" 
later if it turns out to be a problem in practice.

> diff --git a/branch.c b/branch.c
> index 243db7d0fc..98a50d8368 100644
> --- a/branch.c
> +++ b/branch.c
> @@ -442,10 +442,25 @@ static void prepare_checked_out_branches(void)
>   						     &update_refs)) {
>   			struct string_list_item *item;
>   			for_each_string_list_item(item, &update_refs) {
> +				char *resolved_ref;
> +				int flags = 0;
> +
>   				old = strmap_put(&current_checked_out_branches,
>   						 item->string,
>   						 xstrdup(wt->path));
>   				free(old);
> +
> +				resolved_ref = refs_resolve_refdup(
> +					get_main_ref_store(the_repository),
> +					item->string, RESOLVE_REF_READING,
> +					NULL, &flags);
> +				if (resolved_ref && (flags & REF_ISSYMREF)) {
> +					old = strmap_put(
> +						&current_checked_out_branches,
> +						resolved_ref, xstrdup(wt->path));
> +					free(old);
> +				}
> +				free(resolved_ref);
>   			}

After the last commit, when we prepare the todo list don't we skip any 
symbolic refs and only record their target? That would mean there 
shouldn't be any symbolic refs to resolve here. I do wonder if the 
earlier part of this function should be storing the symref and its 
target when it walks all the worktree HEADs. If we have a branch 
"refs/heads/feature" and a symref "refs/heads/symlink-to-feature" is it 
possible to have them checkedout in different worktrees because we only 
add HEAD to the list of checked out branches when we walk all the 
worktree HEADs?

>   			string_list_clear(&update_refs, 1);
>   		}
> diff --git a/sequencer.c b/sequencer.c
> index 63aba60a08..040b5bf645 100644
> --- a/sequencer.c
> +++ b/sequencer.c
> @@ -6459,6 +6459,7 @@ struct todo_add_branch_context {
>   	size_t items_alloc;
>   	struct strbuf *buf;
>   	struct string_list refs_to_oids;
> +	struct string_list symref_update_targets;
>   };
>   
>   static int add_decorations_to_list(const struct commit *commit,
> @@ -6473,6 +6474,7 @@ static int add_decorations_to_list(const struct commit *commit,
>   	while (decoration) {
>   		struct todo_item *item;
>   		const char *path;
> +		const char *checked_ref;
>   		char *resolved_ref;
>   		int flags = 0;
>   		size_t base_offset = ctx->buf->len;
> @@ -6508,6 +6510,17 @@ static int add_decorations_to_list(const struct commit *commit,
>   		}
>   
>   		path = branch_checked_out(decoration->name);
> +		if (!path && resolved_ref && (flags & REF_ISSYMREF)) {
> +			checked_ref = resolved_ref;
> +			path = branch_checked_out(checked_ref);
> +		}
> +		if (!path && resolved_ref && (flags & REF_ISSYMREF) &&
> +		    string_list_has_string(&ctx->symref_update_targets,
> +					   resolved_ref)) {
> +			free(resolved_ref);
> +			decoration = decoration->next;
> +			continue;
> +		}

So we check the to see if the symref or its target are checked out. 
That's necessary because we might have stored a symref rather than its 
target in current_checked_out_branches above (which I think is probably 
a bug). If two branches are symrefs to the same ref we'll only queue the 
update once which is good.

Thanks

Phillip
>   		ALLOC_GROW(ctx->items,
>   			ctx->items_nr + 1,
> @@ -6523,6 +6536,10 @@ static int add_decorations_to_list(const struct commit *commit,
>   					      decoration->name, path);
>   		} else {
>   			struct string_list_item *sti;
> +
> +			if (resolved_ref && (flags & REF_ISSYMREF))
> +				string_list_insert(&ctx->symref_update_targets,
> +						   resolved_ref);
>   			item->command = TODO_UPDATE_REF;
>   			strbuf_addf(ctx->buf, "%s\n", decoration->name);
>   
> @@ -6554,6 +6571,7 @@ static int todo_list_add_update_ref_commands(struct todo_list *todo_list)
>   	struct todo_add_branch_context ctx = {
>   		.buf = &todo_list->buf,
>   		.refs_to_oids = STRING_LIST_INIT_DUP,
> +		.symref_update_targets = STRING_LIST_INIT_DUP,
>   	};
>   
>   	ctx.items_alloc = 2 * todo_list->nr + 1;
> @@ -6579,6 +6597,7 @@ static int todo_list_add_update_ref_commands(struct todo_list *todo_list)
>   	res = write_update_refs_state(&ctx.refs_to_oids);
>   
>   	string_list_clear(&ctx.refs_to_oids, 1);
> +	string_list_clear(&ctx.symref_update_targets, 0);
>   
>   	if (res) {
>   		/* we failed, so clean up the new list. */
> diff --git a/t/t3404-rebase-interactive.sh b/t/t3404-rebase-interactive.sh
> index 11afa8be56..110ed8ae63 100755
> --- a/t/t3404-rebase-interactive.sh
> +++ b/t/t3404-rebase-interactive.sh
> @@ -2024,6 +2024,78 @@ test_expect_success '--update-refs updates refs correctly' '
>   	test_cmp expect err.trimmed
>   '
>   
> +test_expect_success '--update-refs checks resolved non-branch symref target' '
> +	test_when_finished "
> +		git worktree remove --force checked-out-target-wt &&
> +		git symbolic-ref -d refs/heads/non-branch-alias &&
> +		git tag -d checked-out-target
> +	" &&
> +	git tag checked-out-target HEAD~1 &&
> +	git symbolic-ref refs/heads/non-branch-alias refs/tags/checked-out-target &&
> +	git worktree add --detach checked-out-target-wt checked-out-target &&
> +	git -C checked-out-target-wt symbolic-ref HEAD refs/tags/checked-out-target &&
> +
> +	GIT_SEQUENCE_EDITOR="cat >todo" git rebase -i --update-refs HEAD~2 &&
> +
> +	test_grep "^# Ref refs/heads/non-branch-alias checked out at" todo &&
> +	test_write_lines refs/tags/checked-out-target >expect &&
> +	git symbolic-ref refs/heads/non-branch-alias >actual &&
> +	test_cmp expect actual
> +'
> +
> +test_expect_success '--update-refs deduplicates non-branch symref targets' '
> +	test_when_finished "
> +		git symbolic-ref -d refs/heads/non-branch-alias-one &&
> +		git symbolic-ref -d refs/heads/non-branch-alias-two &&
> +		git tag -d shared-non-branch-target
> +	" &&
> +	git tag shared-non-branch-target HEAD~1 &&
> +	git symbolic-ref refs/heads/non-branch-alias-one \
> +		refs/tags/shared-non-branch-target &&
> +	git symbolic-ref refs/heads/non-branch-alias-two \
> +		refs/tags/shared-non-branch-target &&
> +
> +	GIT_SEQUENCE_EDITOR=: git rebase -i --force-rebase --update-refs HEAD~2 &&
> +
> +	test_cmp_rev HEAD~1 refs/heads/non-branch-alias-one &&
> +	test_cmp_rev HEAD~1 refs/heads/non-branch-alias-two &&
> +	test_write_lines refs/tags/shared-non-branch-target >expect &&
> +	git symbolic-ref refs/heads/non-branch-alias-one >actual &&
> +	test_cmp expect actual &&
> +	git symbolic-ref refs/heads/non-branch-alias-two >actual &&
> +	test_cmp expect actual
> +'
> +
> +test_expect_success '--update-refs honors non-branch symref reservations' '
> +	test_when_finished "
> +		test_might_fail git worktree remove --force reserved-target-wt &&
> +		test_might_fail git symbolic-ref -d \
> +			refs/heads/reserved-non-branch-alias-one &&
> +		test_might_fail git symbolic-ref -d \
> +			refs/heads/reserved-non-branch-alias-two &&
> +		test_might_fail git tag -d reserved-non-branch-target
> +	" &&
> +	git tag reserved-non-branch-target HEAD~1 &&
> +	git symbolic-ref refs/heads/reserved-non-branch-alias-one \
> +		refs/tags/reserved-non-branch-target &&
> +	git symbolic-ref refs/heads/reserved-non-branch-alias-two \
> +		refs/tags/reserved-non-branch-target &&
> +	git worktree add --detach reserved-target-wt HEAD &&
> +	wt_gitdir=$(git -C reserved-target-wt rev-parse --absolute-git-dir) &&
> +	mkdir -p "$wt_gitdir/rebase-merge" &&
> +	old_oid=$(git rev-parse refs/heads/reserved-non-branch-alias-one) &&
> +	test_write_lines refs/heads/reserved-non-branch-alias-one \
> +		"$old_oid" "$old_oid" >"$wt_gitdir/rebase-merge/update-refs" &&
> +
> +	GIT_SEQUENCE_EDITOR="cat >todo" git rebase -i --update-refs HEAD~2 &&
> +
> +	test_grep "^# Ref refs/heads/reserved-non-branch-alias-one checked out at" \
> +		todo &&
> +	test_grep "^# Ref refs/heads/reserved-non-branch-alias-two checked out at" \
> +		todo &&
> +	test_grep ! "^update-ref refs/heads/reserved-non-branch-alias" todo
> +'
> +
>   test_expect_success 'respect user edits to update-ref steps' '
>   	git checkout -B update-refs-break no-conflict-branch &&
>   	git branch -f base HEAD~4 &&


      reply	other threads:[~2026-08-07 15:22 UTC|newest]

Thread overview: 25+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-05-28  5:41 [PATCH 0/2] rebase: handle --update-refs branch symrefs Son Luong Ngoc via GitGitGadget
2026-05-28  5:42 ` [PATCH 1/2] t3404: add failing branch symref test Son Luong Ngoc via GitGitGadget
2026-06-01 13:52   ` Phillip Wood
2026-05-28  5:42 ` [PATCH 2/2] rebase: skip branch symref aliases Son Luong Ngoc via GitGitGadget
2026-05-28  7:08   ` Kristoffer Haugsbakk
2026-06-01 14:10   ` Phillip Wood
2026-05-28 20:42 ` [PATCH 0/2] rebase: handle --update-refs branch symrefs Junio C Hamano
2026-06-03 10:27 ` [PATCH v2] rebase: skip branch symref aliases Son Luong Ngoc via GitGitGadget
2026-06-04 15:37   ` Phillip Wood
2026-07-22  8:16     ` Son Luong Ngoc
2026-07-22  8:15   ` [PATCH v3 0/2] rebase: handle --update-refs branch symrefs Son Luong Ngoc via GitGitGadget
2026-07-22  8:15     ` [PATCH v3 1/2] rebase: skip branch symref aliases Son Luong Ngoc via GitGitGadget
2026-07-23 18:58       ` Phillip Wood
2026-07-24  9:55         ` Phillip Wood
2026-07-25 11:21           ` Erik Cervin-Edin
2026-07-26 15:42           ` Junio C Hamano
2026-07-28  9:36             ` Phillip Wood
2026-07-28 14:23               ` Junio C Hamano
2026-07-29  9:31                 ` Phillip Wood
2026-07-29 14:26                   ` Junio C Hamano
2026-07-30 13:10                     ` Phillip Wood
2026-08-06 20:12                       ` Junio C Hamano
2026-08-07 15:28                         ` Phillip Wood
2026-07-22  8:15     ` [PATCH v3 2/2] rebase: guard non-branch symref targets Son Luong Ngoc via GitGitGadget
2026-08-07 15:22       ` Phillip Wood [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=98682fa4-55d9-4829-97f1-02e244b35266@gmail.com \
    --to=phillip.wood123@gmail.com \
    --cc=git@vger.kernel.org \
    --cc=gitgitgadget@gmail.com \
    --cc=kristofferhaugsbakk@fastmail.com \
    --cc=phillip.wood@dunelm.org.uk \
    --cc=sluongng@gmail.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox