Re: [PATCH v4 2/3] branch: description for orphan branch errors

["Rubén Justo" <rjusto@gmail.com>]

Thank you for your review!

On 27/2/23 20:38, Jonathan Tan wrote:
> Firstly, the subject could be more precise. Maybe "branch: check all
> worktrees for orphan branches" (47 characters) or something like that.

The main intention in this series is to stop giving the user a confusing
error "No branch named..." for a branch he may have just created.  I
think the current subject states that better.  But I'm open to change it
in that direction.

> Rubén Justo <rjusto@gmail.com> writes:
>> In bcfc82bd48 (branch: description for non-existent branch errors,
>> 2022-10-08) we checked the current HEAD
> 
> Probably clearer to say "HEAD in the current worktree" instead of
> "current HEAD".

OK. I'll reword with that.

>> to detect if the branch to
>> operate with is an orphan branch, so as to avoid the confusing error:
>> "No branch named...".
>>
>> If we are asked to operate with an orphan branch in a different working
>> tree than the current one, we need to check the HEAD in that different
>> working tree.
> 
> Probably clearer to just say "But there might be orphan branches in
> other worktrees".

That loses important details IMHO, the intention: "avoid the
confusing..", and the reasoning on why we need to check HEAD in all
worktrees.

>> Let's extend the check we did in bcfc82bd48, to all HEADs in the
>> repository, using the helper introduced in 31ad6b61bd (branch: add
>> branch_checked_out() helper, 2022-06-15)
> 
> s/HEADs/worktrees/

I understand your suggestion, but my intention along the message is to
maintain the reasoning on the "HEAD", due to an orphan branch is a HEAD
pointing to a non-existing ref.  Maybe "the HEADs in all worktrees"
could be better?

>> @@ -493,8 +496,9 @@ static void copy_or_rename_branch(const char *oldname, const char *newname, int
>>  	struct strbuf oldsection = STRBUF_INIT, newsection = STRBUF_INIT;
>>  	const char *interpreted_oldname = NULL;
>>  	const char *interpreted_newname = NULL;
>> -	int recovery = 0;
>> +	int recovery = 0, oldref_usage = 0;
>>  	struct worktree **worktrees = get_worktrees();
>> +	struct worktree *oldref_wt = NULL;
> 
> Better to have 2 variables (one for rebased, and one for bisected) to
> avoid the situation in which the last problematic worktree seen was a
> bisected one, but a prior one was a rebased one.

Well seen.  Thanks for reading carefully.

I'll re-roll with that.

>> @@ -818,7 +835,7 @@ int cmd_branch(int argc, const char **argv, const char *prefix)
>>  
>>  		strbuf_addf(&branch_ref, "refs/heads/%s", branch_name);
>>  		if (!ref_exists(branch_ref.buf))
>> -			error((!argc || !strcmp(head, branch_name))
>> +			error((!argc || branch_checked_out(branch_ref.buf))
>>  			      ? _("No commit on branch '%s' yet.")
>>  			      : _("No branch named '%s'."),
>>  			      branch_name);
>> @@ -863,7 +880,7 @@ int cmd_branch(int argc, const char **argv, const char *prefix)
>>  		}
>>  
>>  		if (!ref_exists(branch->refname)) {
>> -			if (!argc || !strcmp(head, branch->name))
>> +			if (!argc || branch_checked_out(branch->refname))
>>  				die(_("No commit on branch '%s' yet."), branch->name);
>>  			die(_("branch '%s' does not exist"), branch->name);
>>  		}
> 
> What is the relevance of these changes?
> 

This is the main intention in the patch: not showing the confusing error
"No branch named..." for orphan branches.  I'm not sure if I understand
your question...