Re: [PATCH 2/2] status: improve rebase todo list parsing
- From
- Phillip Wood <phillip.wood123@gmail.com>
- Date
- Apr 21, 2026, 16:03 UTC
- Message-ID
- <9b239518-7b1c-4125-939e-38ebe69ed004@gmail.com>
- In-Reply-To
- <b62a96c5-fab4-4c6d-9768-ade48a8476ca@malon.dev>
Hi Tian
On 20/04/2026 17:38, Tian Yuchen wrote:
Show 22 quoted lines
> On 4/20/26 23:04, Phillip Wood wrote:
>
>> + if (!starts_with(p, abbrev))
>> + goto out; /* object name was a refname containing only
>> xdigits */
>> + p += strlen(abbrev);
>> + strbuf_remove(line, p - line->buf, end_of_object_name - p);
>> + end_of_object_name = p;
>
>
>> - if ((2 <= string_list_split(&split, line->buf, " ", 2)) &&
>> - !repo_get_oid(r, split.items[1].string, &oid)) {
>> - strbuf_reset(line);
>> - strbuf_addf(line, "%s ", split.items[0].string);
>> - strbuf_add_unique_abbrev(line, &oid, DEFAULT_ABBREV);
>> - for (size_t i = 2; i < split.nr; i++)
>> - strbuf_addf(line, " %s", split.items[i].string);
>
> I noticed that after this patch, refnames shorter than seven characters
> are no longer standardised to the standard seven-character length,
> because the 'start_with()' function always returns FALSE. The code jumps
> directly to 'out', without completing or cutting the refname.Do you mean object id's rather than refnames? We do not want to alter refnames, but we do want to abbreviate object ids. You're right that if the todo list contains an object id that is shorter than the default abbreviation length it will not be expanded but "git rebase" rewrites the todo list to contain full length object ids after the user has edited it. If a user may edit the todo list directly without using "git rebase --edit-todo" and then run "git status" before "git rebase --continue" the todo list could contain shorter object ids but that seems pretty unlikely.
Show 8 quoted lines
> > I’m not sure if this was your intention, but I just want to point it out > for your information. > > (Also noted that there is a very rare scenario where the OID of a > refname longer than 7 characters happens to begin with the refname > itself; in this case, 'start_with' returns TRUE and the string is cut > incorrectly. However, I think we can safely ignore this.)
Yes, I think that's unlikely to happen in practice.
Thanks
Phillip
> Regards, Yuchen