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

Re: [PATCH v2] commit,shallow: unparse commits if grafts changed

From
Derrick Stolee <derrickstolee@github.com>
Date
Jun 3, 2022, 13:29 UTC
Message-ID
<394c054e-e1d2-41a5-a655-2ad3cb7219e0@github.com>
In-Reply-To
<220603.86k09yxf4z.gmgdl@evledraar.gmail.com>
On 6/3/2022 5:30 AM, Ævar Arnfjörð Bjarmason wrote:
Show 21 quoted lines
> 
> On Thu, Jun 02 2022, Jonathan Tan wrote:
> 
>> diff --git a/commit.c b/commit.c
>> index 59b6c3e455..1537ea73d0 100644
>> --- a/commit.c
>> +++ b/commit.c
>> @@ -120,6 +120,17 @@ int commit_graft_pos(struct repository *r, const struct object_id *oid)
>>  		       commit_graft_oid_access);
>>  }
>>  
>> +static void unparse_commit(struct repository *r, const struct object_id *oid)
>> +{
>> +	struct commit *c = lookup_commit(r, oid);
>> +
>> +	if (!c->object.parsed)
>> +		return;
>> +	free_commit_list(c->parents);
>> +	c->parents = NULL;
>> +	c->object.parsed = 0;
>> +}

This looks good. I took a quick inventory of 'struct commit' and agree that this is all we need. There is no need to clear the date or maybe_tree members. The commit_graph_data slab might have some information if there were no grafts before (but are now), but the existence of grafts should clear the commit-graph already and stop that slab from being used.

Show 28 quoted lines
>>  int register_commit_graft(struct repository *r, struct commit_graft *graft,
>>  			  int ignore_dups)
>>  {
>> @@ -145,6 +156,7 @@ int register_commit_graft(struct repository *r, struct commit_graft *graft,
>>  			(r->parsed_objects->grafts_nr - pos - 1) *
>>  			sizeof(*r->parsed_objects->grafts));
>>  	r->parsed_objects->grafts[pos] = graft;
>> +	unparse_commit(r, &graft->oid);
>>  	return 0;
>>  }
>>  
>> @@ -253,8 +265,10 @@ void reset_commit_grafts(struct repository *r)
>>  {
>>  	int i;
>>  
>> -	for (i = 0; i < r->parsed_objects->grafts_nr; i++)
>> +	for (i = 0; i < r->parsed_objects->grafts_nr; i++) {
>> +		unparse_commit(r, &r->parsed_objects->grafts[i]->oid);
>>  		free(r->parsed_objects->grafts[i]);
>> +	}
>>  	r->parsed_objects->grafts_nr = 0;
>>  	r->parsed_objects->commit_graft_prepared = 0;
>>  }
> 
> Are we going to have the same issue with tags, c.f. parse_tag() and
> there being no unparse_tag()?
> 
> (I don't know offhand, just asking)

Grafts are only on commits. You cannot replace what a tag is pointing at with a graft. Replace-objects is a different thing and changes it at the OID level (and I don't think this can happen during the process without concurrent external changes to the filesystem).

Thanks, -Stolee

Previous: Ævar Arnfjörð BjarmasonNext: Jonathan Tan
Message 13 of 16 in “commit-reach: do not parse and iterate minima”
  1. commit-reach: do not parse and iterate minimaJonathan Tan, Mar 23, 2022
  2. Junio C HamanoMar 23, 2022
  3. Derrick StoleeMar 24, 2022
  4. Taylor BlauMar 24, 2022
  5. Derrick StoleeMar 25, 2022
  6. Jonathan TanMar 24, 2022
  7. Bagas SanjayaMar 24, 2022
  8. Jonathan TanMar 24, 2022
  9. Derrick StoleeMar 24, 2022
  10. Jonathan TanMar 24, 2022
  11. commit,shallow: unparse commits if grafts changedJonathan Tan, Jun 2, 2022
  12. Ævar Arnfjörð BjarmasonJun 3, 2022
  13. Derrick StoleeJun 3, 2022
  14. Jonathan TanJun 3, 2022
  15. Jonathan TanJun 3, 2022
  16. commit,shallow: unparse commits if grafts changedJonathan Tan, Jun 6, 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.