Re: [PATCH 2/2] bundle: keep a copy of bundle file name in the in-core bundle header

From: Junio C Hamano <gitster@pobox.com>
Date: Mar 2, 2016, 18:15 UTC
Message-ID: <xmqqfuw84uhb.fsf@gitster.mtv.corp.google.com>
In-Reply-To: <20160302090129.GC30295@sigill.intra.peff.net>

Jeff King <peff@peff.net> writes:

Show 11 quoted lines

> On Tue, Mar 01, 2016 at 03:36:26PM -0800, Junio C Hamano wrote:
>
>> This will be necessary when we start reading from a split bundle
>> where the header and the thin-pack data live in different files.
>> 
>> The in-core bundle header will read from a file that has the header,
>> and will record the path to that file.  We would find the name of
>> the file that hosts the thin-pack data from the header, and we would
>> take that name as relative to the file we read the header from.
>
> Neat. I'm hoping this means you're working on split bundles. :)

Let's just say that during the -rc freeze period, because I can stop looking at or queuing completely new topics to encourage people who are responsible for topics in the upcoming release to focus more on responding to regressions and follow-up fixes necessary, I have a better chance of having some leftover time to look into things myself, at least enough to figure out what needs to be done, ;-)

Show 20 quoted lines

>> -	if (strcmp(cmd, "create") && (bundle_fd =
>> -				read_bundle_header(bundle_file, &header)) < 0)
>> +	header.bundle_file = bundle_file;
>
> What are the memory ownership rules for header.bundle_file?
>
> Here you assign from either an argv parameter or a stack buffer, and
> here...
>
>> @@ -112,6 +111,8 @@ void release_bundle_header(struct bundle_header *header)
>>  	for (i = 0; i < header->references.nr; i++)
>>  		free(header->references.list[i].name);
>>  	free(header->references.list);
>> +
>> +	free((void *)header->bundle_file);
>>  }
>
> You free it.
>
> The call in get_refs_from_bundle does do an xstrdup().

Good eyes.

Show 10 quoted lines

>
> Should we have:
>
>   void init_bundle_header(struct bundle_header *header, const char *file)
>   {
> 	memset(header, 0, sizeof(*header));
> 	header.bundle_file = xstrdup(file);
>   }
>
> to abstract the whole procedure?

Maybe, maybe not. I'll decide after adding the bundle_version field to the structure (which will be read from an existing bundle, but which will have to be set for a bundle being created).

Thanks.