From: Ramkumar Ramachandra <artagnon@gmail.com>
Date: Sun, 13 Nov 2011 10:40:12 GMT
Subject: Re: [PATCH 1/5] sequencer: factor code out of revert builtin
Message-ID: <CALkWK0n7v15n_s3CNq1Qu3LHjYkV-ENAkv2b+oB+VBkyV+Sphw@mail.gmail.com>
In-Reply-To: <20111106001232.GC27272@elie.hsd1.il.comcast.net>

Hi,

Small note.

Jonathan Nieder writes:
>> --- a/sequencer.c
>> +++ b/sequencer.c
>> @@ -1,7 +1,27 @@
>>  #include "cache.h"
>> +#include "object.h"
>> +#include "commit.h"
>> +#include "tag.h"
>> +#include "run-command.h"
>> +#include "exec_cmd.h"
>> +#include "utf8.h"
>> +#include "cache-tree.h"
>> +#include "diff.h"
>> +#include "revision.h"
>> +#include "rerere.h"
>> +#include "merge-recursive.h"
>> +#include "refs.h"
>> -#include "sequencer.h"
>> -#include "strbuf.h"
>>  #include "dir.h"
>> +#include "sequencer.h"
>
> Why did sequencer.h move to after dir.h?

1. I like the convention of including the "foo.h" as the last header
in "foo.c".  I suppose it has to do with the way I include standard
headers in my own code (for pet projects).
2. I didn't want to include many of these headers in sequencer.h again
-- it uses a lot of these data types.  I've noticed that ordering of
header inclusion is important in many parts of Git, so the convention
just stuck.

Thanks.

-- Ram