Re: [PATCH] branch as a builtin (again)

"Kristian Høgsberg" [off-list ref] writes:

Ok, once more without the spaces.  I have to state that it's against
my personal beliefs using pointers as boolean values, but I can go
with the flow here.  For extra bonus, I'm using xrealloc instead of
plain realloc now.

My preferences (pretty much procedural):

 - Documentation/SubmittingPatches
   - Attachments discouraged;
   - With a proper commit log message;
   - With a proper signed-off line;

 - Names in source encoded in utf8 if needed (I think you got
   this one right, but application/octet-stream does not give
   charset information so I cannot really tell).

 - No spaces between a function name and open parenthesis.

Some nitpicks.

static void create_reflog(struct ref_lock *lock)
{
...
}

Probably reflog interface should supply ways to create new ones
(and delete or truncate existing ones) to users like this
program.  Please work with Shawn Pearce to refactor this part.

static void create_branch(const char *name, const char *start,
			  int force, int reflog)
{
	struct ref_lock *lock;
	unsigned char sha1[20];
	char ref[PATH_MAX], msg[PATH_MAX + 20];

You are using snprintf so I think it is safe, but I think using
PATH_MAX for msg length is wrong.  start can be an arbitrary
extended object name expression (HEAD^12~24^2~4^^2~13...) and
does not have much to do with pathname.

	snprintf(ref, sizeof ref, "refs/heads/%s", name);

Maybe barf if snprintf steps over the buffer?