On Tue, May 31, 2016 at 05:54:59PM +0200, Matthieu Moy wrote:
> William <william.duclot@xxxxxxxxxxxxxxxxxxxxxxx> writes:
>
> > On Mon, May 30, 2016 at 11:34:42PM -0700, Junio C Hamano wrote:
> >
> >> As long as your "on stack strbuf" allows lengthening the string
> >> beyond the initial allocation (i.e. .alloc, not .len), the user of
> >> the API (i.e. the one that placed the strbuf on its stack) would not
> >> know when the implementation (i.e. the code in this patch) decides
> >> to switch to allocated memory, so it must call strbuf_release()
> >> before it leaves. Which in turn means that your implementation of
> >> strbuf_release() must be prepared to be take a strbuf that still has
> >> its string on the stack.
> >
> > Well, my implementation does handle a strbuf that still has its
> > string on the stack: the buffer won't be freed in this case (only a
> > reset to STRBUF_INIT).
> > Unless I misunderstood you?
>
> I think Junio meant:
>
> void f()
> {
> struct strbuf sb;
> char buf[N];
> strbuf_wrap_preallocated(&sb, buf, ...);
> strbuf_add(&sb, ...);
>
> // is it safe to just let sb reach the end of its scope?
> }
>
> To answer the last question, the user would need to know too much about
> the allocation policy, so in this case the user should call
> strbuf_release(), and let it chose whether to call "free()"
> (OWNS_MEMORY) or not. This is OK with your implementation, but the doc
> needs to reflect this.
Okay, I understand. The idea was that the only change in the API is the
initialization: the API user can then use the strbuf like they could
before this patch, with the same responsabilities (having to release the
strbuf when they don't need it anymore).
Maybe a clearer documentation about the life and death of a strbuf
(initialization, use and release) would be useful, yes
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html