RE: apr pool realloc?
From: Branko Cibej [mailto:[EMAIL PROTECTED] Sent: Saturday, July 24, 2004 7:17 PM William A. Rowe, Jr. wrote: At 04:06 AM 7/24/2004, Nick Kew wrote: Is there some fundamental reason why there's no apr_prealloc()? [...] I can't help thinking this should be (a) standardised. (b) optimised to work with pools. I've thought that for some time, with the understanding that you can't be playing in the pool you are reallocing. See the apr_psprintf() vformatter function for apr's internal implementation, and I'll offer a huge +1 if you care to abstract that behavior :) I thought Sander had apr_prealloc more or less implemented? Sander? I had, but there was objection against including it. I can dig it up ofcourse. Sander
RE: apr pool realloc?
On Sun, 25 Jul 2004, Sander Striker wrote: I had, but there was objection against including it. I can dig it up ofcourse. Yes please! If you (or anyone) can dig up a URL for the previous discussion and reason for the objection, that would be useful, too. A technical objection (risk of something breaking or less efficient than what I posted) would be important to know about; a philosophical objection can safely be overruled:-) -- Nick Kew
RE: apr pool realloc?
From: Nick Kew [mailto:[EMAIL PROTECTED] Sent: Sunday, July 25, 2004 2:20 PM On Sun, 25 Jul 2004, Sander Striker wrote: I had, but there was objection against including it. I can dig it up ofcourse. Yes please! If you (or anyone) can dig up a URL for the previous discussion and reason for the objection, that would be useful, too. A technical objection (risk of something breaking or less efficient than what I posted) would be important to know about; a philosophical objection can safely be overruled:-) The objection was mostly philosophical. Lack of demand. Sander
apr pool realloc?
Is there some fundamental reason why there's no apr_prealloc()?
I find myself working around this with constructs like:
typedef struct {
char* buf ; /* buffer */
size_t offset ; /* bytes used */
size_t avail ;/* bytes allocated */
} my_realloc_ctx ;
static void my_realloc(apr_pool_t* pool, my_realloc_ctx* ctx, size_t len)
{
char* newbuf ;
while ( len ctx-avail ) {
ctx-avail += BUFSZ ;
}
newbuf = realloc(ctx-buf, ctx-avail) ;
if ( newbuf != ctx-buf ) {
if ( ctx-buf )
apr_pool_cleanup_kill(pool, ctx-buf, (void*)free) ;
apr_pool_cleanup_register(pool, newbuf,
(void*)free, apr_pool_cleanup_null);
ctx-buf = newbuf ;
}
}
But I can't help thinking this should be
(a) standardised.
(b) optimised to work with pools.
--
Nick Kew
Re: apr pool realloc?
At 04:06 AM 7/24/2004, Nick Kew wrote: Is there some fundamental reason why there's no apr_prealloc()? [...] I can't help thinking this should be (a) standardised. (b) optimised to work with pools. I've thought that for some time, with the understanding that you can't be playing in the pool you are reallocing. See the apr_psprintf() vformatter function for apr's internal implementation, and I'll offer a huge +1 if you care to abstract that behavior :) Bill
Re: apr pool realloc?
William A. Rowe, Jr. wrote: At 04:06 AM 7/24/2004, Nick Kew wrote: Is there some fundamental reason why there's no apr_prealloc()? [...] I can't help thinking this should be (a) standardised. (b) optimised to work with pools. I've thought that for some time, with the understanding that you can't be playing in the pool you are reallocing. See the apr_psprintf() vformatter function for apr's internal implementation, and I'll offer a huge +1 if you care to abstract that behavior :) I thought Sander had apr_prealloc more or less implemented? Sander? -- Brane
