More perlish would be:
insert_head => unshift insert_tail => push
Although these method-names are perfect.
I should have mentioned that we want to stay as close as possible to the C API and change things only when things aren't quite perlish (like when C passes its arguments by reference, in order to return more than one argument). Otherwise we try to mostly stick to the C API. That's in order to make the life of those who have to write modules in Perl and C Apache API.
So, in this case, the glue code is this:
static MP_INLINE
void mpxs_APR__Brigade_insert_tail(apr_bucket_brigade *brigade,
apr_bucket *bucket)
{
APR_BRIGADE_INSERT_TAIL(brigade, bucket);
}static MP_INLINE
void mpxs_APR__Brigade_insert_head(apr_bucket_brigade *brigade,
apr_bucket *bucket)
{
APR_BRIGADE_INSERT_HEAD(brigade, bucket);
}where APR_... are C macros privided by APR.
> next($b) => next() > prev($b) => prev() > > I don't know whether my conception is right but $bb is something like an > iterator. Why does one have to pass $bb->next($b) $b, wouldn't > $bb->next() that's how iterator work, at least i'd expect that.
bucket brigade is not an iterator, it's a container for buckets, which internally is just a linked list. So when you say next($b), it does something like $b->next, but it has no way to tell whether it hits the wall or not. In C it looks like:
#define get_bucket(brigade, bucket, fetch) \ (fetch(bucket) == APR_BRIGADE_SENTINEL(brigade) ? \ NULL : fetch(bucket))
static MP_INLINE
apr_bucket *mpxs_APR__Brigade_next(apr_bucket_brigade *brigade,
apr_bucket *bucket)
{
return get_bucket(brigade, bucket, APR_BUCKET_NEXT);
}which can be translated as:
APR_BUCKET_NEXT(bucket) == APR_BRIGADE_SENTINEL(brigade) ? NULL : APR_BUCKET_NEXT(bucket);
gimme the next bucket, unless it's a sentinel bucket.
That wall is artificial and needed because internally it's implemented using a ring:
/**
* The magic pointer value that indicates the head of the brigade
* @remark This is used to find the beginning and end of the brigade, eg:
* <pre>
* while (e != APR_BRIGADE_SENTINEL(b)) {
* ...
* e = APR_BUCKET_NEXT(e);
* }
* </pre>
* @param b The brigade
* @return The magic pointer value
*/
#define APR_BRIGADE_SENTINEL(b) APR_RING_SENTINEL(&(b)->list, apr_bucket, link)I agree with you that ideally, you should be able to just say:
$b_next = $b->next;
and it should work. The problem is that the bucket has no idea which brigade it belongs to. There is not much we can do about it. We just provide the perl glue to the existing C code.
If you are after confusing functions, I find much more confusing the following APR::Bucket APIs:
# insert_after
# insert_before
-- __________________________________________________________________ Stas Bekman JAm_pH ------> Just Another mod_perl Hacker http://stason.org/ mod_perl Guide ---> http://perl.apache.org mailto:[EMAIL PROTECTED] http://use.perl.org http://apacheweek.com http://modperlbook.org http://apache.org http://ticketmaster.com
-- Report problems: http://perl.apache.org/bugs/ Mail list info: http://perl.apache.org/maillist/modperl.html List etiquette: http://perl.apache.org/maillist/email-etiquette.html
