On Sun, 2002-06-30 at 21:30, Justin Erenkrantz wrote: > On Sun, Jun 30, 2002 at 09:14:20PM -0700, Brian Pane wrote: > > The brigade can't be in a filter context, because then it wouldn't be > > accessible from ap_rgetline_core(). The request_rec is the only place > > where we could easily put this brigade. We just need to name it > > something like "temp_working_brigade" and document the fact that any > > function that puts things in that brigade must clear them out before > > it exits. > > This brigade can't be used by anything other than ap_rgetline_core(). > I think allowing it to be used by *anything else* is going to cause > problems later on. I just finished cleaning up most of the invalid > references to req_cfg->bb and now you want to add it back. > Documenting it as temp_working_brigade is just going to declare > open season on misuse. > > *sigh* > > > The pool cleanup registration and cancel operations aren't particularly > > inefficient (though I think there's some room for optimization). The > > problem is the large number of calls to these operations per request, > > due to the multiple brigade creations/deletions per header line. > > It should only be one brigade per-line, right?
It should only be one. But it's not. The speculative read to check for a '\t' at the start of the next line uses a separate brigade. Hmmm...I might as well fix that now; it will speed things up a little while we figure out the best solution for the more general problem of temp brigade management. > How about adding a variant of ap_rgetline_*() that takes in a > temporary brigade and ap_get_mime_headers() can create one brigade to > be used for all headers. +0. If we create the temp brigade in ap_get_mime_headers(), we'll still be doing the brigade creation/deletion for the first line of the request. We really need to create the temp brigade in ap_read_request(), and then pass it through a variant of ap_get_mime_headers() that takes the temp brigade as an input. --Brian
