On Sun, May 13, 2001 at 09:51:37AM +0200, Sander Striker wrote: > I thought that pool usage was mandatory and that malloc/free should > never be called directly in apr (except within apr_pool, and now > in apr_sms).
Nope. Cscope comes up with 43 calls to malloc inside of apr (some of it is aplibtool.c - which doesn't count and some others that are #ifdefd out). There is a bunch more stuff in apr-util and httpd-2.0 that use malloc/free, but that may not be of concern to you guys (it is of concern to me though). If you use the direct calls, you are responsible for managing it (which isn't a big deal). The question is whether we should switch all of these calls outside of apr to apr_sms_t semantics instead of using malloc/free directly. By using apr_sms_t, you will get a more consistent interface than is currently offered (I'd love to see the malloc/memset pattern go away). > *lol* you must have really gotten frustrated with this stuff to actually > buy the standard. It is good to know someone actually has the standard > and can quote from it :-) I'll shut up about standards per Greg's latest email. Whatever works the best for APR is what we should implement. If it happens to be related to any published standards, so be it. -- justin
