Bojan Smojver wrote: > On Fri, 2007-08-31 at 11:17 +0100, Joe Orton wrote: >> it could >> be PR 39075 or 41105. > > After applying that fix in r571651, the test now fails and it's PR 41105 > (not enough space in grbuf, I guess). > > Given comments in that bug report, Bill,are you OK with applying the fix > from the trunk (r532789) to 1.2.x? If you still dislike that large stack > allocation, we can also create a subpool, allocate 8k in there and then > destroy the pool upon exit. Much slower though...
It will be picked up... > If you agree with Joe's fix in r532789, we should also backport r390410, > which is the same for userinfo.c I don't care for it, but it solves the problem -today- instead of tomorrow, and it's certainly legit. To make me happy we need apr_pool_realloc which is hidden within apr_vformatter() - but that can be a 1.3 sort of thing. Let's take it, if someone beats me to the commit, great. Bill
