On Sun, Mar 03, 2002 at 09:55:25PM -0800, Ryan Bloom wrote:
> Justin,
>
> This fixes the second of your two bugs. Thanks for finding those, I'm
> sorry I let them slip through. BTW, the number of bugs that this code
> exposed has convinced me that this is the correct direction to go. The
> reason I say that is that all of the bugs that have been exposed are
> happening because we tried to hack around problems with the filters.
Yeah, I think we hit a hot spot of problems. Better to find and
get it working now. Obviously, this wasn't as easy as we all
thought it should be.
I'm getting a SEGV when requesting / now (not an infinite loop)
at util_filters.c:347.
(gdb) print *r_filters->next->next->next
$12 = {frec = 0x8101800, ctx = 0x0, next = 0x81a6400, prev = 0x8164b40,
r = 0x81a3e40, c = 0x81648d8}
(gdb) print *r_filters->next->next->next->frec
$14 = {name = 0x8101670 "net_time", filter_func = {
out_func = 0x80c7d70 <net_time_filter>,
in_func = 0x80c7d70 <net_time_filter>}, ftype =
AP_FTYPE_HTTP_HEADER,
next = 0x0}
(gdb) print *r_filters->next->next->next->next
$13 = {frec = 0x65746e6f, ctx = 0x6c2d746e, next = 0x75676e61,
prev = 0x3a656761, r = 0xa6c7020, c = 0x746e6f43}
Somehow something is corrupted. I have a few more shows to watch
on my TiVo before I take a look at it tonight. =) If you beat
me to analyzing it, great. If not, I'll try to tackle it. -- justin
