Re: squid 1.1.14+retry

On Thu, 31 Jul 1997, Henrik Nordstrom wrote:

> And did you remember to make clean? (always recommended after applying
> patches)

Actually, even I'm having trouble running it here, so I doubt make clean
will do much good. Get this:

-------------------
97/07/30 16:59:53|(5/3) comm_set_fd_lifetime: FD 30 lft 86400 (870382793)
                        maxfd=650
97/07/30 16:59:53|(5/9) commSetSelect: FD 30 READ handler=0x15ce8
                        data=0x2a2c40
97/07/30 16:59:53|(5/9) commSetSelect: FD 30 timeout=900 (870297293)
                        handler=0x150e8 data=0x2a2c40
97/07/30 16:59:53|(5/6) comm_select: FD 30 ready for reading

Program received signal SIGSEGV, Segmentation fault.
httpReadReply (fd=30, data=0x0) at http.c:549
http.c:549: No such file or directory.
(gdb) where
#0 httpReadReply (fd=30, data=0x0) at http.c:549
#1 0xa40c in comm_select (sec=1) at comm.c:1151
#2 0x1f246 in main (argc=3, argv=0xefbfda88) at main.c:702
(gdb)

--------------
Note that the READ handler data is very clearly set to 0x2a2c40, yet when
the httpReadReply() is called, the data argument is NULL instead of that
value. I'm currently completely mystified as to what's happening to the
fd_table[] entry, where, and why. If it's related to my patch, it's got
to be a pretty peculiar, Rube-Goldberg kind of relationship, since the
only time I access fd_table[] is to touch a variable that I myself added,
and nothing else.

I'm working on it, anyway, so the list will be the first to know when I
come up with anything.

        -Mike Pelletier.
Received on Thu Jul 31 1997 - 07:15:28 MDT