netnews@cuae2.UUCP (Netnews Admin) (05/31/85)
I have seen numerous occasions where rn 4.3 will dump core with SIGSEGV while or just after adding new newsgroups and using "L" and a number to specify the location of the addition. A stack trace from sdb follows. Has anybody else seen this? We are running SVR2 on an AT&T 3B2. Thanks. ----- sdb stack trace begin ----- malloc(15) [malloc.c] safemalloc(size=15) [util.c:87] add_newsgroup(ngn=) [rcstuff.c:361] get_ng(what=,do_reloc=^A ) [rcstuff.c:288] newlist(munged=^A ,checkinlist=^@ ) [addng.c:89] initialize(argc=1,argv=0xc00202b0) [init.c:195] main(argc=1,argv=0xc00202b0,-1073610056) [rn.c60] ----- sdb stack trace end ----- -- Ron Heiby cuae2!netnews /app/eng AT&T Information Systems, Lisle, IL 60532
ekrell@ucla-cs.UUCP (06/04/85)
In article <303@cuae2.UUCP> netnews@cuae2.UUCP (Netnews Admin) writes: >I have seen numerous occasions where rn 4.3 will dump core with >SIGSEGV while or just after adding new newsgroups and using "L" >and a number to specify the location of the addition. A stack >trace from sdb follows. Has anybody else seen this? We are >running SVR2 on an AT&T 3B2. Thanks. I have seen a lot of rn crashes due to SIGSEGVs with the old rn. I haven't seen any crash since I installed rn 4.3. The crashes were all on a call to malloc (probably due to some string allocated via malloc being overwritten and garbling the pointers malloc() and free() use). I was never able to pinpoint the problem since by the time it core dumped it was too late (the actual error happened much sooner). It frequently crashed when using the "K" command. It would say "Badly compiled pattern" and coredump. -- Eduardo Krell UCLA Computer Science Department ekrell@ucla-locus.arpa ..!{sdcrdcf,ihnp4,trwspp,ucbvax}!ucla-cs!ekrell