root@libove.UUCP (The Super User) (07/19/88)
I've been running the pathalias version 9 as modified for SCO Xenix for months now, and all of a sudden (possibly with the advent of a new map file?) it core dumps when it is most of the way through a processing run (it takes about twenty minutes to run when it used to work, and it dies now after about fifteen minutes) with the message: (): Segmentation violation I rebuilt pathalias with -DDEBUG and the error occurred and the message was: (): IOT trap Is anyone else having this problem suddenly? Does anyone know what those empty parenthesis signify? Any help appreciated! -- Jay Libove Internet: libove@cs.cmu.edu libove@andrew.cmu.edu 5313 Ellsworth Avenue formtek!ditka!libove!libove@pt.cs.cmu.edu Pittsburgh, PA 15232 UUCP: cmucspt!formtek!ditka!libove!libove (412) 621-9649 cadre!pitt!darth!libove!libove
jfh@rpp386.UUCP (John F. Haugh II) (07/21/88)
In article <57@libove.UUCP> root@libove.UUCP (The Super User) writes: > >I've been running the pathalias version 9 as modified for SCO Xenix for >months now, and all of a sudden (possibly with the advent of a new map >file?) it core dumps when it is most of the way through a processing >run (it takes about twenty minutes to run when it used to work, and it >dies now after about fifteen minutes) with the message: sounds like the same problem i have running pathalias on the 386. it runs out of memory after processing almost all of the maps. it seems that patalias requires about 2MB of memory (maybe less?) to process the current collection of maps. while on the topic, it seems that SCO Xenix does not actually page a process which is growing. from what i can tell of the pathalias problem, a process which is being grown is swapped, and if the image size is getting near the size of the swapper, you begin to get "warning: out of swap" messages, even though there appears to be plenty of physical memory to hold pages which can't be swapped or don't need to be swapped. grrrrr. - john. -- John F. Haugh II +--------- Cute Chocolate Quote --------- HASA, "S" Division | "USENET should not be confused with UUCP: killer!rpp386!jfh | something that matters, like CHOCOLATE" DOMAIN: jfh@rpp386.uucp | -- with my apologizes
jsilva@cogsci.berkeley.edu (John Silva) (07/22/88)
I had a similar problem with pathalias on my SCO V2.2.0g based 286 system. What hacking the *program* told me was 'out of memory', was in fact an *out of stack space* error. I increased pathalias' stack to something like 5000, and haven't had a problem since. Be aware that pathalias uses a recursive algorithm to generate the paths database. The stack overflow comes from the recursion being too many levels deep on a large path database. John P. Silva Inova Products --- UUCP: ucbvax!cogsci!jsilva DOMAIN: jsilva@cogsci.berkeley.edu
karl@sugar.uu.net (Karl Lehenbauer) (07/24/88)
Those having problems with memory space and/or execution time for pathalias on small, segmented systems may consider running pathalias over a subset only. For people in the USA, for example, only do usa or na, and forward the rest to your nearest smart mailer that does all the maps. -- -- backups: always in season; never out of style. -- karl@sugar.uu.net aka uunet!sugar!karl