joachim@jrix.radig.de (Joachim Riedel) (05/26/91)
Hi, I now want to describe my problems with finding the correct fixstack- values for ported programs and am looking for additional help for a good "algorithm" instead of "try-and-error". ---------------------------------------------------------------------- SMAIL 2.5: I got SMAIL from piggy, installed it here on my coherent machine with the fixstack 1000 ... found in the Makefile. I polled my neighbour with starting uucico from the command line and all things seem to work. Mail arrives without problems. So I added an entry for uucico in crontab to poll my neighbour daily and automatically. There were five mails for me and after receiving them smail 2.5 core- dumped once per mail. Oops, what's that ??? Solved this mysterious bug with fixstack 2000 smail. Esa Ahola gave me also the hint that fixstack 1000 smail is not enough. Geoffrey Mendelson wrote that he uses fixstack 2800 smail. ------------------------------------------------------------------------- TASS 3.2: After successfully compiling TASS it immediate coredumped after starting it. Tried fixstack 1000 tass and it worked without problems with WNEWS 3.0. Now I got GNEWS 1.01b and made some modifications to get TASS 3.2 working with this program. All things seemed to work. So I wrote a test article. Oh, mistyping, wanted to reedit the article. Error Message: Can't open /.article... But I assured that the file was there ! What's wrong ?? Maybe permissions. So I checked the permissions. When I edited the article the first time, the UID/realUID, saved in variables after TASS is started, were ok. But not when trying to reedit the article, then I got UID=108, realUID=108. OK, pointer problem or array too small, someone destroys the values, tried to find that bug. But found none. Who corrupted the saved values ?? Just for fun I tried fixstack 1500 tass (1280 Bytes more) and it worked without problem. ---------------------------------------------------------------------------- I now wasted several hours for "hunting" mysterious bugs solved by "fixstack". Is the correct fixstack-value a riddle, to be solved by try-and-error or is it possible to found a rule for the correct value ? Joachim +---------------------------------------+-------+-------------------+-------+ | Joachim Riedel | @ @ | Don't worry, | @ @ | | Geschwister-Scholl-Strasse 48 | \_/ | keep smiling | \_/ | | D-6050 Offenbach am Main +-------+-------------------+-------+ | Tel. +49 69 85 62 25 | joachim@jrix.radig.de | +---------------------------------------+-----------------------------------+