martin (02/12/83)
Further to what Jim said about uuxqt buffers being to small, i would recomend thath in the newnews documentation, where it talks about putting in the -x and -n flags to ux/uuxqt, it also talk about increasing the buffers correctly. if the site doing the reply does not fix it, then when the mail gets down the link an rmail will find the Zapped mail address, hence wasting network time. even if you fix it at yourn site, all that means is that the mail will propogate thru you. Maybe Jay was right, mail/uucp is the subject for a PHD, but how do you get it running on all these sites?. (there are still people not running 2.9 news) martin.
lepreau (02/12/83)
Re which buffers in rmail to make bigger: assuming you are running rmail.c, just look at the declarations in main() and enlarge all the obvious ones to 256 or 512 or something. You don't have to understand it; there are a bunch called "from", "user", "to" and the like, of teeny size. Make 'em all big and you won't be bothered. Mempry is cheaper than your time. (If your rmail is a link to binmail, dunno.) -Jay
mark (02/13/83)
Note that it wasn't rmail that dumped core, but uuxqt. Thus, the buffer that must be increased is in uuxqt.c. This one has been mentioned a number of times in net.bugs.uucp. My Tektronix UUCP bug list #37 credits Jonathan Hahn at Ampex with finding it first - his fix is to increase MAXFULLNAME from 100 to BUFSIZ.