istvan@hhb.UUCP (Istvan Mohos) (01/02/90)
Thank you very much, Greg A. Woods for posting the source; finally I was able to figure out why our Ultrix V2.2-1 (source-less) 'finger' kept dumping core. One of the gecos fields in /etc/passwd was xxyyzzy_mail_only_account changing 'char buffer[20]' to 'char buffer[BUFSIZ]' in matchcmp() while certainly an overkill, eliminated the problem once and for all. Wouldn't it be nice if everybody could agree on a max string size like 'ANSISTR 512'? Speaking of ANSI, cpp barfed on '#ident'. (Is '#ident' ANSI?) Had to comment it out to compile. -- Istvan Mohos ...uunet!pyrdc!pyrnj!hhb!istvan HHB Systems 1000 Wyckoff Ave. Mahwah NJ 07430 201-848-8000 ====================================================================
woods@robohack.UUCP (Greg A. Woods) (01/03/90)
In article <300@hhb.UUCP> istvan@hhb.UUCP (Istvan Mohos) writes: > Speaking of ANSI, cpp barfed on '#ident'. (Is '#ident' ANSI?) Had to > comment it out to compile. Hmm... I thought I'd been quite careful about the '#ident's. They should only be seen if you have SYSVR3 defined. It was with SysVr3.0 (as far as I can tell) that AT&T introduced the '#ident' pragma for entring .comment sections in the assembler output. There is no other need to define SYSVR3 in the finger.c I posted. As for buffers and such, I didn't do too much cleanup. There are still a lot of "magic numbers" scattered about, some of which place restrictions on data, others which define data, etc. It's not dumped on me yet :-), though I'll change those buffer declarations which deal with user data as you've suggested. -- Greg A. Woods woods@{robohack,gate,tmsoft,ontmoh,utgpu,gpu.utcs.Toronto.EDU,utorgpu.BITNET} +1 416 443-1734 [h] +1 416 595-5425 [w] VE3-TCP Toronto, Ontario; CANADA