On Tue, 25 Nov 2008, Charles Seeger wrote: > +------ Patrick Vervoorn wrote (Tue, 25-Nov-2008, 09:59 +0100): > | On Mon, 24 Nov 2008, Charles Seeger wrote: > | > | > +------ Patrick Vervoorn wrote (Sat, 22-Nov-2008, 18:29 +0100): > | > | I'm still here, even using trn, though it is getting more and more > | > | unstable as time goes by. Using 'valgrind' I also notice trn is quite > | > | sloppy with (de-)allocating it's memory... > | > > | > FWIW, I haven't been seeing these problems (on Solaris 8 SPARC). > | > | Hmmm, interesting. I've been seeing these problems both with my > | 'home-compiled' version of trn, but also with the trn distributed via > | Debian. Al running on Linux-x86. I, unfortunately, do not have access to > | another architecture. > > I would expect compiler, library or source bugs before anything to > do with cpu architecture, per se. Yes, well, I was trying to indicate that re-compiling trn with another/newer compiler and/or other libraries, did not make a difference for the memory-related crashes. Moving to another cpu-architecture, would definately change the compiler- and library-related variables for sure. > Ours is an amazingly old version > (4.test67), built with Sun's compiler (C 4.2 with patches 104667-10 and > 104668-05, according to the elf .comment section reported by 'mcs -p'), > on Solaris 2.6 way back in 1998. test76 was released in 2001. I'm running the latest -test76 version of it. > Perhaps I'm just lucky with the groups that I read and what our > NNTP (INN) server dispenses. The most annoying problem that I see > is the thread selector sometimes dropping to the end of a group after > reading a thread in the middle. It seems to happen repeatedly in > a given group during the same reading session, so I suspect is has > something to do with the set of articles, e.g. broken references or > some such. ISTR that Wayne had some fixes in the later test releases, > but haven't been annoyed enough to hunt it down. I've noticed what you mentioned before, but didn't find it a big nuisance. A simple press of Ctrl-r when you're at the end, puts you back at the beginning. As for my memory-related problems, I do not encounter these when I read 'regular' text-groups, even when they're pretty big and/or have a rather big retention. I do encounter this when skimming through binary groups, with up to several hundred k's of article-overviews being pulled in. So it's possible the usenet-flow wasn't available in april 2001 when test76 was released, so it's never been tested with these amounts of articles...? Regards, Patrick.