On Sun, 08 Nov 2009 18:55:56 +0000, petdog wrote: > Looking at the git log, the "64bit article id" commit seemed the most > likely to be at fault, and in fact, if I compile pan from 10 commits > back, it behaves correctly.
Skimming the diff, I noticed that it changed the logic in Numbers::to_string (pan/usenet-utils/numbers.cc). And indeed, pan/usenet- utils/numbers-test fails at the first test. Reverting to the old code puts things back on track. _______________________________________________ Pan-users mailing list [email protected] http://lists.nongnu.org/mailman/listinfo/pan-users
