On Wed, Jun 24, 2015 at 09:38:20AM +0200, Bill Allombert wrote: > On Tue, Jun 23, 2015 at 11:29:12PM +0100, Dominic Hargreaves wrote: > > Source: pari > > Version: 2.7.4-1 > > Severity: important > > User: [email protected] > > Usertags: perl-5.22-transition > > Tags: sid stretch upstream > > > > This package FTBFS with perl 5.22 (currently in experimental), since > > these warnings are being treated as fatal: > > > > cd ../src/desc && /usr/bin/perl doc_make > ../../doc/usersch3.tex > > Unescaped left brace in regex is deprecated, passed through in regex; > > marked by > > <-- HERE in m/\\var{ <-- HERE flag}/ at doc_make line 41. > > make[2]: *** [usersch3.tex] Error 255 > > Makefile:146: recipe for target 'usersch3.tex' failed > > Thanks for your report, Dominic. > > However there is no warning in sid. Before turning warning to errors, it > would > be nice to let them act as warning for a bit, so that we get warned. > > As a consequence, I have no way to check if a patch fix the warning without > using experimental.
Hi, The message /is/ a warning, and is made fatal through the choice of the upstream software: https://sources.debian.net/src/pari/2.7.4-1/src/desc/doc_make/#L2 I'd be happy to test a patch prior to an upload if you wish (note that you would probably also need to configure [1] in addition to experimental; we don't recommend this on non-disposable systems). Incidentally, you might be interested in reading [2] on the perils of using this pragma. Cheers, Dominic [1] <https://people.debian.org/~dom/perl/test/perl-5.22.0/> [2] <http://www.modernperlbooks.com/mt/2014/01/fatal-warnings-are-a-ticking-time-bomb.html> -- To UNSUBSCRIBE, email to [email protected] with a subject of "unsubscribe". Trouble? Contact [email protected]
