Hello Neil, thank a lot for the patch, it does the trick!
thank you also about the reminder about right style for code. Best regards, Georges. Neil Williams a écrit : > On Fri, 7 Aug 2026 12:06:21 +0200 Georges Khaznadar > <[email protected]> wrote: > > Hi Dominique, hi debian-devel@, > > > > I tried to fix perl-tk, because Bug#1113874, with no success so far. > > Rewrapped slightly to try and improve readability: > > my $do_gridded = ( > $gridded eq 'both' || ( > !$horz == ( > $gridded ne 'x' > ) > ) > ) ? 1 : 0; > > Seems it's the !$horz == ( at fault - it's appalling code. > > > The upstream repository, https://github.com/eserte/perl-tk/ shows no > > change since half a year; I raised an issue about it at > > https://github.com/eserte/perl-tk/issues/123 > > Upstream may not have had recent commits but the commit message for > https://github.com/eserte/perl-tk/commit/3a13aed7a07cd82347695b9595234f629803aef3, > dating from 2025, does seem relevant. > > Tk::Pane: fix precedence warning (new in perl 5.41.x) > Warning looked like this: > > Possible precedence problem between ! and numeric eq (==) at > .../blib/lib/Tk/Pane.pm line 297 > > > It resolves the precedence issue around $horz. > > From (!$horz == to ((!$horz) == > > > Unfortunately, updating the source package to its current head from > > github does not fix the issue. > > Check that this commit is retained after patching etc. during the > build. It's a small change, adding two brackets, easy to miss. > > > Last but no least: I have very poor skills about Perl and its test > > framework. Thank you in advance for any help or suggestion. > > Disclaimer: I'm emeritus for a reason and haven't debugged Perl in > probably 10 years. But I saw the post of -devel and had a look anyway. > I haven't tried to build the package. > > Just check that when you updated to latest upstream git that the effect > of this commit was retained and whether that changed the way that the > build failed or changed the messages. It's entirely possible that code > written this badly will cause secondary bugs. > > PS - this is a horrible way to write code in any language and someone > should be ashamed of themselves to let that into a release. It > desperately needs to be expanded and tested in more detail. > > > Best regards, Georges. > > > > Dominique Dumont a écrit : > > > Hi > > > > > > This bug breaks autopkgtest on libconfig-model-tkui-perl. > > > > > > As a consequence, this package was removed from testing along all > > > its reverse dependencies. > > > > > > Could you create a new release so that these packages can go back > > > to testing ? > > > > > > All the best > > > > > > > > -- > Neil Williams > ============= > [email protected] > https://www.codehelp.co.uk/
signature.asc
Description: PGP signature
