This extends the list of Perls used on the Travis continuous integration service to include up to the current stable Perl version.
I've checked that this change works on Travis, and have found that it doesn't, which is very odd. Namely Perls 5.20 and 5.22 fail (5.24, however passes), and always in the same test (t/01-messages.t line 58). What's more odd is that I can't reproduce the failure on my local Debian (stretch) box: both 5.20 and 5.22 work as expected. Since I can't work out what's really going on here (the test and production code looks fine; nothing obviously wrong) so this PR is in some sense a heads up that this issue exists. At present it's very difficult to tell if the problems turn up on cpantesters, since any test failures are swamped by the misleading OTRS Perl::Critic test failures. Perhaps you have an idea as to why these Perls are failing?
This extends the list of Perls used on the Travis continuous integration service to include up to the current stable Perl version.
I've checked that this change works on Travis, and have found that it doesn't, which is very odd. Namely Perls 5.20 and 5.22 fail (5.24, however passes), and always in the same test (t/01-messages.t line 58). What's more odd is that I can't reproduce the failure on my local Debian (stretch) box: both 5.20 and 5.22 work as expected. Since I can't work out what's really going on here (the test and production code looks fine; nothing obviously wrong) so this PR is in some sense a heads up that this issue exists. At present it's very difficult to tell if the problems turn up on cpantesters, since any test failures are swamped by the misleading OTRS Perl::Critic test failures. Perhaps you have an idea as to why these Perls are failing?