Front page | perl.perl5.porters |
Postings from February 2012
>>>>> On Tue, 28 Feb 2012 20:19:57 -0500, Chris Prather <chris@prather.org> said: > The change log for Carp does include mention of this change. At the > time of the release no more than 35 modules out of >14K tested were > showing any possible change due to this patch, it seems reasonable > that a simple mention was all that was expected to be required. Nope. I'd say we blew it major way. Changing such a heavily used module in a rush a few days before freeze is not as excusable as you put it. Such a relatively irrelevant change should have been postponed to the early phases of the next release cycle. It was put into blead trunk. Instead of a smoke-me branch. And Steffen's parallel smoke setup should have been involved *before* a merge into blead trunk was even considered. > No-one (including Reini) pointed out more than a fraction of a > percentage of affected modules in the two weeks between when the patch > was made and the development release of Perl that this dual-lifed > module was bundled with was released. Sub::Uplevel was effected and reported. Since it is part of the toolchain it broke the smoke process for *everything* for a few days. > (Or if they did it is not recorded in this thread). Reini only > brought up that a group of unnamed people such as yourself had > issues with this change the day before the release. > The nice part about the development process for Perl5 is that any > interested parties, such as yourself, can participate to make sure > that your concerns are noted and dealt with *before* a release > happens. A cheap excuse. We should stand to our mistakes. Development should slow down in February and such changes should not be tried out in the trunk ever. Let's put this into our book of mistakes that should not have happened. -- andreasThread Previous | Thread Next