On Mon, Dec 14, 2009 at 6:44 AM, Toshio Kuratomi <a.badger@xxxxxxxxx> wrote: > Just to make clear, Nicolas's interpretation is correct. Attempting to work > around the problem by language lawyering does not promote better software. Another specific non-library case.... the pytz package was shipping its own timezone definitions separate from the tzdata package that required additional maintenance (stupid shifts in daylight savings time..thanks US congress.) I was told about it, added a patch to pytz and now it reads the tzdata resource files instead of shipping its own. Less work for me as a maintainer long term.. one less aperiodic package update for users...and more consistent timezone handling for users. The intention of the guidelines...is to guide people in using their judgement on how to handle things. Now in the pytz case as soon as I was made aware of the duplication of timezone resource files..it was obvious that I should make a best effort to reuse a common system wide set and it was easily done. But sometimes its not obvious and that's when a peer discussion needs to happen. Or sometimes its obvious but a best effort runs into problems because of upstream customization or tweaking and that's when a peer discussion needs to happen. The guidelines help define the boundary of the grey area when discussion really needs to happen. -jef -- fedora-devel-list mailing list fedora-devel-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/fedora-devel-list
