On Sat, 02 Jun 2007 10:22:52 +0200, Hans de Goede wrote: > >>> http://fedoraproject.org/wiki/Packaging/ReviewGuidelines:: > >>> > >>> - SHOULD: The reviewer should test that the package functions as > >>> described. A package should not segfault instead of running, for > >>> example. > >> I suggested that it the "SHOULD" be changed to "MUST". A package that > >> doesnt even start shouldnt be getting past reviews. > > > > For simple applications, sure, but for libs, modules and servers? And for > > example when the package consist in multiple commands should them all be > > tested? > > > > Exactly, there are very good reasons why this is a SHOULD and not a MUST. I'm > sure almost every reviewer will give a program a test run in the simple program > case / scenarion. "almost every reviewer" is not equal to "all packagers and all reviewers". Which is the reason why the guidelines include this SHOULD item. Many of the guidelines have their origin at fedora.us where the gained experience lead to appropriate policies. [Not even all packagers test their updates before pushing them to multiple dists. One of the biggest risks of %{dist}-aided mass-updates.] > Why must everything by regulated with rules, procedures and > more rules? Why can't we just TRUST each other, I'm getting very tired, sick > even, of this! +1 However, I agree that the second "should" in the guideline really ought to be replaced like this: - SHOULD: The reviewer should test that the package functions as described. A package must not (!) segfault instead of running, for example. -- fedora-devel-list mailing list fedora-devel-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/fedora-devel-list
