Front page | perl.perl5.porters |
Postings from November 2013
* Father Chrysostomos via RT <perlbug-followup@perl.org> [2013-09-17T17:17:30] > On Tue Sep 17 12:18:25 2013, LeonT wrote: > > Can't we just have a policy of removing features that were deprecated in > > 5.000 without further discussion? Pretty please? > > I would like to see a policy that deprecated feature are removed when > they get in the way of further developement, and not otherwise. The > lack of a policy either way is what causes discussion. Deprecated features get in the way of future development in more than one way. Sometimes, as I believe you're thinking of, the code or syntax for the feature is actively in the way of a presently-desired change. In these cases, removing the deprecated feature is generally a no-brainer. The feature is slated for eviction, its space is needed now, so it must go. Another way, though, is by acting as dead weight. Each deprecated feature is one more thing to be understood when one's eyes fall on its code. It's one more thing that might get a bug report which will lead to time wasted discussin whether a bug ought to be fixed even if feature is deprecated. If the code is altered *without* understanding, then the risk of new bugs (and thus new bug reports) is increased. Long deprecated-but-unremoved behaviors are Damoclean swords. Many of these have been deprecated so long that nobody quite remembers why or knows how well-used they are, and so I think it is wise to do a quick check with the list before committing changes. Otherwise, deprecated features should be removed when their deprecation cycle has ended, the debate presumably having already occurred. Heck, I don't even mind one last, "Wait, I have a new objection!" I just object to living in a room with twenty swords dangling over me. -- rjbsThread Previous | Thread Next