Re: [RFC_v3] Handling Backwards Compatibility in PEAR [hopefully the final draft]
| From: | Stefan Neufeind | Date: | Wed, 24 Sep 2003 22:04:06 +0000 |
| Subject: | Re: [RFC_v3] Handling Backwards Compatibility in PEAR [hopefully the final draft] | ||
| References: | 1 | Groups: | php.pear.dev |
| Request: | Send a blank email to pear-dev+get-21949@lists.php.net to get a copy of this message | ||
On 24 Sep 2003 at 23:58, Jan Schneider wrote:
> Zitat von Stefan Neufeind <stefan@neufeind.net>:
>
> > On 24 Sep 2003 at 23:06, Tobias Schlitt wrote:
> >
> > > Am Wednesday, September 24, 2003 8:44 PM [GMT+0100=CET],
> > > artikulierte Matthias Nothhaft <php@mahono.de>:
> > >
> > > Hi Matthias!
> > >
> > > > Well, I remembered another BC issue:
> > > > What about PHP5/6/7/8... -> language/syntax changes ?
> > > > Take the PHPUnit package for example!
> > >
> > > Maybe that should be handled in a kind of new file branch. The
> > > installer should be able to handle the correct package for the PHP
> > > version. After all another open issue would be how long a
> > > maintainer should keep to PHP branches allive.
> > >
> > > What about that?
> >
> > Isn't this "automatically" solved? A new major of PHP comes out only
> > after a very long time / preparation. If there are fixed needed e.g.
> > for PHP5 that also work with PHP4 (stricter code-checking or more
> > warnings or so) then there is no problem fixing them in the current
> > release cycle. And if you need fixes that demand PHP5 you could add
> > a dep on PHP >= 5.0 already, right? For a new major version, which
> > is then directed at PHP5 upward I mean.
>
> There still is the subtle case where you have to rename a public
> identifier because it became a reserved word. In this case a new major
> release should be made even if it's sort of a bugfix. But it's an API
> change after all.
Yeah, that's what I meant by "automatically solved" in my email. Just
follow "the spirit of the RFC" :-)) Renaming a (public) identifier is
an API-change and therefor needs a new major. Nothing to discuss ...
we all know what to do :-))
But fine that you also seem to agree on the thoughts of the RFC.
(Greg - nice work!!!)
Stefan