Re: Could some kind soul review these patches please.
| From: | Matt McClanahan | Date: | Wed, 13 Dec 2000 23:04:19 +0000 |
| Subject: | Re: Could some kind soul review these patches please. | ||
| References: | 1 | Groups: | php.dev |
| Request: | Send a blank email to php-dev+get-41243@lists.php.net to get a copy of this message | ||
On Wed, 13 Dec 2000, Zeev Suraski wrote:
> Personally, I'm not particularly in favour of moving extensions out of PHP,
> unless they're *REALLY* rarely used. I think that one of the good things
> about PHP, and one thing I hated in Perl, is that everything is in
> it. CPAN is impressive, no doubt, but I hated to have to search a zillion
> packages in CPAN just to get a script to run.
>
> I think the question we should be asking ourselves, always, is what the
> gain is. If we can come up with good reasons, then we should stay with the
> way it is, because it works.
>
> I fail to see the advantage of moving extension source to PEAR.
I'd like to see PEAR move in this direction. The benefits of having PHP
include all extensions are definitely strong ones. However, as more
extensions are added which apply to less users, PHP will inevitably grow
larger than is justified by the benefits. Sure, as somebody with no
worries regarding disk or cpu space, or bandwidth, I would probably prefer
to grab everything in one tarball and build it all from there. But I
think the time will come when this becomes impractical for users. In any
case, available resources isn't the only consideration.
It would be beneficial to have an authoritative and automated mechanism
for installing extensions on a system without recompiling PHP itself. It
seems to me that this would go a long way in fostering the development of
more extensions to PHP, assuming this is a direction that interests the
PHP group. If such a system was in place, I think it would be good to
pear down (So to speak) the number of extensions included in PHP. Some
candidates I would consider, for various reasons, as 'non-core' are ccvs,
cybercash, dotnet, fribidi, iisfunc, ircg, printer, and qtdom.
Mainly I'm thinking about which extensions have relatively limited
audiences, either due to an extension's purpose or portability. For
example, one of the key reasons I opted against submitting php_imlib for
inclusion into PHP is that it can't be used on non-Unix platforms. Still,
there are exceptions to everything, so portability can't be the only
criteria.
Matt