Re: [SMARTY-DEV] Strict PHP5 Compatibility
[email protected] (boots)
| Newsgroups | php.smarty.dev |
|---|---|
| Message-ID | <[email protected]> |
--- Ruben Vermeersch <[email protected]> wrote: > On Wed, 2004-09-01 at 17:18 +1000, Robert Amos wrote: > > The question is, aside from private/protected/public variables and > > methods, what benefits and improvements would you make to smarty > > that are php5-only ? > > > > Smarty already benefits from the improved object handles and the > > speed improvements to foreach(), etc. So what else is there that > > would warrant a php5 branch? > > Well, i'd mostly drop all backwards compatibility stuff and see where > it's possible to squeeze out just a little more of smarty, by using > all the new stuff we got. Okay, but what? That's the question that I think bok was asking. A few people (myself included) gave ideas on that topic in the last couple of months, but nothing that compelling over all. Perhaps it is me -- I don't yet have a good enough feel for where the strengths and weaknesses of PHP5's feature set lay but then again, I only have it installed on one dev machine and there are no specific plans to migrate our PHP4 setups (both dev and production). > I'd also prefer error handling based on Exceptions, and the ability > to use PHP5 with a E_STRICT error reporting set (although PEAR is also > very bad in this). > However, I'm not saying we should branch it right away, I love the > current state (and who am i to say there should be a branch) and i'd > like to stay to the HEAD mainline for all development IMO it is way to early to be thinking of a PHP5 only branch. I'm sure it will come, but without any motivating reasons, I can't see anyone spending much time on it. The best thing about the current Smarty is that it is backward and forward compatible with PHP4/5 so everyone wins. Breaking BC to support things like private/protected and Exceptions (which I hear are terribly slow) doesn't really seem like an advancement in the state-of-the-art to me. xo boots > Greetings, > Ruben