Re: Extension publication (RSS)

On Mon, 2012-04-16 at 17:58 +0200, Mattias Lindholm wrote: > > It's easy to do quick implementations for reading RSS feeds, but the > point of having this as an extension is to carry out the entire RSS > validation for the developer. What would happen if you just loaded > something other than a valid RSS feed running the example code you > wrote? With the extension one is supposed to simply initiate a > resource from a source (file, string) and then directly inspect any > error code rather than the entire document for errors with PHP code. > This would minimize the amount of effort required by the developer as > well as improve performance on parsing RSS feeds. (might be minor > differences, but should still be faster to do it through an extension) > As RSS is a quite common feed format, this would also prevent > developers from reinventing the wheel. I think having proper libraries for that using PHP can bring the same benefits, plus the fact that it is simpler to deploy and maintain PHP over C code. Yes, it will be a tiny bit slower, but well... And well, there's not only RSS, but also RDF and Atom and others being used. :-) I scrolled quickly over the code. A thing I noticed was that the library does abort() in OOM situations, in PHP you usually don't want to abort (imagine a threaded web-server, an abort() in one thread and the whole process goes down instead of handling it more graceful) Another thing i noted is that it seems to work on files only, not using PHP streams, while PHP streams are one of the things which make PHP so powerful ... This all is no veto against the package in PECL, if you fix the license it should be ok ... johannes