Re: [PEPr] Proposal for HTTP::HTTP_Spider

On Fri Mar 05, 2004 at 01:1547PM -0500, Matthew Feinberg wrote: > > >* Why is PEAR.php included, if no feature that is provided by this file > > > is used? > > Good question. No good answer. I wrote the class without pear for something > I was working on, and now I'm trying to 'Pearize' it. I felt I should use > PEAR because 'all the other classes are doing it'. I'll stop extending pear. > > Is it okay to just return false instead of a pear error? That depends on the situation: In case you want to provide detailed information about the error, which caused the script to return false, you should use raiseError(). _Personally_ I prefer to always raise PEAR_Error instances in methods, which are directly exposed to the user or the view in MVC applications. I then catch those errors with a application-wide error handler method, which renders a adequate view then. > Is it preferred to > return a pear error even if you only need a boolean so that future > development is easier, There is no strict rule about that, but I'd choose that way. > > >* There is a bunch of cool HTTP packages in PEAR, which should be used > > > within HTTP_Spider. (E.g. HTTP_Request) > Agreed. I am especially interested in HTTP_Client, because of its Cookie > support, though it is listed as beta. I'll look at integrating > > >HTTP_Request. HTTP_Request was just an example. Maybe there is another package, which fits even better into HTTP_Spider. > > >* I do not think that it is a good idea to directly print debugging > > > messages to stdout. Instead, Log should be used to faciliate a > > > fine grained logging mechanism. > > I haven't looked into Log before. Great idea. From what I know, there is no other class, which is using Log for this purpose at the moment. Nonetheless I like the idea :). -- - Martin Martin Jansen http://martinjansen.com/