Universal request objects

Ovid has asked for the wisdom of the Perl Monks concerning the following question:

Replies are listed 'Best First'.
Re: Universal request objects by dragonchild (Archbishop) on Dec 23, 2005 at 19:17 UTC
You're not looking for Catalyst::Request - you're looking for Catalyst::Engine. In the C::R docs, it says that it's filled in by C::E. So, you need C::E subclasses that DWIM, plus some dispatcher that uses the right C::E subclass. Sounds to me like you just want to use Catalyst and go from there. My criteria for good software: Does it work? Can someone else come in, make a change, and be reasonably certain no bugs were introduced?	[reply]
Re^2: Universal request objects by Ovid (Cardinal) on Dec 23, 2005 at 19:19 UTC
Actually, we don't want to use Catalyst. There is a Catalyst component, but it's optional. Thus, if the users don't want to install it, they shouldn't be dependent on it. Cheers, Ovid New address of my CGI Course.	[reply]
Re^3: Universal request objects by dragonchild (Archbishop) on Dec 23, 2005 at 19:38 UTC
Break out the parts you want, create a new CPAN module, and offer it back to the Catalyst devteam as a separate distro that's useful without the rest of Catalyst. Then, Catalyst can use it as a dependency. My criteria for good software: Does it work? Can someone else come in, make a change, and be reasonably certain no bugs were introduced?	[reply]
Re^3: Universal request objects by adamk (Chaplain) on Dec 26, 2005 at 03:07 UTC
The Catalyst::Engine standalone rewrite is called Isotope, and although it's being worked on, the author is apparently somewhat famous for not commiting very often. Some time early in the new year, I'm going be (hopefully) helping to get Isotope $done, so I can replace the engine-layer of my own web application system with it. With you and I and a few others, I reckon we should be able to push it through to completion reasonably promptly. What do you say? :)	[reply]
Re^4: Universal request objects by Ovid (Cardinal) on Dec 27, 2005 at 17:35 UTC
Re: Universal request objects by samtregar (Abbot) on Dec 23, 2005 at 19:02 UTC
CGI.pm works pretty well, with the possible exception of SOAP, depending on how you're taking requests. It definitely works for CGI, mod_perl and on the command-line. -sam	[reply]
Re^2: Universal request objects by Ovid (Cardinal) on Dec 23, 2005 at 19:05 UTC
Right. Doesn't work for SOAP or CLI. I know people have worked on this problem before, so I assume there's something out there. I just don't know if it's open source or not :) Cheers, Ovid New address of my CGI Course.	[reply]
Re^3: Universal request objects by samtregar (Abbot) on Dec 23, 2005 at 19:17 UTC
CGI does work on the command-line. Witness: `$ perl -Ilib htdocs/story.pl 'rm=edit&story_id=1' Content-Type: text/html; charset=ISO-8859-1 <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN" "http +://www.w3.org/TR/html401/loose.dtd"> ...` [download] If you don't like that way of loading params you can load them yourself after calling `CGI->new()`. -sam	[reply] [d/l] [select]
Re^4: Universal request objects by Ovid (Cardinal) on Dec 23, 2005 at 19:24 UTC
Re: Universal request objects by Joost (Canon) on Dec 23, 2005 at 23:23 UTC
CGI.pm works fairly well on the command line, it's very good as the mod_perl and CGI interface. SOAP adds a whole new abstraction level on top of HTTP so it just plain can't be interchangeble with basic CGI unless you're willing to make a lot of assumptions (i.e. constict the SOAP interface). In the same way, CGI.pm's command line interface doesn't allow you to easily simulate POSTs and HTTP headers (though you can, if you're willing to set environment variables and feed STDIN). If I were you I'd probably go with CGI or CGI::Simple's (edit: or CGI::Lite::Request's) input parsing interface and expand that with whatever SOAP functionality you want (but as I don't know what you expect of the SOAP interface this is a hard question to answer in more detail). "What should it profit a man, if he should win a flame war, yet lose his cool?"	[reply]
Re: Universal request objects by nothingmuch (Priest) on Dec 23, 2005 at 23:50 UTC
http://dev.catalyst.perl.org/repos/Catalyst/trunk/Isotope/ -nuffin zz zZ Z Z #!perl	[reply]
Re: Universal request objects by ph713 (Pilgrim) on Dec 23, 2005 at 23:56 UTC
I really wish people would stop recommending CGI.pm for anything. It was a workhorse in its time, but for the love of dog let it die. CGI::Simple is a really good replacement without all the html crap.	[reply]
Re^2: Universal request objects by blahblahblah (Priest) on Dec 24, 2005 at 02:26 UTC
Thanks for pointing this out. It looks interesting. However, in the bottom of the docs it says: BUGS As this is 0.01 there are almost bound to be some. Not recommended for production use. That worries me a little. Is that just the author being humble, or is this a fairly new and untested module? One nice thing about CGI.pm's widespread use is knowing that others have already worked out all the strange bugs.	[reply]
Re^3: Universal request objects by ph713 (Pilgrim) on Dec 24, 2005 at 04:45 UTC
CGI::Simple is fairly robust in my book. That comment has (obviously) been there since 0.01, but it's at version 0.77 now. I've used it in a couple of production environments with great success. (I'm the "Brandon Black" credited at the bottom of the docs for doing some testing and bugfixing on this module way back when).	[reply]
Re^2: Universal request objects by diotalevi (Canon) on Dec 24, 2005 at 16:55 UTC
I thought getting an HTML generator was one of the reasons why CGI.pm is so great! You'll take CGI from me over my dead body. HTML widget construction is supposed to be as easy as CGI.pm makes it and I know of no replacements for it. Care to englighten me? ⠤⠤ ⠙⠊⠕⠞⠁⠇⠑⠧⠊	[reply]