On Fri, May 24, 2013 at 02:03:47PM +0200, Johan Vromans wrote: > Never underestimate the benefits of being a core module. Yes. You now have the perceived advantages of being "blessed", "supported" and "installed". However true those statements aren't. > As a core module, CGI.pm is guaranteed to be there, and it will work. > If you want something quick and simple, CGI.pm is the obvious choice. This is not strictly true, as I believe that at least one point FreeBSD removed it from their base distribution of perl. But I don't know if that was a "glitch", later resolved by removing the need for perl to be in the base distribution (and hence the reason why they really wanted their perl to be smaller). But it's as good as true, because I'm not aware of any other OS or vendor doing something similar. And likely the places where this matters aren't running FreeBSD. > As a CPAN module, you'll have to install it yourself, if you can, and > there's no guarantee it will work. I'm having trouble parsing the last "it". In that, if you can install it (the module), then I'm failing to envisage a situation where the code wouldn't work, as the code is identical. Any more (or less) than your hosting provider (or their OS "vendor") is capable of screwing up packaged modules. But I totally agree with the point - there is no guarantee that you would be able to install a working version. The reason I quibble the detail is that as I can't see what could go wrong with the code itself (distinct from attempting the installation of the code) and hence how to make the code itself more robust. > And, if you want something quick and simple, use CPAN search and find > yourself drowned in houndreds of CGI handling modules. Good luck finding > one that suits your needs. And works. That tyranny of choice also makes it very hard to find a single replacement. > The bottom line is: Do we want to ship perl such that it is easy to get > started with simple web applications? If so we either need to include > CGI.pm in the core, or put it on CPAN *and*make*sure*it*works. > No change in maintenance burden. Yes. Which means that its current "need" for FCGI needs to be resolved first. Nicholas ClarkThread Previous | Thread Next