Now, ActiveState took a pretty big step when they made fork() available in their Perl for Win32 product (I'm not sure if they did the same for Python). But, there's more to life than fork - etc, process info, native threads - the list goes on.
I have noticed that there's a decent selection of Win32 modules on CPAN, but I'm learning that anything that doesn't explicitly fall under the Win32:: namespace either does a poor job of supporting Win32 or doesn't support it at all. And, of course, anything that *does* fall under the Win32:: namespace (*horror*) only works on Win32 systems.
I've been working to rectify this somewhat in Ruby, which is lacking compared to Perl and Python when it comes to Win32 support, IMHO. However, I think all three languages (which can, naturally, more or less share source code) could use improvement in this area. Here are some ideas I've been tossing around:
Well, that's all I can think of at the moment. Oh, I should also mention that I would entirely focus on NT/2000/XP. Win95 and Win98 users can get stuffed - they're dead and no one uses them in any sort of serious production environment anyway. I don't want source code sullied by a bunch of #ifdef's just to make something work on Windows 95, for example. Too much cruft to support for no gain. Of course, the same could be said for BeOS, but hey, I actually *like* BeOS.
Poor Win32 support? (Score:1)
I'm not quite sure what you mean by this and one of the reasons I use Perl is that my experience has been exactly the opposite - most things work well across platforms. Certainly the big-name modules like LWP, DBI, XML::* have not given me any trouble due to Win32 issues.
The biggest problem I've encountered on Win32 is that I don't have
Re:Poor Win32 support? (Score:2)
Part of the problem I think is that a lot of the hard-core C programmers won't, don't or can't program on the Win32 environment and so it ends up an afterthought.