Re: Not-completely-portable modules

On Tue, 26 Oct 1999, you wrote:
> On Tue, 26 Oct 1999 15:00:08 -0400, Adam Turoff <[email protected]>
> wrote:
> 
> >> My feeling on it is that such modules should work in a *lot* of
> >> places.  I don't feel like including four billion Win32::* modules
> >> on the grounds that they're useful and platform-specific is okay.
> >
> >Including 2**32 Win32::* modules would be useful if they were in the
> >perl-win32-sdk, not the (core) perl-sdk.  
> >
> >Iff there is time/need/desire to address a series of mythical 
> >perl-foo-sdks,  that would be the forum to discuss things like 
> >Win32 (etc.) modules.
> 
> There is nothing mythical about the perl-win32-sdk (except that it is
> hidden under the libwin32 moniker):
> 
>     http://search.cpan.org/search?mode=dist&query=libwin32
> 
> -Jan
> 
> PS: And no, I don't think these Win32::* modules should be included in the
> perl-sdk; they are available (in binary form too) in the libwin32-* files
> and are also already preinstalled in the ActivePerl binary distribution.

No. They should be included. The ActivePerl binary distribution is irrelevant.
ActivePerl is a build, not perl itself, and does not define perl on the Win32
platform, despite ambitions.

These are crucial for any win32 perl programmer.

Supposedly there is now one perl. Splitting this up again is rubbing against
the fur. How much would be included that is unix-specific? Anything involving
shared memory, forking, etc. I don't think those should be excluded simply
because they're not completely portable.

Let's not have a double-standard about these things. The mainstream perl
community has a lot of folks who think Win32 perl should crawl in a hole and
die. If we go this route, we would have to exclude anything, from any platform,
including unix/linux, that was not completely portable.