On Thu, Aug 22, 2002 at 09:55:32AM -0400, Jeff wrote:
> Brent Dax wrote:
> >
> > Jeff Goff:
> > # The mass of ICU code that's been added to Parrot. It's taking
> >
> > Have we determined that there are no programming language and
> > portablility issues yet?
> >
> > Are all the bits we need written in C? If not, we can't use it.
as is.
which is an important distinction. I don't know whether you were implying
this, but I think it's important to say it explicitly, because it's
consistent with this:
> The core is C. Some frilly interface bits that we'll need to rewrite are
> in C++ so they don't matter. And even should most of it be in C++, ICU
> is an overall win. Somewhat lamentedly I point out that nobody *else*
> has stepped up to the plate to write all of the code we need to become
> Unicode compliant.
>
> ICU happens to have all that code and more. Should we need to rewrite
> C++ down to C, at least we have a code base to start from instead of
> making all of the Unicode mistakes ab initio. Whether or not this is in
> C++ is immaterial, as we need a Unicode library, and preferably one
> that's already made the mistakes we would and fixed them.
I'd agree with this.
> > Does ICU handle Unix, Windows, VMS and Palm OS? If not, we can't use
> > it. (I figure if it handles those four, it's up to anything. :^) )
>
> Well, we don't handle PalmOS right now anyway, so that's a moot point.
> It handles the other three, and as I pointed out in my last post
> defending ICU's use, it's even got build scripts for OS/400, so right
> now it's more portable than Parrot. That's good enough for me.
Any code we wrote instead would have to be portable to PalmOS. I suspect
that it will take us less effort to port ICU to PalmOS than to start from
scratch. And I hope we'd be in a position to feed back ICU PalmOS patches
to the main ICU tree.
(the same goes for "should we include zlib in the perl5 core so that
Compress::Zlib can be assimilated?" vs "we don't know if zlib compiles on
all platforms that perl does" - if the code is already very portable, I
think we should re-use it, and put our finite efforts into fixing portability
problems we encounter.)
Nicholas Clark
