develooper Front page | perl.perl5.porters | Postings from January 2010

Re: Use ICU for Unicode in Perl?

Thread Previous | Thread Next
From:
Craig A. Berry
Date:
January 29, 2010 07:44
Subject:
Re: Use ICU for Unicode in Perl?
Message ID:
c9ab31fc1001290744x34e2f31cg22091404d220b7db@mail.gmail.com
On Fri, Jan 29, 2010 at 3:18 AM, Nicholas Clark <nick@ccl4.org> wrote:
> On Thu, Jan 28, 2010 at 09:34:35PM -0700, karl williamson wrote:
>> I found only a couple of references to the ICU Unicode C language
>> library in the p5p archives.  Is there a reason not to use some of this
>> open source code when it suits our purposes?
>
> Part of it, maybe. But

And then the question arises about what parts of it can live
independently of other parts.  For example, does one still need C++ if
one only wants the C API, and can any of the code be used without all
of the (huge) data files?

It's a big package; the tarball is bigger than Perl's.  A quick count
of the code shows:

.h, 153,500 lines in 657 files
.c, 216,073 lines in 255 files
.cpp, 575,757 in in 516 files

and it looks like there are a lot more data than code in the package.

> Building International Components for Unicode on UNIX requires:
>
>    * A C++ compiler installed on the target machine (for example: gcc, CC, xlC_r, aCC, cxx, etc...).
>    * An ANSI C compiler installed on the target machine (for example: cc).
>    * A recent version of GNU make (3.80+).

It also includes solution files for building with Microsoft tools, so
the requirement for GNU make would appear to be a fairly thin one.

> Note that:
>
> 1: VMS is not in their list

I've never heard of a port nor could find one in a quick search.
Given that it runs on z/OS and Windows, I'd say it looks portable
enough that it should be possible.

> 2: It doesn't look like they've changed from an m*n configuration system - ie
>
>                              AIX            HP/UX
>   vendor compiler            works          works
>   g++                        works          you need to port!
>
>
>   ie they chose not to abstract out OS specific parts and compiler specific
>   parts from their configuration system
>
> 3: gnu make has its own list of dependencies.
>
>
> And if it's all for UTF-16, then how much does that help us when we're using
> UTF-8 internally?

My impression was that karl is interested in borrowing and adapting
piecemeal rather than trying to outsource the whole Unicode problem or
adding libicu as a dependency.  Grabbing a function or two that do
something useful sounds promising, though it's hard to say without
digging in how possible that is.

Thread Previous | Thread Next


nntp.perl.org: Perl Programming lists via nntp and http.
Comments to Ask Bjørn Hansen at ask@perl.org | Group listing | About