Front page | perl.perl5.porters |
Postings from November 2007
Re: Dropping 5.5 support from my modules.
Thread Previous
|
Thread Next
From:
H.Merijn Brand
Date:
November 22, 2007 12:45
Subject:
Re: Dropping 5.5 support from my modules.
Message ID:
20071122204435.70473d5b@pc09.procura.nl
On Thu, 22 Nov 2007 19:05:07 +0000, Nicholas Clark <nick@ccl4.org> wrote:
> On Wed, Nov 21, 2007 at 05:53:00PM -0600, Brandon Black wrote:
>
> > VMS stands out to me as a target because (a) It's a really small chunk
> > of people still using it and (b) It's very different from the core
> > *nix/wintel platforms that are most widely used and supported, which
> > raises the cost of maintenance. When I say "cost of maintenance", I'm
> > mostly thinking in terms of how much my eyes bleed while I try to read
> > through the source to understand something, and how much that slows me down.
>
> There's a counter cost in portability. If we go "it runs on both kinds of
> operating systems, Windows and Linux", how long before that comes down to
> "both kinds of operating systems, Redhat and Debian", and then
> "both kinds of operating systems, Debian and Ubuntu"
>
> Whereas if we're portable to 4 or 5 distinct families of platforms, we knock
> out a lot of the assumptions that would otherwise prevent perl from
> compiling on the next big thing. Symbian is strange, but Jarkko got Perl to
> build there.
>
> > Contrast this to, say, IRIX. There aren't that many IRIX users out
> > there (compared to the major platforms anyways). IRIX has been
> > officially EOL'd by SGI (although they've said they'll offer support
> > services through 2013). But there's no compelling case for eliminating
> > IRIX support in Perl, because the maintenance cost is very low. The
> > differences between implementing Perl on IRIX and mainstream *nix are
> > very low.
>
> IRIX is a bitch. We've found (at least) one bug in SGI's compiler pre-
> processor that even SGI weren't aware of, and the 64 bit compiler knows
> the ANSI C standard better than most everyone here (probably except Jarkko
> and Abhijit)
But there are a HUGE number of GNU gcc users out here in the wild, and my
testing GNU gcc 64bit on HP-UX with perl revealed several bugs, that were
fixed because of my testing on HP-UX in the post 3.0.4 track.
We interact with Open Source software developers outside the perl business.
See also Schwern's discussion with the nroff/groff folk just recently. We
all learn from eachother, and as long as the other party doesn't deliberately
NOT cooperate (no, I didn't /say/ IBM/, I just thought that), we should get
the most out of that as possible, and that includes feedback from VMS users,
ENCDIC users (however difficult that might be, and reminds us of how many
CPAN modules are NOT EBCDIC compliant (also my Text::CSV_XS)).
> Don't get me wrong - I prefer IRIX to AIX or HP/UX (or aaaaaaaaaaargh NEC
> Super UX) but it's hard work to keep it safe without a test system.
> Likewise actually for AIX and HP/UX. In particular, IBM's xlc is a wonderful
> test of the quality of your code.
>
> The VMS compiler also knows its standard, but its error messages are helpful,
> and we have the availability of test systems to play on.
>
> > Windows is arguably just as burdensome to support as any of the above
> > from a *nix person's perspective, but it's got a tremendously larger
> > audience (half of our userbase on the survey right?), so it just can't
> > be dropped, however much anyone hates it.
>
> Actually right now we have more VMS developers active here than Win32.
Were are the djgpp users? If you think VMS is hard, try djgpp. :)
I gave up, and sticked to Cugwin.
--
H.Merijn Brand Amsterdam Perl Mongers (http://amsterdam.pm.org/)
using & porting perl 5.6.2, 5.8.x, 5.10.x on HP-UX 10.20, 11.00, 11.11,
& 11.23, SuSE 10.1 & 10.2, AIX 5.2, and Cygwin. http://qa.perl.org
http://mirrors.develooper.com/hpux/ http://www.test-smoke.org
http://www.goldmark.org/jeff/stupid-disclaimers/
Thread Previous
|
Thread Next