This is the mail archive of the
mailing list for the Cygwin project.
Re: perl-5.22.0-RC2 / Perl distributions
- From: Ken Brown <kbrown at cornell dot edu>
- To: cygwin-apps at cygwin dot com
- Date: Fri, 10 Jul 2015 15:45:41 -0400
- Subject: Re: perl-5.22.0-RC2 / Perl distributions
- Authentication-results: sourceware.org; auth=none
- References: <87iobhvj4f dot fsf at Rainer dot invalid> <5562881F dot 7000301 at cornell dot edu> <87iobh3zwg dot fsf at Rainer dot invalid> <55631FDB dot 10201 at cornell dot edu> <556350C5 dot 20600 at cornell dot edu> <87wpzw33ws dot fsf at Rainer dot invalid> <55636DAE dot 8020607 at cornell dot edu> <55A00F2B dot 6060401 at cornell dot edu> <1436556813 dot 10920 dot 10 dot camel at cygwin dot com>
On 7/10/2015 3:33 PM, Yaakov Selkowitz wrote:
On Fri, 2015-07-10 at 14:30 -0400, Ken Brown wrote:
On 5/25/2015 2:45 PM, Ken Brown wrote:
BTW, I contacted Philip Kime, the author of Biber, about
Unicode::Normalize. He told me that version 1.18 is unusably slow.
Here's a bug report he filed about it:
I think he's planning to pursue it further.
It doesn't look like anything is going to change in the near future.
According to the bug report I cited, Biber runs slower by a factor of 20
with version 1.18 of Unicode::Normalize than with version 1.17. I don't
think that's acceptable.
The only solution I can think of is to ship Biber as a self-contained
Perl Archive made with PAR::Packer[*], built with version 1.17 of
Unicode::Normalize. I've gotten the latter from backpan and built it
locally using the attached cygport file.
Does anyone have a better idea?
Just ship 1.17 instead as curr:?
Sure, if that's OK with Achim. I'm not sure what impact that might have
on other Perl modules that depend on Unicode::Normalize, if we have any
in the distro.