Another quiet week on perl5-porters; but big patches were proposed, demonstrating that the porters are not dead yet. Read about a revamp of the parser, an in-depth modification of the internals, and other bugs and associated fixes.
While fixing bug #25824, Dave Mitchell thought it would be a good idea to
shave the yacc and to move the perl parser implementation from byacc to
bison. And so did he. He noticed also that the new bison parser actually
seems to be better at error recovery. Byacc is still used for a2p, though.
http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2004-02/msg00331.html
While he was at it, Dave added a verbose flag to the -Dp debugging
command-line option to perl (-Dpv), to display the top 8 elements of
the LR parser stack at each reduce.
Tassilo von Parseval is trying to fix a long standing bug : you can't, currently, assign a glob to a lvalue (e.g. an element of a tied hash.) The primary cause of the bug is that you can't upgrade, technically speaking, a PVLV to hold a GV. He proposed a first patch with a rather hackish approach, but Dave Mitchell raised the opinion that the only sane solution is to expand the PVLV structure so that it can incorporate all the fields of a PVGV. Tassilo then proposed a second patch to implement this. This last patch modifies the order in which the different types of SV are upgraded in the core.
http://groups.google.com/groups?selm=20040214110256.GA14993%40ethan
Tels released Math::BigInt 1.70 and Math::BigRat 0.12. As of this writing,
it's yet unclear whether it fully fixes bug #26559 (problems with big
integers and the & bitwise operator), bug #26522 (some weird
overload-related problem, if I understand correctly), and bug #26508
(problem with the power of a negative big integer.) Those bugs are
described here
http://groups.google.com/groups?selm=rt-3.0.8-26508-77180.15.818688458454%40perl .org
http://groups.google.com/groups?selm=rt-3.0.8-26522-77237.7.75701084030025%40per l.org
http://groups.google.com/groups?selm=rt-3.0.8-26522-77237.7.75701084030025%40per l.org
Rafael checked in bleadperl a patch (which he proposed a while ago) to
implement a little bit of syntactic sugar for chains of filetest
operators; it renders the combined expression -f -w -x $file equivalent
to -x $file && -w _ && -f _.
http://groups.google.com/groups?selm=20040209231251.5988eb2c.rgarciasuarez%40fre e.fr
Dominic Dunlop proposed to remove the support for the obsolescent MachTen operating system. Perl 5.8.x doesn't compile anymore on MachTen, anyway.
Wolfgang Loch reported (bug #26587) that File::Find used to clobber $_.
This was fixed recently, by adding a local($_) at the appropriate
place; but Mark-Jason Dominus points out that this is still a potential
bug, and needs to be a local(*_).
http://groups.google.com/groups?selm=rt-3.0.8-26587-77915.4.38321984527676%40per l.org
This summary was written by Rafael Garcia-Suarez. Weekly summaries are published on http://use.perl.org/ and posted on a mailing list, which subscription address is perl5-summary-subscribe@perl.org . Corrections and comments are welcome.
Re:groups.google.com messing up
rafael on 2004-02-17T10:09:47
I have no idea. I'm pretty sure it used to be there. (Further proof that all email archives suck.)