[perl #113666] [BUG] Memory panic due to string transcoding for every match

[Gerhard R.]

# New Ticket Created by  "Gerhard R." 
# Please include the string:  [perl #113666]
# in the subject line of all future correspondence about this issue. 
# https://rt.perl.org:443/rt3/Ticket/Display.html?id=113666 >


The regex-dna benchmark script (
https://github.com/perl6/perl6-examples/blob/master/shootout/regex-dna.p6.pl
) eats all available memory on 32-bit systems.

The issue can be golfed to

my $string = 'x' x 3;
my @matches = $string.match(/./, :global);

Keeping around a transcoded version of the string for every match is
to blame according to analysis by jnthn and pmichaud:

http://irclog.perlgeek.de/perl6/2012-06-01#i_5667563

(occurs with Rakudo 2012.05-237-geef772d, but hasn't been fixed in master yet)

[perl #115248] [BUG] Cannot read lines from STDIN with current Parrot

[Gerhard R.]

# New Ticket Created by  "Gerhard R." 
# Please include the string:  [perl #115248]
# in the subject line of all future correspondence about this issue. 
# https://rt.perl.org:443/rt3/Ticket/Display.html?id=115248 >


Applies to Parrot RELEASE_4_8_0-130-g6236d2d and earlier.

The issue is described in detail in the following Parrot ticket:

 https://github.com/parrot/parrot/issues/860

It is unclear if the problem lies with the Parrot or Rakudo codebase.

An (untested) workaround would be to pass an explicit value to the
optional record_separator argument of the readline() method of the
Handle PMC, invoked in method get() of class IO.

This of course wouldn't fix the problem of overwiting the handle's
record_separator attribute with garbage.

[perl #116373] [BUG] Error when using (supposedly) built-in subrule in grammar

[Gerhard R.]

# New Ticket Created by  "Gerhard R." 
# Please include the string:  [perl #116373]
# in the subject line of all future correspondence about this issue. 
# https://rt.perl.org:443/rt3/Ticket/Display.html?id=116373 >


Example code:

$ perl6 --version
This is perl6 version 2012.12-87-gc3f565d built on parrot 4.11.0
revision RELEASE_4_11_0-135-g73c6d32

$ perl6 -e '(grammar { token TOP {  } }).parse("42").say'
<1357738945>Can not get attribute '$!pos' declared in class
'Cursor' with this object
  in regex TOP at -e:1
  in method parse at src/gen/CORE.setting:10624
  in block  at -e:1

[perl #117117] [BUG] Module compilation failure with nested packages

[Gerhard R.]

# New Ticket Created by  "Gerhard R." 
# Please include the string:  [perl #117117]
# in the subject line of all future correspondence about this issue. 
# https://rt.perl.org:443/rt3/Ticket/Display.html?id=117117 >


My work on https://github.com/gerdr/m42 revealed a module compilation
bug: While everything works fine without it, after compilation not all
packages will be registered.

A golfed version of the module structure can be found at
https://github.com/gerdr/rakudobug

The module load order is identical between the pm and pir versions and
can be found at https://gist.github.com/gerdr/5130590

-- gerdr

Re: Building MoarVM on Solaris

[Gerhard R.]

Hi.

My other reply to Andy Dougherty is probably still awaiting moderation.

However, I did misunderstand his problem with 'mknoisy' when
specifying a compiler.

The issue he encountered is due to the fact that compilers are
associated with default toolchains, so switching to gcc on Solaris
will generate a Makefile for GNU make.

You can work around that by providing an explicit override via --toolchain.

Alternatively, you can use --cc instead of --compiler, which does not
affect other settings.

-- gerdr

Re: More patches for building MoarVM on Solaris

[Gerhard R.]

Appreciated.

> The only really invasive non-obvious one is patch 4.  Solaris cc doesn't
> support anonymous unions, but MVMOSHandleBody uses them extensively.
> I had to give the anonymous union a name everywhere in order to get it
> to compile.

Is this actually necessary if you pass -features=extensions ? The
manual is a bit unclear on that...

-- gerdr