FAQ
With PL_madskills enabled errors in BEGIN blocks were ignored.
But the opcode tree generated with this is often wrong because of
missing prototypes. And translating it would generate wrong results.

The patch removes the code to ignore the errors wehn PL_madskills is
set.


Gerard Goossen

Search Discussions

  • Rafael Garcia-Suarez at Apr 4, 2007 at 10:54 am

    On 29/03/07, Gerard Goossen wrote:
    With PL_madskills enabled errors in BEGIN blocks were ignored.
    But the opcode tree generated with this is often wrong because of
    missing prototypes. And translating it would generate wrong results.

    The patch removes the code to ignore the errors wehn PL_madskills is
    set.
    There's a similar code a few lines below. It should probably be removed too ?

    Also, why is there PL_minus_c in those tests ? Not being familiar with
    MAD code, I don't see the interaction with -c.
  • Gerard Goossen at Apr 4, 2007 at 2:53 pm

    On Wed, Apr 04, 2007 at 12:54:03PM +0200, Rafael Garcia-Suarez wrote:
    On 29/03/07, Gerard Goossen wrote:
    With PL_madskills enabled errors in BEGIN blocks were ignored.
    But the opcode tree generated with this is often wrong because of
    missing prototypes. And translating it would generate wrong results.

    The patch removes the code to ignore the errors wehn PL_madskills is
    set.
    There's a similar code a few lines below. It should probably be removed too
    ? Yes.
    Also, why is there PL_minus_c in those tests ? Not being familiar with
    MAD code, I don't see the interaction with -c.
    It appears the PL_minus_c is also set when the PERL_XMLDUMP is set
    (which also set the PL_madskills), but you can also set PERL_MADSKILLS
    in which case PL_madskills is set, but not PL_minus_c. Probably so you
    can test that enabling PL_madskills doesn't have any effect on execution of
    the code (though I have never tried PERL_MADSKILLS).

    Gerard Goossen
  • Rafael Garcia-Suarez at Apr 6, 2007 at 1:36 pm

    On 04/04/07, Gerard Goossen wrote:
    On Wed, Apr 04, 2007 at 12:54:03PM +0200, Rafael Garcia-Suarez wrote:
    On 29/03/07, Gerard Goossen wrote:
    With PL_madskills enabled errors in BEGIN blocks were ignored.
    But the opcode tree generated with this is often wrong because of
    missing prototypes. And translating it would generate wrong results.

    The patch removes the code to ignore the errors wehn PL_madskills is
    set.
    There's a similar code a few lines below. It should probably be removed too
    ?
    Yes.
    Thanks. So I applied that as change #30851. (Sorry for being slow...)
    Also, why is there PL_minus_c in those tests ? Not being familiar with
    MAD code, I don't see the interaction with -c.
    It appears the PL_minus_c is also set when the PERL_XMLDUMP is set
    (which also set the PL_madskills), but you can also set PERL_MADSKILLS
    in which case PL_madskills is set, but not PL_minus_c. Probably so you
    can test that enabling PL_madskills doesn't have any effect on execution of
    the code (though I have never tried PERL_MADSKILLS).

    Gerard Goossen

Related Discussions

Discussion Navigation
viewthread | post
Discussion Overview
groupperl5-porters @
categoriesperl
postedMar 29, '07 at 3:59p
activeApr 6, '07 at 1:36p
posts4
users2
websiteperl.org

People

Translate

site design / logo © 2021 Grokbase