Permalink
Commits on Dec 10, 2016
  1. Remove needless call to infix:<,> in [sS]/// codegen

    One more needless call to remove for same reasons as those removed
    in a34f0af
    zoffixznet committed Dec 10, 2016
  2. Remove useless p6store

    zoffixznet committed Dec 10, 2016
  3. Remove outdated comment

    zoffixznet committed Dec 10, 2016
  4. Remove needless calls to infix:<,> in [sS]/// codegen

    After a discussion[^1] and testing the output of old tickets[^2],
    it looks like these are no longer needed. After their removal, the
    bugs referenced by tickets do not re-appear and stresstest passes.
    
    [1] https://irclog.perlgeek.de/perl6/2016-12-10#i_13717904
    [2] https://irclog.perlgeek.de/perl6/2016-12-10#i_13718025
    zoffixznet committed Dec 10, 2016
  5. fix accidental creation of lists in prematch and postmatch

    timo committed Dec 10, 2016
  6. Simplify Match.{pre,post}match

    moritz committed Dec 10, 2016
  7. Make @a[2;2] about 40% faster

    By assuming that the @indices have always already been reified
    completely.  I haven't been able to come up with a way to not have
    them reified (other than calling postcircumfix:<[; ]> directly).
    So the assumption seems safe, for now at least.  Please revert if
    this causes some ecosystem fallout, spectest itself is clean.
    lizmat committed Dec 10, 2016
  8. Add changelog entry

    moritz committed Dec 10, 2016
  9. Corect Match.prematch and Match.postmatch for zero-width matches, ufo…

    …bat++
    moritz committed Dec 10, 2016
  10. Make infix:<..> on Nums 13x faster + fix native num NaN ===

    And anything else that uses === on Nums should be faster too.
    jnthn++ for finding the cause of slowdown.
    
    The presence of &infix:<===>(NaN, NaN) candidate prevents
    caching[1], which causes massive slowdown in infix:<..> on Nums.
    The candidate also does not handle native nums as they go to a
    narrower candidate, so NaN === NaN is broken too.
    
    Fix by removing (NaN, NaN) candidate and adding a check for NaNs
    into other candidates. Any num but NaN is less-or-equal to positive
    infinity, so use that as an `nqp::isnan()` check.
    
    Fixes RT#130190: https://rt.perl.org/Ticket/Display.html?id=130190
    
    [1] https://irclog.perlgeek.de/perl6/2016-12-09#i_13713052
    zoffixznet committed Dec 10, 2016
Commits on Dec 9, 2016
  1. Make .match(:g) return empty List on match failure

    As a side-effect of Great .match Refactor[^1], .match(:g) on failure
    to match was changed from returning an empty List to returning
    an Empty (which is a slip). As a consequence, the idiom of
    nqp::p6store( nqp::call(&infix:<,>, $/), $_) (where $/ is the slip), in
    Perl6/Actions that handle S:g/// resulted in an empty list, rather than
    the actual original string, which is the cause of the bug in the ticket[^3]
    
    During the G.mR, the .comb on failure was also changed to return an Empty,
    instead of empty Seq. This commit affects that behaviour as well,
    returning an empty List instead.
    
    [1] b7201a8
    [2] https://github.com/rakudo/rakudo/blob/c4a60110ef/src/Perl6/Actions.nqp#L7796-L7803
    [3] https://rt.perl.org/Ticket/Display.html?id=130289
    zoffixznet committed Dec 9, 2016
  2. Revert "Simplification of LAST block"

    This reverts commit 58cdfd8.
    
    The change was not semantically identical and caused problems in
    some situations.
    lizmat committed Dec 9, 2016
  3. Merge pull request #943 from MasterDuke17/RT77070

    Add .gist, .perl, and .Str to BOOTSTRAPATTR
    lizmat committed on GitHub Dec 9, 2016
  4. Fix a number of issues in parse-base() error reporting

    - Fix missing sign in error message
    - Fix incorrect char referenced if invalid char is the first char of
        fractional part of the number
    - Fix incorrect char referenced in negative numbers, when the
        invalid char is in fractional portion of the message
    
    Fixes RT#130298: https://rt.perl.org/Ticket/Display.html?id=130298
    zoffixznet committed Dec 9, 2016
  5. Update name of param in a comment

    zoffixznet committed Dec 9, 2016
  6. Add .gist, .perl, and .Str to BOOTSTRAPATTR

    MasterDuke17 committed Dec 9, 2016
Commits on Dec 8, 2016
  1. Fix sink statement prefix not exploding Failures

    In the Actions it code-genned to call .eager. Not sure
    why (copy-pasta from the <eager> in code above?), but
    changing that to call .sink showed no stresstest failures.
    
    Bug finding: https://irclog.perlgeek.de/perl6/2016-12-08#i_13706422
    zoffixznet committed Dec 8, 2016
  2. Fix matcher-less .first not respecting its :k, :p, and :kv adverbs

    The multi handling the matcher-less candidate does not check whether
    any adverbs were passed.
    
    Fix by removing that candidate and assigning `True` default to a
    candidate with Mu matcher.
    
    IRC: https://irclog.perlgeek.de/perl6-dev/2016-12-08#i_13706306
    zoffixznet committed Dec 8, 2016
  3. Fix .first on Numerics

    The logic in the method uses `+self` to get number of elements in
    the list, however, `self` is not guaranteed to be a list and may be
    a Numeric type, in which case we get an incorrect number.
    
    Fix by using `self.elems` instead.
    
    Bug find: https://irclog.perlgeek.de/perl6/2016-12-08#i_13705826
    zoffixznet committed Dec 8, 2016
  4. Merge pull request #942 from Altai-man/empty-emit-fix

    Simplification of LAST block
    lizmat committed on GitHub Dec 8, 2016
  5. Simplification of LAST block

    Altai-man committed Dec 8, 2016
  6. Merge pull request #941 from Altai-man/empty-emit-fix

    Empty emit fix
    zoffixznet committed on GitHub Dec 8, 2016
  7. Tabs->spaces indentation fix

    Altai-man committed Dec 8, 2016
  8. Merge pull request #940 from samcv/nqpbump

    NQP bump: Add nice error message when using \K in regex and suggest
    zoffixznet committed on GitHub Dec 8, 2016
  9. NQP bump: Add nice error message when using \K in regex and suggest

    Using <( or )>
    samcv committed Dec 8, 2016
Commits on Dec 7, 2016
  1. Support superscript powers that are larger then we could store in an …

    …int.
    pmurias committed Dec 7, 2016
  2. (nqp bump) Fix for nqp::pow_I with large exponents

    zoffixznet committed Dec 7, 2016
  3. Merge pull request #939 from lucasbuchala/patch-1

    De-indent 4 spaces the "perl -h" message output
    zoffixznet committed on GitHub Dec 7, 2016
  4. Fix MVMArray Error on .pick with negative numbers

    ...by restricting range to UInt. However, for some ops, like infix:<xx>
    we treat negatives as zero (see
    https://irclog.perlgeek.de/perl6/2016-12-07#i_13698488)
    
    I'm commiting this as a fix for very LTA error, but will leave the
    ticket[^1] open. I will review what we do for most of our methods to see
    if there's any sort of consistency we can make out of it.
    
    [1] https://rt.perl.org/Ticket/Display.html?id=130284
    zoffixznet committed Dec 7, 2016
  5. Merge pull request #938 from rakudo/fix-powers-with-big-exponents

    Stop relying on a bug in infix:<**>(Int, Int).
    pmurias committed on GitHub Dec 7, 2016
  6. De-indent 4 spaces the "perl -h" message output

    lucasbuchala committed Dec 7, 2016
  7. Stop relying on a bug in infix:<**>(Int, Int).

    pmurias committed Dec 7, 2016
  8. Fix poor error with infix:<xx> with non-numeric Str

    When '123aaa' is given, the Failure occurs inside the coercion,
    which we then attempt to store in a native int, triggering a bad error
    
    Fix by moving the coercer into the Num() multi, where such a Failure
    would explode in the .Int method.
    
    Fixes RT#130288: https://rt.perl.org/Ticket/Display.html?id=130288
    
    P.S: this is probably a deeper issue with our coercers; IIRC there's
    a ticket that they don't check the final type.
    zoffixznet committed Dec 7, 2016
  9. Bump to get latest MoarVM.

    * Fixes a couple of premature frees in async sockets when errors
      occur
    * Optimizations for Int -> int coercion
    jnthn committed Dec 7, 2016