Skip to content
Commits on Mar 17, 2016
  1. Bump version to

    committed Mar 17, 2016
  2. Merge pull request #145 from hvr/pr/integer-simple

    committed Mar 17, 2016
    Revert "Rename integer-simple to integer-pure (gh-138)"
  3. Merge pull request #144 from bgamari/master

    committed Mar 17, 2016
    Text.Array: Use unsafeFreezeByteArray# in place of unsafeCoerce#
Commits on Feb 13, 2016
  1. @hvr

    Revert "Rename integer-simple to integer-pure (gh-138)"

    hvr committed Feb 13, 2016
    This basically reverts commit b806d94.
    and therefore resolve #138
Commits on Jan 16, 2016
  1. @bgamari

    Text.Array: Use unsafeFreezeByteArray# in place of unsafeCoerce#

    bgamari committed Jan 16, 2016
    It shouldn't hurt to use the type-safe variant here. Moreover, this
    unsafeCoerce# usage was responsible for STG lint errors
    (GHC Trac #8114).
Commits on Dec 22, 2015
  1. Bump version to

    committed Dec 22, 2015
  2. Update changelog

    committed Dec 22, 2015
  3. Add benchmark for uncons

    committed Dec 22, 2015
  4. Fix mistakes in docs (gh-127)

    committed Dec 22, 2015
  5. Merge pull request #125 from Jubobs/master

    committed Dec 22, 2015
    Fix some typos
  6. Merge pull request #126 from DanBurton/master

    committed Dec 22, 2015
    Propagate decoding failure in the Get monad
  7. Merge pull request #135 from paulcapron/master

    committed Dec 22, 2015
    Implement takeWhileEnd
  8. Merge pull request #137 from yiding/patch-1

    committed Dec 22, 2015
    Fix typo in module comment
Commits on Dec 21, 2015
  1. @23Skidoo @hvr
  2. @hvr

    Fix PrintfArg fallout from 93e3705

    hvr committed Dec 21, 2015
    `PrintfArg` is extensible only from base-4.7 on, prior to that the
    `PrintfArg` class is not extensible.
  3. @hvr

    Define Semigroup instances and make `-Wcompat`-clean

    hvr committed Dec 21, 2015
    See also ekmett/semigroups#56 for the larger
  4. @hvr

    Kill several redundant-import warnings

    hvr committed Dec 21, 2015
    This refactoring avoids adding any additional evil CPP
    This has been tested across all major GHC 7.* releases, as well as GHC
  5. @hvr
Commits on Oct 21, 2015
  1. @yiding

    Fix typo in module comment

    yiding committed Oct 20, 2015
Commits on Oct 7, 2015
  1. @paulcapron

    Implement takeWhileEnd

    paulcapron committed Oct 8, 2015
    This commit adds a takeWhileEnd function to both Data.Text and Data.Text.Lazy,
    and related QuickCheck properties.
    takeWhileEnd, applied to a predicate `p' and a Text `t', returns the longest
    suffix (possibly empty) of elements in `t' that satisfy `p'.
Commits on Aug 12, 2015
  1. @aloiscochard
Commits on Aug 11, 2015
Commits on Jul 30, 2015
  1. @bgamari

    Release text-

    bgamari committed Jul 30, 2015
  2. @bgamari

    Show: Activate literal rewrite rules in simplifier phase 1

    bgamari committed Jul 30, 2015
    When rewriting in a rule, GHC pretends that it is in the phase given by
    the activation phase of the rule (see Note [Simplifying inside stable
    unfoldings] in `SimplUtils` for rationale). This means that while
    rewriting the text literal rules, the simplifier is effectively in phase
    2 due to 86ddf8a.
    As it turns out, the `build` rule provided by `base` is also active in
    phase 2, meaning it still competes with the literal rules (as 7.10.2
    uses it to rewrite the LHS). For this reason, the previous change did
    not resolve the performance issue in 7.10.2; something I did not
    previously catch due to a rather silly issue in my test envirionment.
    This issue is now properly resolved by activating the literal rewrite
    rules only in phase 1, by which point the `build` rule is inactive. See
    comments 22 through 25 in GHC Trac #10528 for details.
  3. @bgamari
Commits on Jul 28, 2015
  1. @bgamari

    Bump version to

    bgamari committed Jul 28, 2015
Commits on Jul 27, 2015
  1. @bgamari

    Show: Activate literal rewrite rules in simplifier phase 2

    bgamari committed Jul 27, 2015
    Previously these RULES were quite fragile as they could compete with the
    `unpack` rule defined in `GHC.Base`. This was tickled by 7.10.2, which
    happened to rewrite the LHS of rules with other rules (which is itself
    a bug in GHC). This caused the literal rules to fail to fire, which then
    resulted in long compilation times.
    To avoid this sort of fragility, we would like the literal rules to fire
    very early in the simplifier. For this reason we set them to rule in
    phase 2.
    See GHC Trac #10528 for further discussion [1].
Something went wrong with that request. Please try again.