Permalink
Commits on Sep 16, 2015
  1. v0.55

      - spelling error, sync with Debian [Lucas Kanashiro]
    committed Sep 16, 2015
  2. changelog++

    committed Sep 16, 2015
Commits on Sep 2, 2015
  1. Gna, confusing typo in ChangeLog...

    Obviously we return *array* not hash refs.
    committed Sep 2, 2015
Commits on Sep 1, 2015
  1. v0.54

      - new function matchr() which returns results as hashref
        instead of hash (RT#106515) [requested by djerius@cpan.org]
    committed Sep 1, 2015
  2. changelog++

    committed Sep 1, 2015
  3. test: for new matchr() function

    The actual match() tests are copied over from
    t/data_dpath.t, so here we only compare whether
    match() and matchr() return the same results.
    committed Sep 1, 2015
  4. docs++

    committed Sep 1, 2015
  5. rt#106515: provide separate function 'match' and 'matchr'

    - matchr() provides an arrayref to results
    - match() dereferences that to provide existing api
    committed Sep 1, 2015
  6. provide empty line for dzil VERSION tweaking

    Avoids message like this:
    
      [@SCHWIGON/PkgVersion] no blank line for  after package Data::DPath::Filters statement on line 24
    committed Sep 1, 2015
  7. separate function '_all' into getting refs and deref

    ...in order to use the ref later on its own.
    committed Sep 1, 2015
Commits on Jun 3, 2015
  1. v0.53

      - no import from POSIX (RT#104914) [reported by ether]
    committed Jun 3, 2015
  2. changelog++

    committed Jun 3, 2015
  3. no import from POSIX

    The default import list is *huge*,
    and pollutes the local namespace.
    committed Jun 3, 2015
Commits on Feb 27, 2015
  1. v0.52

      - Make sure feature.pm is only "test.recommends", as is optional prior
        to 5.10 (GH #4, kentnl)
      - skip NYTProf files in dist (GH #6) [kentnl]
      - doc patches (GH #3) [David Schmidt]
    committed Feb 27, 2015
  2. changelog++

    committed Feb 27, 2015
  3. skip NYTProf files in dist

    committed Feb 27, 2015
  4. Merge branch 'master' of git://github.com/kentfredric/data-dpath into…

    … kentfredric-master
    
    Conflicts:
    	dist.ini
    committed Feb 27, 2015
Commits on Feb 4, 2015
  1. Merge pull request #3 from davewood/master

    doc patches
    committed Feb 4, 2015
  2. v0.51

      - docs: how to do XPath-like attributes access (RT#94211) [Slaven Rezić]
      - avoid warning with perl 5.21+ via newer "aliased" [Doug Bell]
      - speed optimizations (5%) [Reini Urban]
    committed Feb 4, 2015
Commits on Jan 15, 2015
  1. changes++

    committed Jan 15, 2015
  2. Speed optimizations

    - in _any() avoid redundant assignment of arrays,
      instead directly push to final array and also
      drop intermediate map{}s
    
    This takes ~5% less time in reference benchmark.
    
    Credits to Reini Urban for the suggestions.
    committed Jan 15, 2015
Commits on Jan 14, 2015
  1. docs: describe XPath-like attribute access

    Contributed by Slaven Rezić in RT#94211.
    committed Jan 14, 2015
  2. git ignore nytprof files

    committed Jan 14, 2015
  3. Merge pull request #5 from preaction/master

    upgrade to aliased 0.33 to fix warning on perl 5.21
    committed Jan 14, 2015
Commits on Dec 27, 2014
  1. upgrade to aliased 0.33 to fix warning on perl 5.21

    perl 5.21.x has a new warning "Constants from lexical variables
    potentially modified elsewhere are deprecated". This is fixed in aliased
    version 0.33.
    preaction committed Dec 27, 2014
Commits on Aug 22, 2014
Commits on Mar 14, 2014
  1. v0.50

            - micro optimization: _any() saves a dereference
    committed Mar 14, 2014
  2. changelog++

    committed Mar 14, 2014
Commits on Dec 5, 2013
  1. fixed Filters link in POD

    David Schmidt committed Dec 5, 2013
Commits on Apr 24, 2013
  1. doc fix: use 'my' consistently

    David Schmidt committed Apr 24, 2013
  2. doc fix

    David Schmidt committed Apr 24, 2013
Commits on Dec 14, 2012
  1. _any() saves a dereference

    So far _any() returned an array by explicitely
    dereferencing the array-ref after recursion.
    
    We now directly return the array-ref and the
    de-ref is done externally when needed.
    
    This simplifies _any() for later refactoring
    and even brings a slight speed-gain of ~1% in our
    reference benchmarks. Probably more when _any()
    would return lots of results in the benchmarks.
    committed Dec 14, 2012