Stories
Slash Boxes
Comments
NOTE: use Perl; is on undef hiatus. You can read content, but you can't post it. More info will be forthcoming forthcomingly.

All the Perl that's Practical to Extract and Report

The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
 Full
 Abbreviated
 Hidden
More | Login | Reply
Loading... please wait.
  • This is great news, and very welcome indeed.

    But I also have to add that part of the reduced runtime is due to many changes in the tests that leads to parse failures and thus early exists.

    The biggest chunk is the changed POD. We're slowly converting the old style POD from

      =pod

      ...

      =cut

    to new style POD

      =begin description

      ...

      =end description

    Since pugs can't parse that, the number of failing tests rose significantly.
    • Hmm? I don't see why that would be the case; Pugs parsed the begin/end POD comments just fine for quite a while now.

      For example, t/spec/S02-builtin_data_types/subtypes.t has such a block, and pugs passes it just fine (with two unexpected successes).

      Is it possible for you to give a single test file for me to try against?

      • Sorry for the noise, I tried with an outdated version of pugs. Very Outdated.

        But there's something different that came to my mind - are you actually using fudge before running the tests? some of the :todo<feature> markers were removed because they are are implementation specific, and substituted by fudge markers.

        In the parrot repository languages/perl6/t/harness integrates fudge, and languages/perl6/tools/autounfudge.pl is a script for autotmatically removing fudge markers from passing tests (which br
        • Sure, that'd be excellent. I did notice the unTODOing, but that did not trouble me as much because as long as the passing tests are still passing, things are going fine.

          In retrospect, I regret the insistence of marking :todo<feature> and making sure that all platforms fails with an equal number with tests before each Pugs release. Much of release engineering effort were spent tracking down platform-specific bugs, while it would make more sense simply to let them fail.

          But sure, a fudge script livi

          • The trouble with unTUDOed tests is just that they terminate pugs sometimes, and thus hide other tests.

            The fudge and fudgeall scripts already live in t/spec/, so you just have to integrate them into the pugs test harness.
            It's the unfudge scripts (that basically writes patches that remove '#?rakudo' or '#?pugs' fudge lines for you) that I considered for moving.

            I think that we also need to specify the functionality of Test.pm a bit more since rakudo uses its own (simpler) Test.pm, with slightly different is() semantics (it uses string comparison for now). I'll raise that question on p6c after my vacations.