1 =head1 These are things that need doin'
3 Please let me know if you start hacking on one of these or if you have any amazing ideas.
5 * Be able to run any test file, not just PIR tests
7 Run /bin/sh -c or ./file, and then default to parrot if neither work ?
9 Turns out /bin/sh -c foo.t means foo.t must be executable.
11 We also have the option of using proc_exec from the global parrot config
13 * Use Getopt to have some nice command line options, like
18 * Parse todo/skips/bailout correctly
20 http://cpansearch.perl.org/src/ANDYA/Test-Harness-3.17/lib/TAP/Parser/Grammar.pm
22 * Exit codes are not being dealt with correctly. This will require some internal redesign.
23 stream.'is_pass' should take exit codes into account
25 * More detailed statistics in test summary
27 exit codes if nonzero and test runtime
29 * Aggregation step in t/harness.pir needs to be abstracted out into Tapir::Harness and tested properly
31 http://cpansearch.perl.org/src/ANDYA/Test-Harness-3.17/lib/TAP/Parser/Aggregator.pm
33 * Benchmark tapir against Test::Harness and Test::Run, with euler_bench
34 Test::Run http://web-cpan.berlios.de/modules/Test-Run/
35 euler_bench http://github.com/notbenh/euler_bench
37 * Make the output format prettier
38 line up "passed" in the same column/etc
40 * Running tests in parallel
41 How to do it? There should be examples in the Parrot test suite
43 * Tests and implementation for parsing invalid plans
45 * Tests and implementation for out of order test notification
47 * Submitting smoke reports to smolder
48 TAP::Harness::Archive for Parrot
50 * How to deal with differences been TAP versions?