[Boost-testing] Spirit unit and regression test reorganization

Subject: [Boost-testing] Spirit unit and regression test reorganization
From: Bryce Lelbach (admin_at_[hidden])
Date: 2010-12-30 09:02:03

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hey folks. I reorganized the Spirit test suites the other day. Previously, the
qi targets weren't prefix with qi_, though the karma targets and lex targets
were. This resulted in ambigious names for half of the tests.

Currently, while the old Bjam targets are being rotated out of the test matrix,
the Spirit regression page looks terrible because both the old target
identitifers and the new target identifiers are there, as some machines are on
the new suite, and some are still on the old suite.

P.S. I've also started breaking up bigger Spirit test files into smaller
translation units, so that they don't timeout on (mostly Intel) slow compilers.
I apologize in advance to test runners if this puts a strain on your resources,
but having all the Intel tests fail because of the timeout makes it hard for me
to get a real picture of the state of Spirit support on Intel's compiler. 300
or even 600 seconds isn't a feasible timeout for anything more than trivial
Spirit code on Intel.

- -- 
Bryce Lelbach aka wash
boost-spirit.com
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (GNU/Linux)

iEYEARECAAYFAk0ckNsACgkQO/fqqIuE2t79UgCgscstSgV1y4nkS9Y/OBq5pNdx
yXcAnj1smja5Thl9yp4Visko7cXKKV2f
=v067
-----END PGP SIGNATURE-----