Stabilize dev-vcs/hgsubversion-1.2.1.
does not pass test ( i believe is known ) but the rest is ok.
amd64; Curious, flag test set for emerge, then emerged in no time flat. ran ebuild $PORTDIR/dev-vcs/hgsubversion/hgsubversion-1.2.1.ebuild test to force the running of test phase. , svnadmin loa~test/testrepo ran for many minutes, output seemingly to /dev/null; the test flag seemingly not invoking the test phase. * Testing of dev-vcs/hgsubversion-1.2.1 with CPython 2.7... test_externals: skipping .hgsub tests ............EE..................................................................................FF..., then did fail., as stated by above description. /var/tmp/portage/dev-vcs/hgsubversion-1.2.1/temp/environment: line 3423: 504 Segmentation fault PYTHONPATH="../build-${PYTHON_ABI}/lib" "$(PYTHON)" run.py * ERROR: dev-vcs/hgsubversion-1.2.1 failed (test phase): * Testing failed with CPython 2.7 in testing() function This package's PKGINFO notes doesn't match its content. According to the notes, invoke with hg version --svn to test if it's correctly installed/emerged.It yields: hg version: option --svn not recognized hg version output version and copyright information use "hg -v help version" to show global options In addition, according to equery b hg, the hg exectuable belongs to dev-vcs/mercurial-1.7 The optarg --svn, supposedly representing this package content, is reported as foreign! Any help info is derived from hg mercurial. This just doesn't add up. The package has no executable, so it is presumably an extension to hg. The notes also cite subvertpy as 'required' which is not a listed package under DEPEND. The curious thing about the notes is the stress on running the test suite. Once again there is a mis-match between what is stated and what occurs. The notes say that large volumes of info is 'spat out' in running it. On the contrary, running it outputs the running dots, as if with a " | tee "dots" > /dev/null", which means a user is blinded from what occurs in the test suite. Just perhaps the running with a manual make in the directory might match and the influence of the ebuild src_test() changes it. So ran it a second time, with the benefit of the addition of subvertpy, =dev-python/subvertpy-0.8.0. Grab a book; cpu flat strap, both cores, then, ............................................ ====================================================================== FAIL: test_rename (test_fetch_renames.TestFetchRenames) ---------------------------------------------------------------------- Traceback (most recent call last): File "/var/tmp/portage/dev-vcs/hgsubversion-1.2.1/work/hgsubversion-1.2.1/tests/test_fetch_renames.py", line 57, in test_rename self._test_rename(False) File "/var/tmp/portage/dev-vcs/hgsubversion-1.2.1/work/hgsubversion-1.2.1/tests/test_fetch_renames.py", line 48, in _test_rename 'copy records differ for %s in %d' % (f, rev)) AssertionError: copy records differ for c1 in 4 ====================================================================== FAIL: test_rename_stupid (test_fetch_renames.TestFetchRenames) ---------------------------------------------------------------------- Traceback (most recent call last): File "/var/tmp/portage/dev-vcs/hgsubversion-1.2.1/work/hgsubversion-1.2.1/tests/test_fetch_renames.py", line 60, in test_rename_stupid self._test_rename(True) File "/var/tmp/portage/dev-vcs/hgsubversion-1.2.1/work/hgsubversion-1.2.1/tests/test_fetch_renames.py", line 48, in _test_rename 'copy records differ for %s in %d' % (f, rev)) AssertionError: copy records differ for c1 in 4 ---------------------------------------------------------------------- Ran 474 tests in 7448.146s FAILED (failures=2) * ERROR: dev-vcs/hgsubversion-1.2.1 failed (test phase): * Testing failed with CPython 2.7 in testing() function Little more to be gained from the complete log. gentoo64 ~ # ls -ld /var/tmp/portage/dev-vcs/hgsubversion-1.2.1/temp/build.log -rw-rw---- 1 portage portage 7705 May 3 20:47 /var/tmp/portage/dev-vcs/hgsubversion-1.2.1/temp/build.log
Created attachment 272035 [details] my emerge --info
x86 stable (tests pass here). IAN, use FEATURES='test' to run the tests, the USE-flag test is not for this.
amd64 done. Thanks Ian and Agostino. All done
