Closed GoogleCodeExporter closed 9 years ago
That's odd. I applied the patch to my "staging" clone but I'm now seeing a test
failure (on Ubuntu natty).
In an effort to assist diagnosis, I have tweaked the test assertion so that it
outputs the value of rdfXml. The CI build failure log has the details -
Python2.6:
http://bel-epa.com/hudson/job/rdflib-test/TOXENV=py26/lastCompletedBuild/testRep
ort/test.test_serializexml/TestXMLSerializer/test_result_fragments_with_base/
Python2.7:
http://bel-epa.com/hudson/job/rdflib-test/TOXENV=py27/lastCompletedBuild/testRep
ort/test.test_serializexml/TestXMLSerializer/test_result_fragments_with_base/
and I've included the relevant fragment of the error log in an attached file.
Original comment by gjhigg...@gmail.com
on 24 Sep 2011 at 7:36
Attachments:
Thanks for the feedback detail.
Hmmm... I wonder if my patch generation went wrong (I'm a bit new to this
aspect of Mercurial). It appears that the XMLSerializer is NOT adding the
xml:base statement. ( Is it possible to browse code in the staging repository?)
Looking at the patch source, it looks as if there may be 2 lines missing in the
update to serialize for XMLSerializer. I think I should try and regenerate the
patch against the current repository tip.
Original comment by gk-goo...@ninebynine.org
on 24 Sep 2011 at 11:35
OK, I'm now fairly sure it was my patch: as far as I can tell, there were two
commits making up my changes, but my patch file only covered one of them.
Apologies for the inconvenience caused.
I've generated a new patch file (attached) covering both the commits. I've
tested the patch file by:
(1) creating a new clone of the main code repo (hg clone ...)
(2) running the original testsuite 297 tests all pass (python run_tests.py)
(3) applying the new patch to my new working copy (hg import ...)
(4) re-running the test suite. 303 tests all pass: (python run_tests.py)
Ran 303 tests in 103.897s
Is there more I can do to validate the patch?
Original comment by gk-goo...@ninebynine.org
on 24 Sep 2011 at 12:18
Attachments:
++ Is it possible to browse code in the staging repository?
Ach, that would have been useful to you, sorry for the omission.
++ I've tested the patch file by ...
++ [ ... ]
++ Is there more I can do to validate the patch?
Your approach is impeccable. I was able to replicate it
successfully and confirm the test success. Job done,
revision 4abe48469431 committed.
http://code.google.com/p/rdflib/source/detail?r=4abe48469431fb6a3598dfb447477675
e52b553e
(http://bel-epa.com/hudson/job/rdflib/37/)
Thank you for the contribution.
I've been caught out a number of times with Mercurial (and
friends) --- I now habitually arrange matters so that stuff
is pushed through the Hudson CI as a matter of course so
that I get an independent confirmation from a different
machine. It's overkill to a degree but I'm still feeling
around for an approach that compensates for my rather loose
working style but which also keeps the busywork to a minimum.
Original comment by gjhigg...@gmail.com
on 24 Sep 2011 at 9:35
Original issue reported on code.google.com by
gk-goo...@ninebynine.org
on 23 Sep 2011 at 2:54Attachments: