Augie Fackler <augie@google.com> [Sat, 14 Apr 2018 00:08:59 -0400] rev 37752
hook: also use pprint on lists for stable output on py2/3
Differential Revision: https://phab.mercurial-scm.org/D3361
Augie Fackler <augie@google.com> [Fri, 13 Apr 2018 23:56:11 -0400] rev 37751
hook: use stringutil.pprint instead of reinventing it
Differential Revision: https://phab.mercurial-scm.org/D3360
Augie Fackler <augie@google.com> [Fri, 13 Apr 2018 23:57:57 -0400] rev 37750
stringutil: make b prefixes on string output optional
I need this to preserve some behavior in hook.py.
Differential Revision: https://phab.mercurial-scm.org/D3359
Augie Fackler <augie@google.com> [Fri, 13 Apr 2018 23:55:37 -0400] rev 37749
stringutil: teach pprint how to format None
Differential Revision: https://phab.mercurial-scm.org/D3358
Augie Fackler <augie@google.com> [Fri, 13 Apr 2018 23:50:17 -0400] rev 37748
tests: update inline extensions in test-bundle2-exchange.t to py3
The test doesn't pass for what superficially look like good
reasons. We'll need to come back to it later.
# skip-blame because it's b prefixes and a couple of b'N' instead of str(N)
Differential Revision: https://phab.mercurial-scm.org/D3357
Augie Fackler <augie@google.com> [Fri, 13 Apr 2018 23:11:46 -0400] rev 37747
wsgicgi: un-do some prior porting work that is now wrong
The Python 3 WSGI behavior is that the environ dict should be full of
unicodes. We previously tried Too Hard here, so we unwind that bit of porting.
Also add some bytesurl() encodes on status and headers.
test-clone-cgi.t now passes.
Differential Revision: https://phab.mercurial-scm.org/D3356
Augie Fackler <augie@google.com> [Fri, 13 Apr 2018 22:36:54 -0400] rev 37746
hgweb_mod: inform hgweb class about paths actually being bytes
Differential Revision: https://phab.mercurial-scm.org/D3355
Augie Fackler <augie@google.com> [Fri, 13 Apr 2018 22:33:38 -0400] rev 37745
hgweb: inform hgweb.hgweb() entrypoint that paths should be bytes
Differential Revision: https://phab.mercurial-scm.org/D3354
Augie Fackler <augie@google.com> [Fri, 13 Apr 2018 22:27:16 -0400] rev 37744
tests: port inline cgi script in test-largefiles.t to python 3
Differential Revision: https://phab.mercurial-scm.org/D3353
Augie Fackler <augie@google.com> [Fri, 13 Apr 2018 22:26:57 -0400] rev 37743
lfcommands: use %d on known-int in format string
Differential Revision: https://phab.mercurial-scm.org/D3352
Augie Fackler <augie@google.com> [Fri, 13 Apr 2018 22:12:30 -0400] rev 37742
py3: fix test-shelve.t on Python 3
We get a better error message out of the abort here, which is fine.
Differential Revision: https://phab.mercurial-scm.org/D3351
Augie Fackler <augie@google.com> [Fri, 13 Apr 2018 21:53:07 -0400] rev 37741
tests: fix up a couple of minor bytes inconsistencies in run-tests.py
Only in the automatic bisection code, so fortunately nothing major is
amiss. Fixes test-run-tests.t under Python 3.
Differential Revision: https://phab.mercurial-scm.org/D3350
Augie Fackler <augie@google.com> [Fri, 13 Apr 2018 21:51:10 -0400] rev 37740
tests: make sure test-run-tests.t actually runs run-tests.py under Python 3
I'm fairly certain it hasn't been until now. Mercifully, there doesn't
appear to be any ninja breakage.
Differential Revision: https://phab.mercurial-scm.org/D3349
Augie Fackler <augie@google.com> [Fri, 13 Apr 2018 21:30:55 -0400] rev 37739
py3: another three passing
Differential Revision: https://phab.mercurial-scm.org/D3348
Augie Fackler <augie@google.com> [Fri, 13 Apr 2018 21:11:28 -0400] rev 37738
httppeer: work around API differences on urllib Request objects
Since this is only a problem in httppeer, I'd rather keep this a
local-to-the-module kludge rather than pile more on pycompat. We'll
still find it easily to clean up later because it checks
pycompat.ispy3.
Differential Revision: https://phab.mercurial-scm.org/D3347
Augie Fackler <augie@google.com> [Fri, 13 Apr 2018 21:07:18 -0400] rev 37737
httppeer: no matter what Python 3 might think, http headers are bytes
Differential Revision: https://phab.mercurial-scm.org/D3346