util: work around Python 3 returning None at EOF instead of '' Differential Revision: https://phab.mercurial-scm.org/D2561

util: add missing r prefix on some __setattr__ calls # skip-blame just a pair of r prefixes Differential Revision: https://phab.mercurial-scm.org/D2560

tests: add some re and globs for test-revset on python3 At this point we're down to two deprecation warnings (which I suspect are showing a bug in the test?) and one weird-looking failure. Progress! Differential Revision: https://phab.mercurial-scm.org/D2559

tests: add missing b prefixes and fix a %s to %d in test-revset.t # skip-blame just b prefixes and a %d instead of %s Differential Revision: https://phab.mercurial-scm.org/D2558

revset: use %d to turn an int into a bytestr Differential Revision: https://phab.mercurial-scm.org/D2557

revset: use {force,}bytestr to fix some %r formatting issues Differential Revision: https://phab.mercurial-scm.org/D2556

py3: sixteen more passing tests Differential Revision: https://phab.mercurial-scm.org/D2552

debugcommands: fix repr in debugignore print with pycompat.bytestr Differential Revision: https://phab.mercurial-scm.org/D2549

verify: fix exception formatting bug in Python 3 Differential Revision: https://phab.mercurial-scm.org/D2541

get-with-headers: use bytes stdout thoroughly On Python 3, sys.stdout.buffer is backed by a separate buffer from sys.stdout. We should choose one.

exchange: remove dead assignment or forcebundle1 Differential Revision: https://phab.mercurial-scm.org/D2548

templatekw: fix dict construction in _showlist to not mix bytes and strs What we had was fine on Python 2, but was slightly wrong on Python 3. This works on both. Differential Revision: https://phab.mercurial-scm.org/D2554

templatefilters: avoid infinite recursion bug in stringify This doesn't ever happen on Python 2, but it's been a persistent pain on Python 3. Adding this helped produce some of my upcoming Python 3 fixes. Differential Revision: https://phab.mercurial-scm.org/D2553

match: some minimal pycompat fixes guided by test-hgignore.t Differential Revision: https://phab.mercurial-scm.org/D2551

purge: apply byteskwargs to opts, fixing all python3 issues here Differential Revision: https://phab.mercurial-scm.org/D2547

tests: port test-bookmarks.t extension to Python 3 # skip-blame it's just b prefixes Differential Revision: https://phab.mercurial-scm.org/D2546

scmutil: fix a repr in an error message on Python 3 Differential Revision: https://phab.mercurial-scm.org/D2545

bookmarks: fix a repr in a message on Python 3 Differential Revision: https://phab.mercurial-scm.org/D2544

py3: add missing b prefixes in test-debugextensions.t # skip-blame just some b prefixes Differential Revision: https://phab.mercurial-scm.org/D2543

tests: help dummysmtpd work on python 3 Differential Revision: https://phab.mercurial-scm.org/D2542

tests: port helper script revlog-formatv0.py to python 3 Differential Revision: https://phab.mercurial-scm.org/D2540

tests: add missing b prefix in test python in test-issue2137.t # skip-blame just a b prefix Differential Revision: https://phab.mercurial-scm.org/D2539

templatefilters: convert arguments to sysstrs for unicode() ctor Differential Revision: https://phab.mercurial-scm.org/D2538

py3: use util.forcebytestr() to convert IOErrors to bytes Differential Revision: https://phab.mercurial-scm.org/D2536

py3: whitelist three more passing tests Differential Revision: https://phab.mercurial-scm.org/D2530

convert: fix two %r output formats with pycompat.bytestr() wrapping Differential Revision: https://phab.mercurial-scm.org/D2529

convert: use our shlex wrapper in filemap to avoid Python 3 tracebacks Differential Revision: https://phab.mercurial-scm.org/D2527

convcmd: use our shlex wrapper to avoid Python 3 tracebacks Differential Revision: https://phab.mercurial-scm.org/D2526

convert: add some utility code for working with shlex on Python 3 This could have gone in pycompat, but it's only needed in convert, so I figured it made more sense here. It's got py3 in the name and checks pycompat.ispy3, so we'll find it whenever we decide to drop Python 2 support in 20x6. # no-check-commit because of required foo_bar naming on the proxy class Differential Revision: https://phab.mercurial-scm.org/D2525

pycompat: add support for encoding argument to our wrapper This only works on Python 3, but I'm about to need it for a regrettable hack in the convert code. Differential Revision: https://phab.mercurial-scm.org/D2524

py3: port tests/test-wireproto.py to Python 3 Differential Revision: https://phab.mercurial-scm.org/D2534

py3: use pycompat.bytestr() to convert None to bytes Differential Revision: https://phab.mercurial-scm.org/D2533

webcommands: use explicit integer division for Python 3 compat Differential Revision: https://phab.mercurial-scm.org/D2532

templatefilters: stop using str as a variable name str() is an inbuilt function. Differential Revision: https://phab.mercurial-scm.org/D2531

py3: fix string slicing in util.parsetimezone()

py3: replace type 'str' by 'bytes' in templater.py

py3: don't crash when re-raising encoding error

py3: mark all string literals in test-command-template.t as bytes # skip-blame because just b'' prefixes

py3: byte-stringify ValueError of unescapestr() to reraise as ParseError

py3: fix type of string literals in templater.tokenize() # skip-blame because just b'' prefixes

py3: drop b'' from error message generated by templater.runmember()

py3: fix join(), min(), and max() template functions over string It's silly to split a string into characters and concatenate them, but that should work and test-command-template.t has one. min() and max() had the same issue on Python 3, so fixed too.

py3: use startswith() to check existence of trailing '\n' in .hgtags file

py3: silence return value of file.write() in test-command-template.t

error: fix isinstnace check to use bytes instead of str Differential Revision: https://phab.mercurial-scm.org/D2528

py3: whitelist another nine passing tests I now see 347 tests passing[0] and 336 failing, so we're past halfway there. Hooray! 0: A few tests appear to regress in small ways (doctest output changes, for example) on Python 3.7.0a1, which is what I'm now using to test. That said, I'm pleased to report no major regressions. Differential Revision: https://phab.mercurial-scm.org/D2439

util: use pycompat.bytestr() on repr() in date parse abort Avoids the b'' prefix. Differential Revision: https://phab.mercurial-scm.org/D2438

py3: whitelist 14 new tests passing 12 new tests passed because of parent changeset and 2 were passing before that too. Differential Revision: https://phab.mercurial-scm.org/D2523