byteify-strings: do not rewrite system string literals to u'' It would make things worse on Python 2 because unicode processing is generally slower than byte string. We should just leave system strings unmodified.

byteify-strings: do not rewrite iteritems() and itervalues() by default We can't do that automatically due to performance concerns.

byteify-strings: drop import-line hack This is ugly, and valid only for Python 3. We'll need to find a different way if we want to get rid of the code transformer at all.

byteify-strings: add --inplace option to write back result

byteify-strings: add basic command interface

byteify-strings: fork py3 code transformer to make it a standalone command I'm thinking of making a one-off s/''/b''/g change for overall codebase to make linter happy. We could do that without maintaining the script, but I think it will be somewhat useful for extension authors. So it is in contrib.

highlight: get around tmpl.load() which now returns a parsed tree Broken at e637dc0b3b1f, "templater: parse template string to tree by templater class."

changegroup: use progress helper Differential Revision: https://phab.mercurial-scm.org/D3783

archival: use progress helper Differential Revision: https://phab.mercurial-scm.org/D3782

copystore: use progress helper Differential Revision: https://phab.mercurial-scm.org/D3781