Pierre-Yves David <pierre-yves.david@octobus.net> [Mon, 04 Mar 2019 15:46:54 +0100] rev 41857
revlog: preserve `_lazydelta` attribute in `revlog.clone`
The attribute was introduce in
688fc33e105d, Yuya Nishihara pointed out that
this preservation was missing. This changeset fixes the preservation and make
sure we set the attribute according the modes.
Pierre-Yves David <pierre-yves.david@octobus.net> [Mon, 04 Mar 2019 16:08:44 +0100] rev 41856
localrepo: explicit `_lock` arguments in `lock`
The argument for `_lock` are non-trivial, having them passed explicitly makes
thing clearer in my opinion. This is a Gratuitous change, I expect it will save
me (and others) time in the future.
Gregory Szorc <gregory.szorc@gmail.com> [Sun, 03 Mar 2019 17:22:03 -0800] rev 41855
inno: remove w9xpopen.exe
w9xpopen.exe is a utility program shipped with Python <3.4
(https://bugs.python.org/
issue14470 tracked its removal).
The program was used by subprocess to wrap invoked processes
on Windows 95 and 98 or when command.com was used in order to
work around a redirect bug.
The workaround is only used on ancient Windows versions -
versions that we shouldn't see in 2019.
While Python 2.7's subprocess module still references
w9xpopen.exe, not shipping it shouldn't matter unless we're
running an ancient version of Windows. Python will raise
an exception if w9xpopen.exe can't be found.
It's highly unlikely anyone is using current Mercurial releases
on these ancient Windows versions. So remove w9xpopen.exe
from the Inno installer.
.. bc::
The 32-bit Windows Inno installers no longer distribute
w9xpopen.exe. This should only impact people running
Mercurial on Windows 95, 98, or ME.
Differential Revision: https://phab.mercurial-scm.org/D6068
Gregory Szorc <gregory.szorc@gmail.com> [Sun, 03 Mar 2019 15:53:27 -0800] rev 41854
inno: stop shipping pywin32
Ancient versions of Mercurial relied on pywin32 and I suspect
that's why we have this dependency.
We also ship the "keyring" package, which has a dependency
on "pywin32-ctypes" (providing the "win32ctypes" package).
This is a stripped down version of pywin32 that doesn't have
as many dependencies.
Since we don't have a dependency on pywin32 and since pywin32
is a bit annoying to package, let's get rid of it.
With this change, py2exe no longers picks up DLL dependencies
on various UCRT DLLs (because we no longer have a .pyd file
beloning to pywin32 which was pulling them in). So, we were
able to remove code in support of the UCRT DLLs.
.. bc::
The Windows Inno installers no longer ship the pywin32 package.
This package was being bundled for historical reasons. Mercurial
stopped using pywin32 several years ago and the disappearance
of this package should not have any meaningful impact.
Differential Revision: https://phab.mercurial-scm.org/D6067
Gregory Szorc <gregory.szorc@gmail.com> [Sun, 03 Mar 2019 18:19:07 -0800] rev 41853
inno: script to automate building Inno installer
The official Inno installer build process is poorly documented.
And attempting to reproduce behavior of the installer uploaded
to www.mercurial-scm.org has revealed a number of unexpected
behaviors.
This commit attempts to improve the state of reproducibility
of the Inno installer by introducing a Python script to
largely automate the building of the installer.
The new script (which must be run from an environment with the
Visual C++ environment configured) takes care of producing an
Inno installer. When run from a fresh Mercurial source checkout
with all the proper system dependencies (the VC++ toolchain,
Windows 10 SDK, and Inno tools) installed, it "just works."
The script takes care of downloading all the Python
dependencies in a secure manner and manages the build
environment for you. You don't need any additional config
files: just launch the script, pointing it at an existing
Python and ISCC binary and it takes care of the rest.
The produced installer creates a Mercurial installation with
a handful of differences from the existing 4.9 installers
(produced by someone else):
* add_path.exe is missing (this was removed a few changesets ago)
* The set of api-ms-win-core-* DLLs is different (I suspect this
is due to me using a different UCRT / Windows version).
* kernelbase.dll and msasn1.dll are missing.
* There are a different set of .pyc files for dulwich,
keyring, and pygments due to us using the latest versions of
each.
* We include Tcl/Tk DLLs and .pyc files (I'm not sure why these
are missing from the existing installers).
* We include the urllib3 and win32ctypes packages (which are
dependencies of dulwich and pywin32, respectively). I'm not
sure why these aren't present in the existing installers.
* We include a different set of files for the distutils package.
I'm not sure why. But it should be harmless.
* We include the docutils package (it is getting picked up as
a dependency somehow). I think this is fine.
* We include a copy of argparse.pyc. I'm not sure why this was
missing from existing installers.
* We don't have a copy of sqlite3/dump.pyc. I'm not sure why. The
SQLite C extension code only imports this module when
conn.iterdump() is called. It should be safe to omit.
* We include files in the email.test and test packages. The set of
files is small and their presence should be harmless.
The new script and support code is written in Python 3 because
it is brand new and independent code and I don't believe new
Python projects should be using Python 2 in 2019 if they have
a choice about it.
The readme.txt file has been renamed to readme.rst and overhauled
to reflect the existence of build.py.
Differential Revision: https://phab.mercurial-scm.org/D6066
Gregory Szorc <gregory.szorc@gmail.com> [Sun, 03 Mar 2019 14:08:25 -0800] rev 41852
setup: exclude some internal UCRT files
When attempting to build the Inno installer locally, I was getting
several file not found errors when py2exe was crawling DLL
dependencies. The missing DLLs appear to be "internal" DLLs
used by the Universal C Runtime (UCRT). In many cases, the
missing DLLs don't appear to exist on my system at all!
Some of the DLLs have version numbers that appear to be N+1
of what the existing version number is. Maybe the "public" UCRT
DLLs are probing for version N+1 at load time and py2exe is
picking these up? Who knows.
This commit adds the non-public UCRT DLLs as found by
py2exe on my system to the excluded DLLs set. After this
change, I'm able to produce an Inno installer with an
appropriate set of DLLs.
Differential Revision: https://phab.mercurial-scm.org/D6065
Gregory Szorc <gregory.szorc@gmail.com> [Sun, 03 Mar 2019 15:46:26 -0800] rev 41851
setup: include additional packages in py2exe distribution
I'm attempting to reproduce the Inno installers on my local
machine. As part of auditing differences between installer output,
I noticed that the existing Inno installers include various 3rd
party packages.
There is no mention of this in the build instructions nor on
the wiki. This must be something that is done by the installer
producer.
This commit teaches setup.py to include these 3rd party packages
in py2exe's library. After this change, I am able to produce
Inno installers that have a nearly identical set of Python
modules.
It's worth noting that pywin32 is included even though it
probably shouldn't be. But including it is necessary in order
to achieve parity with existing Inno installers.
Differential Revision: https://phab.mercurial-scm.org/D6064
Gregory Szorc <gregory.szorc@gmail.com> [Sun, 03 Mar 2019 10:31:23 -0800] rev 41850
setup: define build_doc command
Currently, various processes for packaging Mercurial state to
manually invoke `make -C doc` in order to generate the documentation.
This Makefile merely invokes `gendoc.py` and `runrst` to produce
man pages and HTML pages.
Not all environments may have the ability to easily run
Makefiles. Windows is notably in this set.
This commit ports the man page and HTML generation logic from
doc/Makefile to setup.py. We introduce a new build_doc command
which generates documentation by calling gendoc.py and runrst.
The documentation can now be built via pure Python by running
`python setup.py build_doc`.
We don't implement dependency tracking because IMO it is more
effort than it is worth.
We could potentially remove the duplicated functionality in
doc/Makefile. But I'm not sure what all is depending on it. So
I plan to keep it around.
# no-check-commit because forced foo_bar function names
Differential Revision: https://phab.mercurial-scm.org/D6063