Mercurial > hg
view contrib/win32/win32-build.txt @ 30442:41a8106789ca
util: implement zstd compression engine
Now that zstd is vendored and being built (in some configurations), we
can implement a compression engine for zstd!
The zstd engine is a little different from existing engines. Because
it may not always be present, we have to defer load the module in case
importing it fails. We facilitate this via a cached property that holds
a reference to the module or None. The "available" method is
implemented to reflect reality.
The zstd engine declares its ability to handle bundles using the
"zstd" human name and the "ZS" internal name. The latter was chosen
because internal names are 2 characters (by only convention I think)
and "ZS" seems reasonable.
The engine, like others, supports specifying the compression level.
However, there are no consumers of this API that yet pass in that
argument. I have plans to change that, so stay tuned.
Since all we need to do to support bundle generation with a new
compression engine is implement and register the compression engine,
bundle generation with zstd "just works!" Tests demonstrating this
have been added.
How does performance of zstd for bundle generation compare? On the
mozilla-unified repo, `hg bundle --all -t <engine>-v2` yields the
following on my i7-6700K on Linux:
engine CPU time bundle size vs orig size throughput
none 97.0s 4,054,405,584 100.0% 41.8 MB/s
bzip2 (l=9) 393.6s 975,343,098 24.0% 10.3 MB/s
gzip (l=6) 184.0s 1,140,533,074 28.1% 22.0 MB/s
zstd (l=1) 108.2s 1,119,434,718 27.6% 37.5 MB/s
zstd (l=2) 111.3s 1,078,328,002 26.6% 36.4 MB/s
zstd (l=3) 113.7s 1,011,823,727 25.0% 35.7 MB/s
zstd (l=4) 116.0s 1,008,965,888 24.9% 35.0 MB/s
zstd (l=5) 121.0s 977,203,148 24.1% 33.5 MB/s
zstd (l=6) 131.7s 927,360,198 22.9% 30.8 MB/s
zstd (l=7) 139.0s 912,808,505 22.5% 29.2 MB/s
zstd (l=12) 198.1s 854,527,714 21.1% 20.5 MB/s
zstd (l=18) 681.6s 789,750,690 19.5% 5.9 MB/s
On compression, zstd for bundle generation delivers:
* better compression than gzip with significantly less CPU utilization
* better than bzip2 compression ratios while still being significantly
faster than gzip
* ability to aggressively tune compression level to achieve
significantly smaller bundles
That last point is important. With clone bundles, a server can
pre-generate a bundle file, upload it to a static file server, and
redirect clients to transparently download it during clone. The server
could choose to produce a zstd bundle with the highest compression
settings possible. This would take a very long time - a magnitude
longer than a typical zstd bundle generation - but the result would
be hundreds of megabytes smaller! For the clone volume we do at
Mozilla, this could translate to petabytes of bandwidth savings
per year and faster clones (due to smaller transfer size).
I don't have detailed numbers to report on decompression. However,
zstd decompression is fast: >1 GB/s output throughput on this machine,
even through the Python bindings. And it can do that regardless of the
compression level of the input. By the time you have enough data to
worry about overhead of decompression, you have plenty of other things
to worry about performance wise.
zstd is wins all around. I can't wait to implement support for it
on the wire protocol and in revlogs.
author | Gregory Szorc <gregory.szorc@gmail.com> |
---|---|
date | Fri, 11 Nov 2016 01:10:07 -0800 |
parents | e999ed2192ef |
children |
line wrap: on
line source
The standalone Windows installer for Mercurial is built in a somewhat jury-rigged fashion. It has the following prerequisites. Ensure to take the packages matching the mercurial version you want to build (32-bit or 64-bit). Python 2.6 for Windows http://www.python.org/download/releases/ A compiler: either MinGW http://www.mingw.org/ or Microsoft Visual C++ 2008 SP1 Express Edition http://www.microsoft.com/express/Downloads/Download-2008.aspx Python for Windows Extensions http://sourceforge.net/projects/pywin32/ mfc71.dll (just download, don't install; not needed for Python 2.6) http://starship.python.net/crew/mhammond/win32/ Visual C++ 2008 redistributable package (needed for >= Python 2.6 or if you compile with MSVC) for 32-bit: http://www.microsoft.com/downloads/details.aspx?FamilyID=9b2da534-3e03-4391-8a4d-074b9f2bc1bf for 64-bit: http://www.microsoft.com/downloads/details.aspx?familyid=bd2a6171-e2d6-4230-b809-9a8d7548c1b6 The py2exe distutils extension http://sourceforge.net/projects/py2exe/ GnuWin32 gettext utility (if you want to build translations) http://gnuwin32.sourceforge.net/packages/gettext.htm Inno Setup http://www.jrsoftware.org/isdl.php#qsp Get and install ispack-5.3.10.exe or later (includes Inno Setup Processor), which is necessary to package Mercurial. ISTool - optional http://www.istool.org/default.aspx/ add_path (you need only add_path.exe in the zip file) http://www.barisione.org/apps.html#add_path Docutils http://docutils.sourceforge.net/ CA Certs file http://curl.haxx.se/ca/cacert.pem And, of course, Mercurial itself. Once you have all this installed and built, clone a copy of the Mercurial repository you want to package, and name the repo C:\hg\hg-release. In a shell, build a standalone copy of the hg.exe program. Building instructions for MinGW: python setup.py build -c mingw32 python setup.py py2exe -b 2 Note: the previously suggested combined command of "python setup.py build -c mingw32 py2exe -b 2" doesn't work correctly anymore as it doesn't include the extensions in the mercurial subdirectory. If you want to create a file named setup.cfg with the contents: [build] compiler=mingw32 you can skip the first build step. Building instructions with MSVC 2008 Express Edition: for 32-bit: "C:\Program Files\Microsoft Visual Studio 9.0\VC\vcvarsall.bat" x86 python setup.py py2exe -b 2 for 64-bit: "C:\Program Files\Microsoft Visual Studio 9.0\VC\vcvarsall.bat" x86_amd64 python setup.py py2exe -b 3 Copy add_path.exe and cacert.pem files into the dist directory that just got created. If you are using Python 2.6 or later, or if you are using MSVC 2008 to compile mercurial, you must include the C runtime libraries in the installer. To do so, install the Visual C++ 2008 redistributable package. Then in your windows\winsxs folder, locate the folder containing the dlls version 9.0.21022.8. For x86, it should be named like x86_Microsoft.VC90.CRT_(...)_9.0.21022.8(...). For x64, it should be named like amd64_Microsoft.VC90.CRT_(...)_9.0.21022.8(...). Copy the files named msvcm90.dll, msvcp90.dll and msvcr90.dll into the dist directory. Then in the windows\winsxs\manifests folder, locate the corresponding manifest file (x86_Microsoft.VC90.CRT_(...)_9.0.21022.8(...).manifest for x86, amd64_Microsoft.VC90.CRT_(...)_9.0.21022.8(...).manifest for x64), copy it in the dist directory and rename it to Microsoft.VC90.CRT.manifest. Before building the installer, you have to build Mercurial HTML documentation (or fix mercurial.iss to not reference the doc directory): cd doc mingw32-make html cd .. If you use ISTool, you open the C:\hg\hg-release\contrib\win32\mercurial.iss file and type Ctrl-F9 to compile the installer file. Otherwise you run the Inno Setup compiler. Assuming it's in the path you should execute: iscc contrib\win32\mercurial.iss /dVERSION=foo Where 'foo' is the version number you would like to see in the 'Add/Remove Applications' tool. The installer will be placed into a directory named Output/ at the root of your repository. If the /dVERSION=foo parameter is not given in the command line, the installer will retrieve the version information from the __version__.py file. If you want to build an installer for a 64-bit mercurial, add /dARCH=x64 to your command line: iscc contrib\win32\mercurial.iss /dARCH=x64 To automate the steps above you may want to create a batchfile based on the following (MinGW build chain): echo [build] > setup.cfg echo compiler=mingw32 >> setup.cfg python setup.py py2exe -b 2 cd doc mingw32-make html cd .. iscc contrib\win32\mercurial.iss /dVERSION=snapshot and run it from the root of the hg repository (c:\hg\hg-release).