view contrib/Makefile.python @ 38732:be4984261611

merge: mark file gets as not thread safe (issue5933)

In default installs, this has the effect of disabling the thread-based
worker on Windows when manifesting files in the working directory. My
measurements have shown that with revlog-based repositories, Mercurial
spends a lot of CPU time in revlog code resolving file data. This ends
up incurring a lot of context switching across threads and slows down
`hg update` operations when going from an empty working directory to
the tip of the repo.

On mozilla-unified (246,351 files) on an i7-6700K (4+4 CPUs):

before: 487s wall
after:  360s wall (equivalent to worker.enabled=false)
cpus=2: 379s wall

Even with only 2 threads, the thread pool is still slower.

The introduction of the thread-based worker (02b36e860e0b) states that
it resulted in a "~50%" speedup for `hg sparse --enable-profile` and
`hg sparse --disable-profile`. This disagrees with my measurement
above. I theorize a few reasons for this:

1) Removal of files from the working directory is I/O - not CPU - bound
   and should benefit from a thread pool (unless I/O is insanely fast
   and the GIL release is near instantaneous). So tests like `hg sparse
   --enable-profile` may exercise deletion throughput and aren't good
   benchmarks for worker tasks that are CPU heavy.
2) The patch was authored by someone at Facebook. The results were
   likely measured against a repository using remotefilelog. And I
   believe that revision retrieval during working directory updates with
   remotefilelog will often use a remote store, thus being I/O and not
   CPU bound. This probably resulted in an overstated performance gain.

Since there appears to be a need to enable the thread-based worker with
some stores, I've made the flagging of file gets as thread safe
configurable. I've made it experimental because I don't want to formalize
a boolean flag for this option and because this attribute is best
captured against the store implementation. But we don't have a proper
store API for this yet. I'd rather cross this bridge later.

It is possible there are revlog-based repositories that do benefit from
a thread-based worker. I didn't do very comprehensive testing. If there
are, we may want to devise a more proper algorithm for whether to use
the thread-based worker, including possibly config options to limit the
number of threads to use. But until I see evidence that justifies
complexity, simplicity wins.

Differential Revision: https://phab.mercurial-scm.org/D3963

PYTHONVER=2.7.14
PYTHONNAME=python-
PREFIX=$(HOME)/bin/prefix-$(PYTHONNAME)$(PYTHONVER)
SYMLINKDIR=$(HOME)/bin

help:
	@echo
	@echo 'Make a custom installation of a Python version'
	@echo
	@echo 'Common make parameters:'
	@echo '  PYTHONVER=...    [$(PYTHONVER)]'
	@echo '  PREFIX=...       [$(PREFIX)]'
	@echo '  SYMLINKDIR=...   [$(SYMLINKDIR) creating $(PYTHONNAME)$(PYTHONVER)]'
	@echo
	@echo 'Common make targets:'
	@echo '  python    - install Python $$PYTHONVER in $$PREFIX'
	@echo '  symlink   - create a $$SYMLINKDIR/$(PYTHONNAME)$$PYTHONVER symlink'
	@echo
	@echo 'Example: create a temporary Python installation:'
	@echo '  $$ make -f Makefile.python python PYTHONVER=${PYTHONVER} PREFIX=/tmp/p27'
	@echo '  $$ /tmp/p27/bin/python -V'
	@echo '  Python 2.7'
	@echo
	@echo 'Some external libraries are required for building Python: zlib bzip2 openssl.'
	@echo 'Make sure their development packages are installed systemwide.'
# fedora: yum install zlib-devel bzip2-devel openssl-devel
# debian: apt-get install zlib1g-dev libbz2-dev libssl-dev
	@echo
	@echo 'To build a nice collection of interesting Python versions:'
	@echo '  $$ for v in 2.{6{,.1,.2,.9},7{,.8,.10}}; do'
	@echo '    make -f Makefile.python symlink PYTHONVER=$$v || break; done'
	@echo 'To run a Mercurial test on all these Python versions:'
	@echo '  $$ for py in `cd ~/bin && ls $(PYTHONNAME)2.*`; do'
	@echo '    echo $$py; $$py run-tests.py test-http.t; echo; done'
	@echo

export LANGUAGE=C
export LC_ALL=C

python: $(PREFIX)/bin/python docutils
	printf 'import sys, zlib, bz2, docutils, ssl' | $(PREFIX)/bin/python

PYTHON_SRCDIR=Python-$(PYTHONVER)
PYTHON_SRCFILE=$(PYTHON_SRCDIR).tgz

$(PREFIX)/bin/python:
	[ -f $(PYTHON_SRCFILE) ] || wget http://www.python.org/ftp/python/$(PYTHONVER)/$(PYTHON_SRCFILE) || curl -OL http://www.python.org/ftp/python/$(PYTHONVER)/$(PYTHON_SRCFILE) || [ -f $(PYTHON_SRCFILE) ]
	rm -rf $(PYTHON_SRCDIR)
	tar xf $(PYTHON_SRCFILE)
	# Debian/Ubuntu disables SSLv2,3 the hard way, disable it on old Pythons too
	-sed -i 's,self.*SSLv[23]_method(),0;//\0,g' $(PYTHON_SRCDIR)/Modules/_ssl.c
	# Find multiarch system libraries on Ubuntu and disable fortify error when setting argv
	LDFLAGS="-L/usr/lib/`dpkg-architecture -qDEB_HOST_MULTIARCH`"; \
	BASECFLAGS=-U_FORTIFY_SOURCE; \
	export LDFLAGS BASECFLAGS; \
	cd $(PYTHON_SRCDIR) && ./configure --prefix=$(PREFIX) && make all SVNVERSION=pwd && make install
	printf 'import sys, zlib, bz2, ssl' | $(PREFIX)/bin/python
	rm -rf $(PYTHON_SRCDIR)

DOCUTILSVER=0.12
DOCUTILS_SRCDIR=docutils-$(DOCUTILSVER)
DOCUTILS_SRCFILE=$(DOCUTILS_SRCDIR).tar.gz

docutils: $(PREFIX)/bin/python
	@$(PREFIX)/bin/python -c 'import docutils' || ( set -ex; \
	[ -f $(DOCUTILS_SRCFILE) ] || wget http://downloads.sourceforge.net/project/docutils/docutils/$(DOCUTILSVER)/$(DOCUTILS_SRCFILE) || [ -f $(DOCUTILS_SRCFILE) ]; \
	rm -rf $(DOCUTILS_SRCDIR); \
	tar xf $(DOCUTILS_SRCFILE); \
	cd $(DOCUTILS_SRCDIR) && $(PREFIX)/bin/python setup.py install --prefix=$(PREFIX); \
	$(PREFIX)/bin/python -c 'import docutils'; \
	rm -rf $(DOCUTILS_SRCDIR); )

symlink: python $(SYMLINKDIR)
	ln -sf $(PREFIX)/bin/python $(SYMLINKDIR)/$(PYTHONNAME)$(PYTHONVER)

.PHONY: help python docutils symlink