Mercurial > hg
view tests/test-config.t @ 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
author | Gregory Szorc <gregory.szorc@gmail.com> |
---|---|
date | Wed, 18 Jul 2018 09:49:34 -0700 |
parents | a22915edc279 |
children | edbcf5b239f9 |
line wrap: on
line source
hide outer repo $ hg init Invalid syntax: no value $ cat > .hg/hgrc << EOF > novaluekey > EOF $ hg showconfig hg: parse error at $TESTTMP/.hg/hgrc:1: novaluekey [255] Invalid syntax: no key $ cat > .hg/hgrc << EOF > =nokeyvalue > EOF $ hg showconfig hg: parse error at $TESTTMP/.hg/hgrc:1: =nokeyvalue [255] Test hint about invalid syntax from leading white space $ cat > .hg/hgrc << EOF > key=value > EOF $ hg showconfig hg: parse error at $TESTTMP/.hg/hgrc:1: key=value unexpected leading whitespace [255] $ cat > .hg/hgrc << EOF > [section] > key=value > EOF $ hg showconfig hg: parse error at $TESTTMP/.hg/hgrc:1: [section] unexpected leading whitespace [255] Reset hgrc $ echo > .hg/hgrc Test case sensitive configuration $ cat <<EOF >> $HGRCPATH > [Section] > KeY = Case Sensitive > key = lower case > EOF $ hg showconfig Section Section.KeY=Case Sensitive Section.key=lower case $ hg showconfig Section -Tjson [ { "name": "Section.KeY", "source": "*.hgrc:*", (glob) "value": "Case Sensitive" }, { "name": "Section.key", "source": "*.hgrc:*", (glob) "value": "lower case" } ] $ hg showconfig Section.KeY -Tjson [ { "name": "Section.KeY", "source": "*.hgrc:*", (glob) "value": "Case Sensitive" } ] $ hg showconfig -Tjson | tail -7 }, { "name": "*", (glob) "source": "*", (glob) "value": "*" (glob) } ] Test empty config source: $ cat <<EOF > emptysource.py > def reposetup(ui, repo): > ui.setconfig(b'empty', b'source', b'value') > EOF $ cp .hg/hgrc .hg/hgrc.orig $ cat <<EOF >> .hg/hgrc > [extensions] > emptysource = `pwd`/emptysource.py > EOF $ hg config --debug empty.source read config from: * (glob) none: value $ hg config empty.source -Tjson [ { "name": "empty.source", "source": "", "value": "value" } ] $ cp .hg/hgrc.orig .hg/hgrc Test "%unset" $ cat >> $HGRCPATH <<EOF > [unsettest] > local-hgrcpath = should be unset (HGRCPATH) > %unset local-hgrcpath > > global = should be unset (HGRCPATH) > > both = should be unset (HGRCPATH) > > set-after-unset = should be unset (HGRCPATH) > EOF $ cat >> .hg/hgrc <<EOF > [unsettest] > local-hgrc = should be unset (.hg/hgrc) > %unset local-hgrc > > %unset global > > both = should be unset (.hg/hgrc) > %unset both > > set-after-unset = should be unset (.hg/hgrc) > %unset set-after-unset > set-after-unset = should be set (.hg/hgrc) > EOF $ hg showconfig unsettest unsettest.set-after-unset=should be set (.hg/hgrc) Test exit code when no config matches $ hg config Section.idontexist [1] sub-options in [paths] aren't expanded $ cat > .hg/hgrc << EOF > [paths] > foo = ~/foo > foo:suboption = ~/foo > EOF $ hg showconfig paths paths.foo:suboption=~/foo paths.foo=$TESTTMP/foo edit failure $ HGEDITOR=false hg config --edit abort: edit failed: false exited with status 1 [255] config affected by environment variables $ EDITOR=e1 VISUAL=e2 hg config --debug | grep 'ui\.editor' $VISUAL: ui.editor=e2 $ VISUAL=e2 hg config --debug --config ui.editor=e3 | grep 'ui\.editor' --config: ui.editor=e3 $ PAGER=p1 hg config --debug | grep 'pager\.pager' $PAGER: pager.pager=p1 $ PAGER=p1 hg config --debug --config pager.pager=p2 | grep 'pager\.pager' --config: pager.pager=p2 verify that aliases are evaluated as well $ hg init aliastest $ cd aliastest $ cat > .hg/hgrc << EOF > [ui] > user = repo user > EOF $ touch index $ unset HGUSER $ hg ci -Am test adding index $ hg log --template '{author}\n' repo user $ cd .. alias has lower priority $ hg init aliaspriority $ cd aliaspriority $ cat > .hg/hgrc << EOF > [ui] > user = alias user > username = repo user > EOF $ touch index $ unset HGUSER $ hg ci -Am test adding index $ hg log --template '{author}\n' repo user $ cd ..