Joey Hess [Wed, 16 Mar 2022 16:37:09 +0000 (12:37 -0400)]
clean up concurrent output of tests
Using concurrent-output this is easy. Just have to check if tasty has
color enabled, and propagate it into the worker processes, some of which
will be run without a controlling console.
Also added a call to installSignalHandlers; I noticed that interrupting
the test suite could leave the console in a bad state and this fixes
that.
The ansi-terminal dependency is free, since tasty also depends on it.
Joey Hess [Mon, 14 Mar 2022 21:23:52 +0000 (17:23 -0400)]
allow tests to be split for more granularity
Unit tests are the main bulk of runtime, so splitting them into 2 or 3
parts should help.
For now, the number of parts is still 1, because on my 4 core laptop,
2 was a little bit slower, and 3 slower yet. However, this probably does
vary based on the number of cores, so needs to be revisited, and perhaps
made dynamic.
Since each test mode gets split into the specified number of parts,
plus property and remote tests, 2 gives 8 parts, and 3 gives 11 parts.
Load went to maybe 18, so there was probably contention slowing things
down.
So probably it needs to start N workers with some parts, and when a
worker finishes, run it with the next part, until all parts are
processed.
Joey Hess [Mon, 14 Mar 2022 19:24:37 +0000 (15:24 -0400)]
parallelize git-annex test for 25% speedup
Note the very weird workaround for what appears to be some kind of tasty
bug, which causes a segfault. This is not new to this modification,
I was seeing a segfault before at least intermittently when limiting
git-annex test -p to only run a single test group.
Also, the path from one test repo to a remote test repo used to be
"../../foo", which somehow broke when moving the test repos from .t to
.t/N. I don't actually quite understand how it used to work, but
"../foo" seems correct and works in the new situation.
Test output from the concurrent processes is not yet serialized.
Should be easy to do using concurrent-output.
More test groups will probably make the speedup larger. It would
probably be best to have a larger number of test groups and divvy them
amoung subprocesses numbered based on the number of CPU cores, perhaps
times 2 or 3.
Joey Hess [Mon, 7 Mar 2022 18:01:45 +0000 (14:01 -0400)]
hard link git-annex to fix ./git-annex test
2fc46e1871e2f793689cc98523744040c9010bdc broke running ./git-annex test
when git-annex-shell is not in PATH from another install, and also
caused the wrong git-annex-shell to be tested.
Using a symlink made programPath return the .cabal/ directory where the
git-annex binary is, but git-annex-shell is not in that directory.
Changing to a hard link avoids that problem.
Joey Hess [Wed, 2 Mar 2022 22:22:38 +0000 (18:22 -0400)]
add back lost packString
The patch that removed it did not break anything, since the strings it's
used on are all ASCII not unicode. But I like making sure to use
packString everywhere just in case the code later changes in a way that
needs it.
sternenseemann [Wed, 2 Mar 2022 20:16:10 +0000 (21:16 +0100)]
allow building with aeson >= 2.0
In aeson 2.0, Text has been replaced by the Key type and HashMap by the
KeyMap interface. Accomodating this required adding some CPP in order to
still be able to compile with aeson < 2.0. The required changes were:
* Prevent Key from being re-exported by Utilities.Aeson, as it clashes
with git-annex's own Key type.
* Fix up convertion from String/Text to Key (or Text in aeson 1.*) in a
couple of places
* Import Data.Aeson.KeyMap instead of Data.HashMap.Strict, as they are
mostly API-compatible. insertWith needs to be replaced by unionWith,
however, as KeyMap lacks the former function.
Joey Hess [Tue, 1 Mar 2022 20:08:55 +0000 (16:08 -0400)]
git-annex test from standalone speedup
Avoid git-annex test being very slow when run from within the standalone
linux tarball or OSX app.
It may not really be necessary to add to PATH the directory where the
git-annex binary resides, but it can't hurt. Most places where the test
suite or git-annex run git-annex, they use programPath, so won't need
a modified PATH. But I'm not sure if that's always the case.
Joey Hess [Tue, 1 Mar 2022 18:26:04 +0000 (14:26 -0400)]
remove broken test
After ce91f10132805d11448896304821b0aa9c6d9845, unannex on a non-annexed
file is expected to fail, unless annex.skipunknown were globally
overridden to true. So this test that used to succeed should probably
fail, but it seems better to just remove it than expect it to fail.
Joey Hess [Mon, 28 Feb 2022 17:26:26 +0000 (13:26 -0400)]
remove 127/128 exit status mention
I cannot find any indication that git-annex uses these exit statuses,
and I see I didn't write this doc so I don't know where those numbers
came from. Odd.
Joey Hess [Mon, 28 Feb 2022 16:54:56 +0000 (12:54 -0400)]
fix annex.skipunknown false error propagation
Propagate nonzero exit status from git ls-files when a specified file does
not exist, or a specified directory does not contain any files checked into
git.
The recent completion of the annex.skipunknown transition exposed this
bug, that has unfortunately been lurking all along.
It is also possible that git ls-files errors out for some other reason
-- perhaps a permission problem -- and this will also fix error propagation
in such situations.
Joey Hess [Thu, 24 Feb 2022 18:21:23 +0000 (14:21 -0400)]
avoid accidentally thawing git-annex symlink
It did nothing, since at this point the link is dangling. But when there
is a thaw hook, it would probably not be happy to be asked to run on a
symlink, or might do something unexpected.
projects / git-annex.git / log