This adds a __nonzero__ method to manifestdict. This isn't strictly necessary in
the vanilla Mercurial implementation, since Python will handle nonzero checks by
using __len__, but having it implemented here makes it easier for alternative
implementations to implement __nonzero__ and have them be plug-n-play with the
normal implementation.
sys.argv returns unicodes on Python 3. We need a bytes version for us.
There was also a python bug/feature request which wanted then to implement
one. They rejected and it is quoted in one of the comments that we can use
fsencode() to get a bytes version of sys.argv. Though not sure about its
correctness.
Link to the comment: http://bugs.python.org/issue8776#msg217416
After this patch we will have pycompat.sysargv which will return us bytes
version of sys.argv. If this patch goes in, i will like to make transformer
rewrite sys.argv with pycompat.argv because there are lot of occurences.
This patch replaces lyhash with the hash algorithm used by diffutils.
The algorithm has its origins in Git commit 2e9d1410, which is all the
way back from 1992. The license header in the code at that revision
in GPL v2.
I have not performed an extensive analysis of the distribution
(and therefore buckets) of hash output. However, `hg perfbdiff`
gives some clear wins. I'd like to think that if it is good enough
for diffutils it is good enough for us?
From the mozilla-unified repository:
$ perfbdiff -m 3041e4d59df2
! wall 0.053271 comb 0.060000 user 0.060000 sys 0.000000 (best of 100)
! wall 0.035827 comb 0.040000 user 0.040000 sys 0.000000 (best of 100)
$ perfbdiff 0e9928989e9c --alldata --count 100
! wall 6.204277 comb 6.200000 user 6.200000 sys 0.000000 (best of 3)
! wall 4.309710 comb 4.300000 user 4.300000 sys 0.000000 (best of 3)
From the hg repo:
$ perfbdiff 35000 --alldata --count 1000
! wall 0.660358 comb 0.660000 user 0.660000 sys 0.000000 (best of 15)
! wall 0.534092 comb 0.530000 user 0.530000 sys 0.000000 (best of 19)
Looking at the generated assembly and statistical profiler output
from the kernel level, I believe there is room to make this function
even faster. Namely, we're still consuming data character by character
instead of at the word level. This translates to more loop iterations
and more instructions.
At this juncture though, the real performance killer is that we're
hashing every line. We should get a significant speedup if we change
the algorithm to find the longest prefix, longest suffix, treat those
as single "lines" and then only do the line splitting and hashing on
the parts that are different. That will require a lot of C code,
however. I'm optimistic this approach could result in a ~2x speedup.
The statprof sampling profiler runs with significantly less overhead.
Its data is therefore more useful. Furthermore, its default output
shows the hotpath by default, which I've found to be way more useful
than the default profiler's function time table.
There is one behavioral regression with this change worth noting:
the statprof profiler currently doesn't profile individual hgweb
requests like lsprof does. This is because the current implementation
of statprof only profiles the thread that started profiling.
The ability for lsprof to profile individual hgweb requests is
relatively new and likely not widely used. Furthermore, I have plans
to modify statprof to support profiling multiple threads. I expect
that change to go through several iterations. I'm submitting this
patch first so there is more time to test statprof. Perfect is the
enemy of good.
Now that the statprof module is vendored and suitable for use, we
switch our statprof profiler to use it. This required some minor
changes because of drift between the official statprof profiler
and the vendored copy.
We also incorporate Facebook's improvements from the "statprofext"
extension at
https://bitbucket.org/facebook/hg-experimental, notably support for
different display formats.
Because statprof output is different, this is marked as BC. Although
most users likely won't notice since most users don't profile.
It appears crecord.py has its own termsize() function. I want to get rid of it.
The fallback height is chosen from the default of cmd.exe on Windows, and
VT100 on Unix.
I was a bit confused since we didn't add 1 to the width, which is different
from the example shown in StackOverflow.
http://stackoverflow.com/a/12642749
The array module must exist. It's sufficient to suppress the ImportError of
termios. Also salvaged the comment why we have to handle AttributeError, from
54db81f689bd.
I'm going to get rid of sys.stderr|out|in references from posix.termwidth().
In order to do that, termwidth() needs to take a ui, but functions in util.py
shouldn't depend on a ui object. So moves termwidth() to scmutil.py.
`plast = a + len - 1`. So, this "for" loop iterates from "a" to "plast",
inclusive. So, `p == plast` can only be true on the final iteration
of the loop. So checking for it on every loop iteration is wasteful.
This patch simply decreases the upper bound of the loop by 1 and
adds an explicit check after iteration for the `p == plast` case.
We can't simply add 1 to the initial value for "i" because that
doesn't do the correct thing on empty input strings.
`perfbdiff -m 3041e4d59df2` on the Firefox repo becomes significantly
faster:
! wall 0.072763 comb 0.070000 user 0.070000 sys 0.000000 (best of 100)
! wall 0.053221 comb 0.060000 user 0.060000 sys 0.000000 (best of 100)
For the curious, this code has its origins in 3605ecc924ef, which is
the changeset that introduced bdiff.c in 2005.
Also, GNU diffutils is able to perform a similar line-based diff in
under 20ms. So there's likely more perf wins to be found in this code.
One of them is the hashing algorithm. But it looks like mpm spent
some time testing hash collisions in c22788816627. I'd like to do the
same before switching away from lyhash, just to be on the safe side.
This patch does not exactly solve issue5228 but it results in a better
condition on this issue. For disabled extensions, we used to parse the
module and get the first occurrences of docstring and then return the first
line of that as an introductory heading of extension. This is what we get
today.
This patch returns the whole docstring of the module as a help for extension,
which is more informative. There are some modules which don't have much
docstring at top level except the heading so those are unaffected by this
change. To follow the existing trend of showing commands either we have to
load the extension or have a very ugly parsing method which don't even assure
correctness.
This patch make sure scmutil.rcpath() returns bytes independent of
which platform is used on Python 3. If we want to change type for windows we
can just conditionalize the return variable.
Certain instances of os.sep has been converted to pycompat.ossep where it was
sure to use bytes only. There are more such instances which needs some more
attention and will get surely.
os.name returns unicodes on py3. Most of our checks are like
os.name == 'nt'
Because of the transformer, on the right hand side we have b'nt'. The
condition will never satisfy even if os.name returns 'nt' as that will be an
unicode.
We either need to encode every occurence of os.name or have a
new variable which is much cleaner. Now we have pycompat.osname.
There are around 53 occurences of os.name in the codebase which needs to
be replaced by pycompat.osname to support Python 3.
In this patch we make util.datapath a bytes variable, but we have to pass a
unicode to gettext.translation otherwise it will cry. Used pycompat.fsdecode()
to decode it back to unicode as it was converted to bytes using
pycompat.fsencode().
test-convert-darcs.t suddenly started failing on my Debian sid machine. The
reason was Darcs was upgraded from 2.12.0 to 2.12.4 so the original pattern
got to match the last two digits. Fix the pattern to match 2.2+.
The old manifest had different functions for performing shallow reads, shallow
readdeltas, and shallow readfasts. Since a lot of the code is duplicate (and
since those functions don't make sense on a normal manifestctx), let's unify
them into flags on the existing readdelta and readfast functions.
A future diff will change consumers of these functions to use the manifestctx
versions and will delete the old apis.
In the last patch we added a get() function that allows fetching directory level
treemanifestctxs. It didn't handle caching at directory level though, so we need to
change our mancache to support multiple directories.
Previously manifestlog only allowed obtaining root level manifests. Future
patches will need direct access to subdirectory manifests as part of changegroup
creation, so let's add a get() function that knows how to deal with
subdirectories.
When accessing a manifest via manifestlog[node], let's verify that the node
actually exists and throw a LookupError if it doesn't. This matches the old read
behavior, so we don't accidentally return invalid manifestctxs.
We do this in manifestlog instead of in the manifestctx/treemanifestctx
constructors because the treemanifest code currently relies on the fact that
certain code paths can produce treemanifests without touching the revlogs (and
it has tests that verify things work if certain revlogs are missing entirely, so
they break if we add validation that tries to read them).
In many cases, _chunkraw() is called with startrev==endrev. When
this is true, we can avoid an extra index lookup and some other
minor operations.
On the mozilla-unified repo, `hg perfrevlogchunks -c` says this
has the following impact:
! read w/ reused fd
! wall 0.371846 comb 0.370000 user 0.350000 sys 0.020000 (best of 27)
! wall 0.337930 comb 0.330000 user 0.300000 sys 0.030000 (best of 30)
! read batch w/ reused fd
! wall 0.014952 comb 0.020000 user 0.000000 sys 0.020000 (best of 197)
! wall 0.014866 comb 0.010000 user 0.000000 sys 0.010000 (best of 196)
So, we've gone from ~25x slower than batch to ~22.5x slower.
At this point, there's probably not much else we can do except
implement an optimized function in the index itself, including in C.
When I implemented `hg perfrevlogchunks`, one of the things that
stood out was N * _chunk() calls was ~38x slower than 1
_chunks() call. Specifically, on the mozilla-unified repo:
N*_chunk: 0.528997s
1*_chunks: 0.013735s
This repo has 352,097 changesets. So the average time per changeset
comes out to:
N*_chunk: 1.502us
1*_chunks: 0.039us
If you extrapolate these numbers to a repository with 1M changesets,
that comes out to 1.502s versus 0.039s, which is significant.
At these latencies, Python attribute lookups and function calls
matter. So, this patch inlines some code to cut down on that overhead.
The impact of this patch on N*_chunk() calls is clear:
! wall 0.528997 comb 0.520000 user 0.500000 sys 0.020000 (best of 19)
! wall 0.367723 comb 0.370000 user 0.350000 sys 0.020000 (best of 27)
So, we go from ~38x slower to ~27x. A nice improvement. But there's
still a long way to go.
It's worth noting that functionality like revsets perform changelog
lookups one revision at a time. So this code path is worth optimizing.
Index entries are ordered tuples. We have accessors in the revlog
class to map tuple offsets to names. To help reinforce the order,
reorder the methods so they match the order of elements in the
tuple. While I'm here, also sneak in some minimal documentation.
The 'hg debugcolor' command gains a '--style' flag to display all the configured
labels and their styles. This have many benefits:
* discovering documented label,
* checking consistency between label's style,
* showing the actual style of a label.
The previous ordering were provided by the set. The new output is more stable
and rational. In addition we have some logic to keep the '_background' version
together to help readability.
We are about to introduce a second mode for 'hg debugcolor' that would list the
known label and their configuration, so we split the code related to color and
effect out of the main function.
Before this change, running 'debugcolor' would destroy all color style for the
rest of the process life. We now properly backup and restore the variable
content. Using a global variable is sketchy in general and could probably be
removed. However, this is a quest for another adventure.
That test file is very small and is merge with the new 'test-push.t'. No logic
is changed.
We don't register this as a copy because is actually a "ypoc" merging two file
together without replacing the destination and Mercurial cannot express that.
That test file is very small and is merge with the new 'test-push.t'. No logic
is changed but repository name are update to avoid collision.
We don't register this as a copy because is actually a "ypoc" merging two file
together without replacing the destination and Mercurial cannot express that.
We do not have a simple test for 'hg push' but we have multiple tiny tests for
various aspect of it. We'll unify them into a single file, and we start with
'test-push-r.t'. The code is unchanged but we renamed the repository used to
avoid collision with other tests we'll import in coming changesets.
Test timing for the record:
start end cuser csys real Test
1.850 2.640 0.650 0.090 0.790 test-push-validation.t
2.640 3.520 0.760 0.090 0.880 test-push-hook-lock.t
0.000 1.850 1.560 0.210 1.850 test-push-r.t
I came across this code by chance. The script of this test is a bit messy with a
lot of unnecessary intermediate commands. We simplify the script and unify
repository access through '-R'.
In the process the update after the unbundle is dropped as it does not add
anything to the tests.