##// END OF EJS Templates
revlog: use compression engine APIs for decompression...
revlog: use compression engine APIs for decompression Now that compression engines declare their header in revlog chunks and can decompress revlog chunks, we refactor revlog.decompress() to use them. Making full use of the property that revlog compressor objects are reusable, revlog instances now maintain a dict mapping an engine's revlog header to a compressor object. This is not only a performance optimization for engines where compressor object reuse can result in better performance, but it also serves as a cache of header values so we don't need to perform redundant lookups against the compression engine manager. (Yes, I measured and the overhead of a function call versus a dict lookup was observed.) Replacing the previous inline lookup table with a dict lookup was measured to make chunk reading ~2.5% slower on changelogs and ~4.5% slower on manifests. So, the inline lookup table has been mostly preserved so we don't lose performance. This is unfortunate. But many decompression operations complete in microseconds, so Python attribute lookup, dict lookup, and function calls do matter. The impact of this change on mozilla-unified is as follows: $ hg perfrevlogchunks -c ! chunk ! wall 1.953663 comb 1.950000 user 1.920000 sys 0.030000 (best of 6) ! wall 1.946000 comb 1.940000 user 1.910000 sys 0.030000 (best of 6) ! chunk batch ! wall 1.791075 comb 1.800000 user 1.760000 sys 0.040000 (best of 6) ! wall 1.785690 comb 1.770000 user 1.750000 sys 0.020000 (best of 6) $ hg perfrevlogchunks -m ! chunk ! wall 2.587262 comb 2.580000 user 2.550000 sys 0.030000 (best of 4) ! wall 2.616330 comb 2.610000 user 2.560000 sys 0.050000 (best of 4) ! chunk batch ! wall 2.427092 comb 2.420000 user 2.400000 sys 0.020000 (best of 5) ! wall 2.462061 comb 2.460000 user 2.400000 sys 0.060000 (best of 4) Changelog chunk reading is slightly faster but manifest reading is slower. What gives? On this repo, 99.85% of changelog entries are zlib compressed (the 'x' header). On the manifest, 67.5% are zlib and 32.4% are '\0'. This patch swapped the test order of 'x' and '\0' so now 'x' is tested first. This makes changelogs faster since they almost always hit the first branch. This makes a significant percentage of manifest '\0' chunks slower because that code path now performs an extra test. Yes, I too can't believe we're able to measure the impact of an if..elif with simple string compares. I reckon this code would benefit from being written in C...

File last commit:

r29747:aba2bb2a default
r30817:2b279126 default
Show More
requirements.txt
108 lines | 3.3 KiB | text/plain | TextLexer
Repositories contain a file (``.hg/requires``) containing a list of
features/capabilities that are *required* for clients to interface
with the repository. This file has been present in Mercurial since
version 0.9.2 (released December 2006).
One of the first things clients do when opening a repository is read
``.hg/requires`` and verify that all listed requirements are supported,
aborting if not. Requirements are therefore a strong mechanism to
prevent incompatible clients from reading from unknown repository
formats or even corrupting them by writing to them.
Extensions may add requirements. When they do this, clients not running
an extension will be unable to read from repositories.
The following sections describe the requirements defined by the
Mercurial core distribution.
revlogv1
========
When present, revlogs are version 1 (RevlogNG). RevlogNG was introduced
in 2006. The ``revlogv1`` requirement has been enabled by default
since the ``requires`` file was introduced in Mercurial 0.9.2.
If this requirement is not present, version 0 revlogs are assumed.
store
=====
The *store* repository layout should be used.
This requirement has been enabled by default since the ``requires`` file
was introduced in Mercurial 0.9.2.
fncache
=======
The *fncache* repository layout should be used.
The *fncache* layout hash encodes filenames with long paths and
encodes reserved filenames.
This requirement is enabled by default when the *store* requirement is
enabled (which is the default behavior). It was introduced in Mercurial
1.1 (released December 2008).
shared
======
Denotes that the store for a repository is shared from another location
(defined by the ``.hg/sharedpath`` file).
This requirement is set when a repository is created via :hg:`share`.
The requirement was added in Mercurial 1.3 (released July 2009).
dotencode
=========
The *dotencode* repository layout should be used.
The *dotencode* layout encodes the first period or space in filenames
to prevent issues on OS X and Windows.
This requirement is enabled by default when the *store* requirement
is enabled (which is the default behavior). It was introduced in
Mercurial 1.7 (released November 2010).
parentdelta
===========
Denotes a revlog delta encoding format that was experimental and
replaced by *generaldelta*. It should not be seen in the wild because
it was never enabled by default.
This requirement was added in Mercurial 1.7 and removed in Mercurial
1.9.
generaldelta
============
Revlogs should be created with the *generaldelta* flag enabled. The
generaldelta flag will cause deltas to be encoded against a parent
revision instead of the previous revision in the revlog.
Support for this requirement was added in Mercurial 1.9 (released
July 2011). The requirement was disabled on new repositories by
default until Mercurial 3.7 (released February 2016).
manifestv2
==========
Denotes that version 2 of manifests are being used.
Support for this requirement was added in Mercurial 3.4 (released
May 2015). The requirement is currently experimental and is disabled
by default.
treemanifest
============
Denotes that tree manifests are being used. Tree manifests are
one manifest per directory (as opposed to a single flat manifest).
Support for this requirement was added in Mercurial 3.4 (released
August 2015). The requirement is currently experimental and is
disabled by default.