This copies the performance hack from encoding.lower (e7a5733d533f).
The case-folding logic that kicks in on case-insensitive filesystems
hits encoding.upper hard: with a repository with 75k files, the
timings went from
hg perfstatus
! wall 3.156000 comb 3.156250 user 1.625000 sys 1.531250 (best of 3)
to
hg perfstatus
! wall 2.390000 comb 2.390625 user 1.078125 sys 1.312500 (best of 5)
This is a 24% decrease. For comparison, Mercurial 2.0 gives:
hg perfstatus
! wall 2.172000 comb 2.171875 user 0.984375 sys 1.187500 (best of 5)
so we're only 10% slower than before we added the extra case-folding
logic.
The same decrease is seen when executing 'hg status' as normal, where
we go from:
hg status --time
time: real 4.322 secs (user 2.219+0.000 sys 2.094+0.000)
to
hg status --time
time: real 3.307 secs (user 1.750+0.000 sys 1.547+0.000)
When calling encode on a str, the string is first decoded using the
default encoding and then encoded. So
s.encode('ascii') == s.decode().encode('ascii')
We don't care about the encode step here -- we're just after the
UnicodeDecodeError raised by decode if it finds a non-ASCII character.
This way is also marginally faster since it saves the construction of
the extra str object.
The usercache hardlinks files for filesystems that support hardlinks. So when
this test is run on a filesystem that supports hardlinking, we have a link
in r7 and in the usercache, pointing to the same file contents (4cdac4d8...).
vfat does not support hardlinks, the file in the cache and in the store of
r7 are thus independent.
For the test to pass on vfat, we need to corrupt *both* the largefile in the
usercache and in the store of r7. Corrupting only one of them is not sufficient.
Fixes:
--- /home/buildslave/mercurial/vfat_hg_tests__stable_/build/tests/test-largefiles.t
+++ /home/buildslave/mercurial/vfat_hg_tests__stable_/build/tests/test-largefiles.t.err
@@ -997,9 +997,11 @@
$ hg push -R r7 http://localhost:$HGPORT1
pushing to http://localhost:$HGPORT1/
searching for changes
- remote: largefiles: failed to put 4cdac4d8b084d0b599525cf732437fb337d422a8 into store: largefile contents do not match hash
- abort: remotestore: could not put $TESTTMP/r7/.hg/largefiles/4cdac4d8b084d0b599525cf732437fb337d422a8 to remote store http://localhost:$HGPORT1/ (glob)
- [255]
+ searching for changes
+ remote: adding changesets
+ remote: adding manifests
+ remote: adding file changes
+ remote: added 2 changesets with 2 changes to 2 files
$ rm -rf empty
Push a largefiles repository to a served empty repository
An easy way to force this (and cause a traceback) prior to the fix for 3507 was
$ touch large
$ hg add --large large
$ hg ci -m "add"
$ hg remove large
$ touch large
$ hg addremove --config largefiles.patterns=**large
This patch also detected (and corrected) a previous test where a standin got
added as a largefile (without a traceback).
The problem only occurred if a file was removed with 'hg rm' (as opposed to the
OS utilities), and then addremove was run before a commit. Both normal and
large files were affected.
Ensuring that the file exists prior to an lstat() for size seems like the Right
Thing. But oddly enough, the missing file that was causing lstat() to blow up
was a standin when a largefile was removed, which seems fishy, because a standin
should never be added as a largefile. I was then able to get a standin added as
a largefile (whose name is 'large') with
hg addremove --config largefiles.patterns=**large
which also causes a backtrace. That will be fixed next.
The example in comment #9 of the bug writeup must be run exactly- it was the
commit after the rm and prior to the addremove that screwed things up, because
that commit noticed that the largefile was missing, called drop(), and then the
original commit function did nothing (due to the file in the '!' state). The
addremove command properly put it into the 'R' state, but it remained stuck in
that state (because commit insisted 'nothing changed'). Without the commit
prior to addremove, the problem didn't occur.
Maybe this is an indication that lfdirstate needs to take a few more hints from
the regular dirstate, regardless of what _it_ thinks the state is- similar
inconsistency is probably still possible with this patch if the original commit
succeeds but the lfdirstate write fails.
If a file was missing, the missing list contained a path relative to the repo.
When building the matcher from that list, the file name ended up concatenated to
cwd, causing the command to abort with '<file> not under root'. This rebuilds
the missing list with paths relative to cwd.
The "worst" extension still is the one tested with the lowest tested version
below the current version of Mercurial, but if an extension with was only
tested with newer versions, it is considered a candidate for a bad extension,
too. In this case extensions which have been tested with higher versions of
Mercurial are considered better. This allows finding the oldest extension if
ct can't be calculated correctly and therefore defaults to an empty tuple, and
it involves less changes to the comparison logic during the current code
freeze.
Changeset 8c1e21a3407c caused this when the "from win32 import *" line
was replaced with explicit import statements: the wildcard import was
at the bottom of the file and so windows.termwidth was overwritten by
win32.termwidth as indented, but the new explicit import statements
were at the top and so win32.termwidth got lost.
With the switch to ctypes, win32 can always be imported and so the
fallback termwidth in windows is no longer needed.
If we push some successors they will likely create a new head on
remote. However as the obsoleted head will disappear after the push we
are not really increasing the number of heads.
There is several case which will lead to extra being actually pushed. But this
first changeset aims to be simple. See the inline comment for details.
Without this change, you need to push --force every time you want to
push a newer version which is very error prone.
The remote side still display +n heads on unbundle because it does not have the
obsolete marker at unbundle time.
This function yield every nodes which succeed to a group of nodes.
The first user will be checkheads who need to know if we push successors for
remote extra heads.
The checkheads function is far too complicated. This extract help to explicite
what part of the preprocessing are reused by the actual check.
This the first step toward a wider refactoring.
The repo.hiddenrevs set is updated with all extinct() changesets which aren't
descendants of either:
- the current working copy,
- a bookmark,
- a tag.
This set is always accessed through the repo for now. Having this set
carried by the changelog make it complicated to:
- initialize it, computing hidden set may involve revset call
- lazy compute it, (1) only the changelog can detect someone access it,
(2) only the repo have enought knowledge to compute it.
In later version I expect he changelog to apply filtering itself and the set to
be carried by changelog again.
Extinct changesets are excluded from all exchange operations. This is a silent
exclusion because the user should not need to be aware of them.
There is no reason to strongly enforce this exclusion except implementation
simplicity. User should be able to explicitly request an extinct changeset in
the future.
They were previously inside the mercurial.phases module, but obsolete
logic will need them to exclude `extinct` changesets from pull and
push.
The proper and planned way to implement such filtering is still to apply a
changelog level filtering. But we are far to late in the cycle to implement and
push such a critical piece of code (changelog filtering). With Matt Mackall
approval I'm extending this quick and dirty mechanism for obsolete purpose.
Changelog level filtering should come during the next release cycle.
bffd8f8dfc85 claims this was needed "to avoid cyclic dependency", but there is
no cyclic dependency.
windows.py already imports encoding, posix.py can import it too, so we can
simply use encoding.upper in windows.py and in posix.py.
(this is a partial backout of bffd8f8dfc85)
The function is now able to write the version header as necessary. The function
now yield bytes to be written to a stream.
This should ease later use of this function for wireprotocol based exchanged.
Prepare the public use of the writemarker by wireprotocol function.
Adds new web command to the core, ``comparison``, which enables colorful
side-by-side change display, which for some might be much easier to work with
than the standard line diff output. The idea how to implement comes from the
SonicHq extension.
The web interface gets a new link to call the comparison functionality. It lets
users configure the amount of context lines around change blocks, or to show
full files - check help (also in this changeset) for details and defaults. The
setting in hgrc can be overridden by adding ``context=<value>`` to the request
query string. The comparison creates addressable lines, so as to enable sharing
links to specific lines, just as standard diff does.
Incorporates updates to all web related styles.
Known limitations:
* the column diff is done against the first parent, just as the standard diff
* this change allows examining diffs for single files only (as I am not sure if
examining the whole changeset in this way would be helpful)
* syntax highlighting of the output changes is not performed (enabling the
highlight extension has no influence on it)
There may be a more generic way that would add revset support to more commands
by adding revset support to addbranchrevs(), but given the proximity of the next
code freeze, a minimal change seems like the better choice.
When loading a python hook with file syntax fails, there is no
information that this happened while loading a hook. When the python
file does not exist even the file name is not printed. (Only that a
file is missing.)
This patch adds this information and a test for loading a non existing file and
a directory not being a python module.
When the source repository had a revision renaming "$new -> $old",
but the filemap a "$old -> $new" rename, the converted revision could
use either $new (deleting the file) or $old (keeping the file) when
getting the file data, depending on the lexicographical order of
those names. So the resulting revision would leave some files
untouched (as expected), but delete others arbitrarely.
Plain 'hg help rollback' now looks like this:
$ hg help rollback
hg rollback
roll back the last transaction (dangerous)
This command should be used with care. There is only one level of
rollback, and there is no way to undo a rollback. It will also restore
the dirstate at the time of the last transaction, losing any dirstate
changes since that time. This command does not alter the working
directory.
Transactions are used to encapsulate the effects of all commands that
create new changesets or propagate existing changesets into a repository.
This command is not intended for use on public repositories. Once changes
are visible for pull by other users, rolling a transaction back locally
is ineffective (someone else may already have pulled the changes).
Furthermore, a race is possible with readers of the repository; for
example an in-progress pull from the repository may fail if a rollback is
performed.
Returns 0 on success, 1 if no rollback data is available.
options:
-n --dry-run do not perform actions, just print output
-f --force ignore safety measures
--mq operate on patch repository
use "hg -v help rollback" to show more info
state == 'a' implies check
I fail to see what the point of this check parameter is. Near as I can see,
the only _addpath call where it was set to True was in add(), but there, state
is 'a'.
This is a follow-up to 24a646d9943a.
This change separates peer implementations from the repository implementation.
localpeer currently is a simple pass-through to localrepository, except for
legacy calls, which have already been removed from localpeer. This ensures that
the local client code only uses the most modern peer API when talking to local
repos.
Peers have a .local() method which returns either None or the underlying
localrepository (or descendant thereof). Repos have a .peer() method to return
a freshly constructed localpeer. The latter is used by hg.peer(), and also to
allow folks to pass either a peer or a repo to some generic helper methods.
We might want to get rid of .peer() eventually.
The only user of locallegacypeer is debugdiscovery, which uses it to pose as a
pre-setdiscovery client. But we decided to leave the old API defined in
locallegacypeer for clarity and maybe for other uses in the future.
It might be nice to actually define the peer API directly in peer.py as stub
methods. One problem there is, however, that localpeer implements
lock/addchangegroup, whereas the true remote peers implement unbundle.
It might be desireable to get rid of this distinction eventually.
This introduces a peer method into all repository classes, which currently
simply returns self. It also changes hg.repository so it now raises an
exception if the supplied paths does not resolve to a localrepo or descendant.
Finally, all call sites are changed to use the peer and local methods as
appropriate, where peer is used whenever the code is dealing with a remote
repository (even if it's on local disk).