Unlike changeset_printer, it does not hide the manifest field because JSON
output will be parsed by machine where explicit "null" will be more useful
than nothing.
A style name should not contain "/", "\", "." and "..". Otherwise, templates
could be loaded from outside of the specified templates directory by invalid
?style= parameter. hgweb should not allow such requests.
This change means subdir/name is also rejected.
We should consider add HTML rendering of the RST into the response as a
follow-up. I attempted to do this, but there was an empty array
returned by the rstdoc() template function. Not sure what's going on.
Will deal with it later.
These are the same dispatch function under the hood. The only difference
is the default number of entries to render and the template to use. So
it makes sense to use a shared template.
Format for {changelistentry} is similar to {changeset}. However, there
are differences to argument names and their values preventing us from
(easily) using the same template. (Perhaps there is room to consolidate
the templates as a follow-up.)
We're currently not recording some data in {changelistentry} that exists
in {changeset}. This includes the branch name. This should be added in
a follow-up. For now, something is better than nothing.
In git 2.2.0 and higher, removing files and directories is changed:
removing an object that does not exist returns success rather than failure.
As a result, even though .hg/hgrc does not exist, success is returned
and the .hg/ directory is removed.
To handle this correctly, use 'rm -rf' to allow successful removing
for all git versions.
The exact changeset where this was introduced in git:
1054af7d04aef64378d69a0496b45cdbf6a0bef2
wrapper.c: remove/unlink_or_warn: simplify, treat ENOENT as success
The new manifest format is designed to be smaller, in particular to
produce smaller deltas. It stores hashes in binary and puts the hash
on a new line (for smaller deltas). It also uses stem compression to
save space for long paths. The format has room for metadata, but
that's there only for future-proofing. The parser thus accepts any
metadata and throws it away. For more information, see
http://mercurial.selenic.com/wiki/ManifestV2Plan.
The current manifest format doesn't allow an empty filename, so we use
an empty filename on the first line to tell a manifest of the new
format from the old. Since we still never write manifests in the new
format, the added code is unused, but it is tested by
test-manifest.py.
While it should be safe to switch to the new manifest format on an
existing repo, let's keep it simple for now and make the configuration
have any effect only at repo creation time. If the configuration is
enabled then (at repo creation), we add an entry to requires and read
that instead of the configuration from then on.
It's pretty surprising phase wasn't part of this template call already.
We now expose {phase} to the {changeset} template and we expose this
data to JSON.
This brings JSON output in line with the output from `hg log -Tjson`.
The lone exception is hweb doesn't print the numeric rev. As has been
stated previously, I don't believe hgweb should be exposing these
unstable identifiers. (We can add them later if we really want them.)
There is still work to bring hgweb in parity with --verbose and
--debug output from the CLI.
There were no tests for the various code paths in manifestdict.matches(), so I
added some. This also adds a more complex testing manifest so that any bugs
relating to traversal of directories are more likely to be caught.
Tags is pretty easy to implement. Let's start there.
The output is slightly different from `hg tags -Tjson`. For reference,
the CLI has the following output:
[
{
"node": "e2049974f9a23176c2addb61d8f5b86e0d620490",
"rev": 29880,
"tag": "tip",
"type": ""
},
...
]
Our output has the format:
{
"node": "0aeb19ea57a6d223bacddda3871cb78f24b06510",
"tags": [
{
"node": "e2049974f9a23176c2addb61d8f5b86e0d620490",
"tag": "tag1",
"date": [1427775457.0, 25200]
},
...
]
}
"rev" is omitted because it isn't a reliable identifier. We shouldn't
be exposing them in web APIs and giving the impression it remotely
resembles a stable identifier. Perhaps we could one day hide this behind
a config option (it might be useful to expose when running servers
locally).
The "type" of the tag isn't defined because this information isn't yet
exposed to the hgweb templater (it could be in a follow-up) and because
it is questionable whether different types should be exposed at all.
(Should the web interface really be exposing "local" tags?)
We use an object for the outer type instead of Array for a few reasons.
First, it is extensible. If we ever need to throw more global properties
into the output, we can do that without breaking backwards compatibility
(property additions should be backwards compatible). Second, uniformity
in web APIs is nice. Having everything return objects seems much saner than
a mix of array and object. Third, there are security issues with arrays
in older browsers. The JSON web services world almost never uses arrays
as the main type for this reason.
Another possibly controversial part about this patch is how dates are
defined. While JSON has a Date type, it is based on the JavaScript Date
type, which is widely considered a pile of garbage. It is a non-starter
for this reason.
Many of Mercurial's built-in date filters drop seconds resolution. So
that's a non-starter as well, since we want the API to be lossless where
possible. rfc3339date, rfc822date, isodatesec, and date are all lossless.
However, they each require the client to perform string parsing on top of
JSON decoding. While date parsing libraries are pretty ubiquitous, some
languages don't have them out of the box. However, pretty much every
programming language can deal with UNIX timestamps (which are just
integers or floats). So, we choose to use Mercurial's internal date
representation, which in JSON is modeled as float seconds since UNIX
epoch and an integer timezone offset from UTC (keep in mind
JavaScript/JSON models all "Numbers" as double prevision floating point
numbers, so there isn't a difference between ints and floats in JSON).
Many have long wanted hgweb to emit a common machine readable output.
We start the process by defining a stub json template.
Right now, each endpoint returns a stub "not yet implemented" string.
Individual templates will be implemented in subsequent patches.
Basic tests for templates have been included. Coverage isn't perfect,
but it is better than nothing.
Upcoming patches will add support for JSON output from hgweb.
Because JSON output from the templater is hard to read and because it
is easy to introduce malformed JSON, we introduce a JSON processing
mode to get-with-headers.py that will parse and pretty print JSON
from HTTP responses. This will make tests easier to read and write
and it will ensure hgweb is emitting well-formed JSON.
Caches should be transparent. If a cache is damaged, it should
silently be rebuilt, much like if it were invalid. No one seems to
have ever hit this in the wild.
Now that command running is part of Test, we no longer need to pass
a list of replacements down through various call layers.
The impetus for this change is to fetch replacements after
command execution, not before. This will allow replacements to be
defined as part of test execution.
Upcoming patches will change how the replacements system works
to make it more flexible. To prepare for this, eliminate the one-off
use of replacements to perform newline normalization on Windows.
Future patches will change how replacements work. Since the logic in
run() is strongly tied to the operation of individual tests and since
there is potential to make the implementation simpler by giving the
function access to Test attributes, move it into Test.
The test runner has the ability to stop on first error.
Tests are executed in new Python threads. The test runner starts new
threads when it has capacity to do so. Before this patch, the "stop on
first error" logic would return immediately from the "run tests"
function, without waiting on test threads to complete. There was thus
a race between the test runner thread doing cleanup work and the test
thread performing activity. For example, the test thread could be in
the middle of executing a test shell script and the test runner
could remove the test's temporary directory. Depending on timing, this
could result in any number of output from the test runner.
This patch eliminates the race condition by having the test runner
explicitly wait for test threads to complete before continuing.
I discovered this issue as I modified the test harness in a subsequent
patch and was reliably able to tickle the race condition.
As part of testing code coverage output, I noticed some files were
being reported twice: there was an entry for the file in the install
location and for the file in the source tree. I'm not sure why this
is. But it resulted in under-reporting of coverage data since some
lines weren't getting covered in both locations.
I also noticed that files in the source directory and outside the
"mercurial" and "hgext" packages were getting included in the
coverage report. Cosmetically, this seemed odd to me. It's not
difficult to filter paths from the report. But I figure this data
can be useful (we could start reporting run-tests.py coverage,
for example).
This patch switches the coverage API to report code coverage from
the source directory. It registers a path alias so that data from
the install location is merged into data from the source directory.
We now get merged results for files that were being reported in
multiple locations.
Since code coverage reporting now relies on the profiled install
now being in sync with the source tree, an additional check to
disallow code coverage when --with-hg is specified has been added.
This should have been present before, as --local was previously
disallowed for the same reasons.
Merging the paths raises our aggregate line coverage from ~60 to
81%.
Before this patch, every Python process during a code coverage run was
writing coverage data to the same file. I'm not sure if the coverage
package even tries to obtain a lock on the file. But what I do know is
there was some last write wins leading to loss of code coverage data, at
least with -j > 1.
This patch changes the code coverage mechanism to be multiple process
safe. The mechanism for initializing code coverage via sitecustomize.py
has been tweaked so each Python process will produce a separate coverage
data file on disk. Unless two processes generate the same random value,
there are no race conditions writing to the same file. At the end of the
test run, we combine all written files into an aggregate report.
On my machine, running the full test suite produces a little over
20,000 coverage files consuming ~350 MB. As you can imagine, it takes
several seconds to load and merge these coverage files. But when it is
done, you have an accurate picture of the aggregate code coverage for the
entire test suite, which is ~60% line coverage.
Before, we were invoking the "coverage" program provided by the
"coverage" module. This patch changes the code to go through the
Python API. This makes the next patch a little bit easier to reason
about.
A side effect of this patch is that writing code coverage reports
will be slightly faster, as we won't have to redundantly load
coverage data.
The only difference for the first two was to add the globs, but the third line
of output on Windows was '..\dir2\copy'. I'm not sure why 'copy' is output on
Windows instead of '*'.
The xunit writer was passing a str to a minidom API. An implicit
.decode('ascii') was performed somewhere, causing UnicodeDecodeError
if test output contained non-ascii sequences.
This patch converts test output to utf-8 before passing it to minidom.
We use the "replace" strategy to ensure invalid utf-8 sequences get
munged into �.
It seems better to leave the actual output in place instead of globbing
everything but 'abort:', in case it starts aborting for other reasons.
It isn't clear the purpose for reversing the file name position, but that
originates in windows.posixfile.
manifest.intersectfiles() is just a utility used by manifest.matches(), and
a future commit removes intersectfiles for treemanifest for optimization
purposes.
This commit makes the intersectfiles methods on manifestdict and treemanifest
internal, and converts its test to a more generic testMatches(), which has the
exact same coverage.
displayer doesn't want a fctx but a ctx. It failed with -Tdefault template.
Traceback (most recent call last):
...
File "mercurial/templatekw.py", line 212, in showbookmarks
bookmarks = args['ctx'].bookmarks()
AttributeError: 'filectx' object has no attribute 'bookmarks'
This patch makes wrapping "commands.update()" by largefiles extension
useless, because "cmdutil.bailifchanged()" can detect changes of
largefiles in the working directory.
This patch also changes test-update-branches.t, because
"cmdutil.bailifchanged()" shows more detailed information about
dirty-ness of the working directory than "workingctx.dirty()".
This patch newly adds "dirtyreason()" to centralize composing dirty
reason message like "uncommitted changes in subrepository 'xxxx'".
There are 3 similar messages below, and this patch is a part of
preparations for unifying them into (1), too.
1. uncommitted changes in subrepository 'XXXX'
2. uncommitted changes in subrepository XXXX
3. uncommitted changes in subrepo XXXX
This patch chooses adding new method "dirtyreason()" instead of making
"dirty()" return "reason string", because:
- some of existing "dirty()" implementation is too complicated to do
so simply, and
- ill-mannered 3rd party subrepo classes, of which "dirty()" doesn't
return "reason string", cause meaningless message (even though it
is rare case)
We have a UI bug where toggling a newly added file twice in the curses
interface didn't mark it as selected. This test checks that the underlying
logic is working as expected, the next patch of the series fixes the UI bug.
When m.text() is called after setting a nodeid with a suffix (such as
'+'), manifestdict uses the suffix-less nodeid for the text, while
treemanifest includes the suffix. It would perhaps make most sense to
raise an exception so the bug is found, but since the two
implementations behave differently, let's just not test the behavior
for now.
By rewriting test-manifest.py in terms of manifestdict instead of
_lazymanifest, the tests can be run on treemanifests too. There are
still a few tests that fail on treemanifests. They will be addressed
in the next few patches.
Reverting to a revision where the subrepo didn't exist will now abort, and
matching subrepos against the working directory is consistent with how filesets
are evaluated since dd1c701aad4d.
If self is a smartset and other is a fullreposet, nothing should be necessary.
A small win for trivial query in mozilla-central repo:
revset #0: (0:100000)
0) wall 0.017211 comb 0.020000 user 0.020000 sys 0.000000 (best of 163)
1) wall 0.001324 comb 0.000000 user 0.000000 sys 0.000000 (best of 2160)
Before this change, amending a merge would lose the rename information for file
adding in the second parents and implicitly renamed into a new directory.
In case of the merge, we also look for directory rename data from the second
parent. This seems to fix the bug and does not show other issues from the test
suite.
Even if largefiles extension is enabled in a repository, "repo"
object, which isn't "largefiles.reposetup()"-ed, is passed to
overridden functions in the cases below unexpectedly, because
extensions are enabled for each repositories strictly.
(1) clone without -U:
(2) pull with -U:
(3) pull with --rebase:
combination of "enabled@src", "disabled@dst" and
"not-required@src" cause this situation.
largefiles requirement
@src @dst @src result
-------- -------- --------------- --------------------
enabled disabled not-required aborted unexpectedly
required requirement error (intentional)
-------- -------- --------------- --------------------
enabled enabled * success
-------- -------- --------------- --------------------
disabled enabled * success (only for "pull")
-------- -------- --------------- --------------------
disabled disabled not-required success
required requirement error (intentional)
-------- -------- --------------- --------------------
(4) update/revert with a subrepo disabling largefiles
In these cases, overridden functions cause accessing to largefiles
specific fields of not "largefiles.reposetup()"-ed "repo" object, and
execution is aborted.
- (1), (2), (4) cause accessing to "_lfstatuswriters" in
"getstatuswriter()" invoked via "updatelfiles()"
- (3) causes accessing to "_lfcommithooks" in "overriderebase()"
For safe accessing to these fields, this patch examines whether passed
"repo" object is "largefiles.reposetup()"-ed or not before accessing
to them.
This patch chooses examining existence of newly introduced
"_largefilesenabled" instead of "_lfcommithooks" and
"_lfstatuswriters" directly, because the former is better name for the
generic "largefiles is enabled in this repo" mark than the latter.
In the future, all other overridden functions should avoid largefiles
specific processing for efficiency, and "_largefilesenabled" is better
also for such purpose.
BTW, "lfstatus" can't be used for such purpose, because some code
paths set it forcibly regardless of existence of it in specified
"repo" object.
The default joinfmt, "x.values()[0]", can't be used here because it picks
either 'bookmark' or 'current' randomly.
I got wrong result with PYTHONHASHSEED=1 on my amd64 machine.
My Mac test machine has 'echo' in '/opt/local/libexec/gnubin/echo' since, well,
GNU is not BSD.
Also, I feel it need to be said about using regexes:
Some people, when confronted with a problem, think "I know, I'll use regular
expressions." Now they have two problems.
Before this patch, revset predicate "tag()" and "named('tags')" differ
from each other, because the former doesn't include "tip" but the
latter does.
For equivalence, "named('tags')" shouldn't include the revision
corresponded to "tip". But just removing "tip" from the "tags"
namespace causes breaking backward compatibility, even though "tip"
itself is planned to be eliminated, as mentioned below.
http://selenic.com/pipermail/mercurial-devel/2015-February/066157.html
To mask specific names ("tip" in this case) for "named()" predicate,
this patch introduces "deprecated" into "namespaces", and makes
"named()" predicate examine whether each names are masked by the
namespace, to which they belong.
"named()" will really work correctly after 3.3.1 (see a3c326a7f57a for
detail), and fixing this on STABLE before 3.3.1 can prevent initial
users of "named()" from expecting "named('tags')" to include "tip".
It is reason why this patch is posted for STABLE, even though problem
itself isn't so serious.
This may have to be flagged as "(BC)", if applied on DEFAULT.
There are currently no tests for change/remove conflicts of subrepos,
and it's pretty broken. Add some tests demonstrating some of the
breakages and fix the most obvious one (a KeyError when trying to look
up a subrepo in the wrong context).
Before this patch, revset predicate "named()" uses each nodes gotten
from target namespaces directly.
This causes problems below:
- combination of other predicates doesn't work correctly, because
they assume that revisions are listed up in number
- "hg log" doesn't show any revisions for "named()" result, because:
- "changeset_printer" stores formatted output for each revisions
into dict with revision number (= ctx.rev()) as a key of them
- "changeset_printer.flush(rev)" writes stored output for
the specified revision, but
- "commands.log" invokes it with the node, gotten from "named()"
- "hg debugrevspec" shows nodes (= may be binary) directly
Difference between revset predicate "tag()" and "named('tags')" in
tests is fixed in subsequent patch.
Before this patch, "bookmark()", "named()" and "tag()" predicates
raise "Abort", when the specified pattern doesn't match against
existing ones.
This prevents "present()" predicate from continuing the query, because
it only catches "RepoLookupError".
This patch raises "RepoLookupError" instead of "Abort", to make
"present()" predicate continue the query, even if "bookmark()",
"named()" or "tag()" in the sub-query of it are aborted.
This patch doesn't contain raising "RepoLookupError" for "re:" pattern
in "tag()", because "tag()" treats it differently from others. Actions
of each predicates at failure of pattern matching can be summarized as
below:
predicate "literal:" "re:"
---------- ----------- ------------
bookmark abort abort
named abort abort
tag abort continue (*1)
branch abort continue (*2)
---------- ----------- ------------
"tag()" may have to abort in the (*1) case for similarity, but this
change may break backward compatibility of existing revset queries. It
seems to have to be changed on "default" branch (with "BC" ?).
On the other hand, (*2) seems to be reasonable, even though it breaks
similarity, because "branch()" in this case doesn't check exact
existence of branches, but does pick up revisions of which branch
matches against the pattern.
This patch also adds tests for "branch()" to clarify behavior around
"present()" of similar predicates, even though this patch doesn't
change "branch()".
The previous code was adding standin files to the matcher's file list when
neither the standin file nor the original existed in the context. Somehow, this
was confusing the logging code into behaving differently from when the extension
wasn't loaded.
It seems that this was an attempt to support naming a directory that only
contains largefiles, as a test fails if the else clause is dropped entirely.
Therefore, only append the "standin" if it is a directory. This was found by
running the test suite with --config extensions.largefiles=.
The first added test used to log an additional cset that wasn't logged normally.
The only relation it had to file 'a' is that 'a' was the source of a move, but
it isn't clear why having '.hglf/a' in the list causes this change:
@@ -47,6 +47,11 @@
Make sure largefiles doesn't interfere with logging a regular file
$ hg log a --config extensions.largefiles=
+ changeset: 3:2ca5ba701980
+ user: test
+ date: Thu Jan 01 00:00:04 1970 +0000
+ summary: d
+
changeset: 0:9161b9aeaf16
user: test
date: Thu Jan 01 00:00:01 1970 +0000
The second added test used to complain about a file not being in the parent
revision:
@@ -1638,10 +1643,8 @@
Ensure that largefiles doesn't intefere with following a normal file
$ hg --config extensions.largefiles= log -f d -T '{desc}' -G
- @ c
- |
- o a
-
+ abort: cannot follow file not in parent revision: ".hglf/d"
+ [255]
$ hg log -f d/a -T '{desc}' -G
@ c
|
Note that there is still something fishy with the largefiles code, because when
using a glob pattern like this:
$ hg log 'glob:sub/*'
the pattern list would contain '.hglf/glob:sub/*'. None of the tests show this
(this test lives in test-largefiles.t at 1349), it was just something that I
noticed when the code was loaded up with print statements.
The conditional was a bit too narrow and produced buggy result when a node was
present in both common and heads (because it pleased the discovery) and it was
locally known but filtered.
This resulted in buggy getbundle request and server side crash.
Before this patch, failure of updating subrepos may cause inconsistent
".hgsubstate". For example:
1. dirstate entry for ".hgsubstate" of the parent repo is filled
with valid size/date (via "hg state" or so)
2. "hg update" is invoked at the parent repo
3. ".hgsubstate" of the parent repo is updated on the filesystem as
a part of "g"(et) action in "merge.applyupdates"
4. it is assumed that size/date of ".hgsubstate" on the filesystem
aren't changed from ones at (1)
this is not so difficult condition, because just changing hash
ids (every ids are same in length) in ".hgsubstate" doesn't
change the file size of it
5. "subrepo.submerge()" is invoked to update subrepos
6. failure of updating in one of subrepos raises exception
(e.g. "untracked file differs")
7. "hg update" is aborted without updating dirstate of the parent repo
dirstate entry for ".hgsubstate" still holds size/date at (1)
Then, ".hgsubstate" of the parent repo is treated as "CLEAN"
unexpectedly, because updating ".hgsubstate" at (3) doesn't change
size/date of it on the filesystem: see assumption at (4).
This inconsistent ".hgsubstate" status causes unexpected behavior, for
example:
- "hg revert" forgets to revert ".hgsubstate"
- "hg update" misunderstands that (not yet updated) subrepos diverge
(then, it shows the prompt to confirm user's decision)
To avoid inconsistent ".hgsubstate" status above, this patch marks
".hgsubstate" as possibly dirty before "submerge" invocation.
"normallookup"-ed (= dirty) dirstate should be written out, even if
processing is aborted by failure.
This patch marks ".hgsubstate" as possibly dirty before "submerge",
also when it is removed or merged while merging, for safety. This
should prevent Mercurial from misunderstanding inconsistent
".hgsubstate" as clean.
To satisfy conditions at (1) and (4) above, this patch uses "hg status
--config debug.dirstate.delaywrite=2" (to fill valid size/date into
dirstate) and "touch" (to fix date of the file).
Before this changeset rebase was getting very confused if any revision in the
rebase set became hidden. This was fairly easy to achieve if a rebased revision
was made visible by the working copy location. The rebase process would update
somewhere else and the revision would become hidden.
To work around this issue, we ensure rebased revisions remain visible for the
whole process.
This is a simple change suitable for stable. More subtle usage of unfiltered
repository in rebase may solve this issue more cleanly.
b8552020f458 broke things ... and the following cleanups didn't fix all issues.
It didn't work with the diffargs shipped in mergetools.rc with explicit
quoting. Parameters would end up with being quoted twice - especially if they
really needed quoting.
To work around that, look for explicit quotes around the variables that will be
substituted with proper quoting. Also accept an additional prefix so we can
handle both
--foo='$parent'
and
'--foo=$parent'
It will however still fail if the user intentionally place the variable inside
a quoted string, as in
'parent $parent is on the left'
There is currently no good way to handle that, short of knowing exactly which
quoting mechanism will be used.
This effectively reverts 9fc565fa1621, which caused some normal file copies to
not be displayed as copies. Other normal file copies could be displayed- the
exact reason isn't clear. This also adds two tests that were failing prior to
this backout, so that this can be sorted out next cycle.
The difference between copy cases that worked and those that didn't seemed to be
in copies.pathcopies(). When largefiles isn't enabled for the changed test, or
lfstatus is not set in the commands.status() override, 'y.ancestor(x) == x'.
That wasn't true otherwise, which fell through to the _chain() method. In this
case, the copy is removed in the criss cross loop.
'y.ancestor(x)' returns a context.changectx type, while 'x' is a lfilesctx type
in the failing case. I tried adding the ancestor method to the lfilesctx class
to change the type of the ancestor context, however the context when printed as
a string then gains a '+'. This points to it being a context.committablectx,
which clearly isn't correct for an ancestor. Possibly the problem is the
lfilesctx needs to subclass context.committablectx in some cases, but
context.changectx in others, within the same invocation? I'm not sure how to
pull that off, and backing out this change is safer during the freeze.
As to the status changing when a path is specified, I haven't looked into it
yet.
This can simplify the conversion from numeric revision to string. Without it,
we have to handle -1 specially because repo['-1'] != repo[-1].
The -1 revision is not officially documented, but this change makes sense
assuming that "rev(%d)" exists for scripting or third-party tools.
Before this patch, test-tag.t can't run successfully on Windows,
because:
- quoted hg command ('"hg"') prevents "hg.bat" from working correctly
(only at testing with pure Python build)
"%~f0" and "%~dp0hg" in "hg.bat" cause unexpected result in this
case. BTW, quoted "\path\to\hg" works correctly.
- "`pwd`" in the command line is expanded unexpectedly
not "C:\path\to\TESTTMP" but "C;C:\path\to\TESTTMP"
When running "hg log 'set:added()'", we create two matchers: one used
for producing the revset and one used for finding files to match. In
185b6b930e8c (graphlog: evaluate FILE/-I/-X filesets on the working
dir, 2012-02-26), we started passing a revision argument along from
what's currently in cmdutil._makelogrevset() to
revset._matchfiles(). When the revision was an empty string, it
referred to the working copy. This was subtly done with "repo[rev or
None]". Then, in 5ff5c5c9e69f (revset: avoid recalculating filesets,
2014-10-22), that conversion from empty string to None was lost. Note
that repo[''] is equivalent to repo['.'], not repo[None].
The consequence of this, to the user, is that when running "hg log
'set:added()'", the file matcher matches files added in the working
copy, while the revset matcher matches revisions that touch files
added in the parent of the working copy. As a result, only revisions
that touch any files added in the parent of the working copy will be
considered, but they will only be included if they also touch files
added in the working copy.
Fix the bug by converting '' to None again, but make it a little more
explicit this time (plus, we now have tests for it).
On IRC, rom1dep reported a traceback[1] from setting
experimental.strip-bundle2-version to True. This diff catches
unexpected values and falls back to the non-experimental bundle1
implementation after issuing a warning.
[1] http://gist.tamytro.org/_admin/gists/qXcdQLwtApgy6e3NwWgl
Previously, git subrepo diffs did not have a newline at the end.
This caused multiple subrepo diffs to be joined on the same line.
Additionally, the command prompt after the diff still contained
a part of the diff.
Before this patch, "test-run-tests.t" doesn't test "run-tests.py" with
"--pure", even if outer "run-tests.py" is executed with it.
This patch uses not "HG_RUN_TESTS_PURE" but "HGTEST_RUN_TESTS_PURE",
because "HG_" prefixed environments are forcibly dropped in "_getenv()".
This is also useful to run "run-tests.py" successfully by
"run-tests.py --pure" on Windows without any compilation tools (like
VisualStudio).
When "hg.bat" is invoked via interactive shell "cmd.exe" on Windows,
it can store own exit code into ERRORLEVEL correctly, regardless of
explicit "exit" statement in it: "cmd.exe" seems to hold ERRORLEVEL
updated by the last command in the batch file (= "python hg", in
"hg.bat" case).
On the other hand, "hg.bat" is invoked indirectly via
"subprocess.Popen" (e.g. shell alias, hooks, hgclient and so on), the
parent process always receives exit code 0 from spawned "hg.bat":
batch files on Windows seem not to be really spawned like as shell
scripts on UNIX, but to be executed in the "cmd.exe" process.
This patch returns exit code explicitly for indirect invocation.
"/b" should be specified for "exit" to prevent "cmd.exe" from being
terminated when "hg.bat" is invoked interactively from it.
Before this patch, "run-tests.py" executes "hghave" process without
any modifications for environment variables, even though actual tests
are executed with LC_ALL, LANG and LANGUAGE explicitly assigned "C".
When "run-tests.py" is executed:
- with non-"C" locale environment variables on any platforms, or
- without any explicit locale environment setting on Windows
(only for "outer-repo" feature using "hg root")
external commands indirectly executed by "hghave" may show translated
messages.
This causes incorrect "hghave" result and skipping tests, because some
regexp matching of "hghave" expect external commands to show
un-translated messages.
To prevent external commands from showing translated messages, this
patch makes "run-tests.py" execute "hghave" with same environment
variables as ones for actual tests.
This patch doesn't make "hghave" execute external commands forcibly
with LC_ALL, LANG and LANGUAGE explicitly assigned "C", because
changing "run-tests.py" is cheaper than changing "hghave":
- "os.popen" should be replaced by "subprocess.Popen" or so, and
- setting up environment variables should be newly added
Previously, we had weird, nonsensical behavior when committing a file move
with a missing source. This removes that weird logic and tests that the bug
this strange behavior caused is fixed. Also adds a longish comment to prevent
some poor soul from accidentally re-implementing the bug in the future.
When a binary source has been copied or renamed into a non-binary
file, we don't check whether the copy source was binary. There is a
code comment explaining that a git mode will be forced anyway in order
to capture the copy record (i.e. losedatafn() will be called). This is
true, but forcing git mode is not the only effect binary files have:
when git mode was already requested, we use the binary-ness to tell us
whether to use a regular unified diff or a git binary diff. The user
sees this as a "Binary file $file has changed" instead of the binary
When a directory is named in the commit file list, the previous behavior was to
walk the list, and if no normal files in the directory were also named, add the
corresponding standin for each largefile in that directory. The directory is
then dropped from the list, so that committing a directory with no normal file
changes works. It then added the corresponding standin directory for the first
largefile seen, by prefixing it with '.hglf/'.
The latter is unnecessary since each affected largefile is explicitly referenced
by its standin in the list. It also caused an abort if there were no changed
largefiles in the directory, because none of its standins changed:
abort: .hglf/foo/bar: no match under directory!
This list of files is used to tweak a matcher in lfutil.updatestandinsbymatch(),
which is what is passed to commit().
The status() call that is ultimately done in the commit code with this matcher
seems to have some OS specific differences. It is not necessary to append '.'
for Windows to run the largefiles tests cleanly. But if '.' is not added to the
list, the match function isn't called on Linux, so status() would miss any
normal files that were also in a named directory. The commit then proceeds
without those normal files, or says "nothing changed" if there were no changed
largefiles in the directory. This is not filesystem specific, as VFAT on Linux
had the same behavior as when run on ext4. It is also not an issue with
lfilesrepo.status(), since that only calls the overridden implementation when
paths are passed to commit. I dont have access to an OS X machine ATM to test
there.
Maybe there's a better way to do this. But since the standin directory for the
first largefile was previously being added, and that caused the same walk in
status(), there's no preformance change to this. There is no danger of
erroneously committing files in '.', because the original match function is
called, and if it fails, the lfutil.updatestandinsbymatch() tweaked matcher only
indicates a match if the file is in the list of standins- and '.' never is. The
added tests confirm this.
Once the transaction is closed, we now write transaction related data for
possible future undo. For now, we only do it for full file "backup" because
their were not handle at all in that case. In the future, we could move all the
current logic to set undo up (that currently exists in localrepository) inside
transaction itself, but it is not strictly requires to solve the current
situation.
This adds an experimental option 'strip-bundle2-version' which causes backup
bundles to use bundle2 formatting. Especially for generaldelta repositories,
this should provide significant performance gains for any operation that needs
to write a backup.
This diff adds support to writebundle to generate a bundle2 wrapper; upcoming
diffs will add an option to write a v2 changegroup part instead of v1 in these
bundles.
The --large, --normal and --lfsize args couldn't be passed to a subrepo before,
and files in the subrepos would be added silently (if -v wasn't specified) as
normal files. As an added bonus, 'hg add --dry-run' no longer prints that
largefiles would also be added as normal files as well.
Extensions can declare compatibility with Mercurial versions. If an
error occurs, Mercurial will attempt to pin blame on an extension that
isn't marked as compatible.
While all bets are off when it comes to the internal API, my experience
has shown that a monthly/patch release of Mercurial has never broken any
of the extensions I've written. I think that expecting extensions to
declare compatibility with every patch release of Mercurial is asking a
bit much and adds little to no value.
This patch changes the blame logic from exact version matching to only
match on the major and minor Mercurial versions. This means that
extensions only need to mark themselves as compatible with the major,
quarterly releases and not the monthly ones in order to stay current and
avoid what is almost certainly unfair blame. This will mean less work
for extension authors and almost certainly fewer false positives in the
blame attribution.
I still get the following diff on Windows 7 with NTFS. I'm not sure if it is
expected and we should wildcard the link count, or if something needs to be
fixed.
@@ -58,7 +58,7 @@
$ ln bar baz
$ f bar -n baz -l --hexdump -t --sha1 --lines=9 -B 20
- bar: file, links=2, newer than baz, sha1=612ca68d0305c821750a
+ bar: file, links=0, newer than baz, sha1=612ca68d0305c821750a\r (esc)
0000: 31 0a 32 0a 33 0a 34 0a 35 0a 36 0a 37 0a 38 0a |1.2.3.4.5.6.7.8.|
0010: 39 0a |9.|
When the manifest revision is stored as a delta against a non-parent revision,
'_adjustlinkrev' could miss some file update because it was using the delta
only. We now use the 'fastread' method that uses the delta only when it makes
sense.
A test showcasing on the of possible issue have been added.
The cache files are usually append only but will automatically be truncated and
recover in exceptional situations. Add a debug notice when such exceptional
situations are encountered.
The tests that were added with the revision branch cache in 6aa81b0c4658 had
suffered from bit rot in the development iterations. They were no longer that
"good".
Now, the tests are rewritten and reworked to be more readable and maintainable
and relevant for the actual implementation.
This also utilizes the new 'f' helper tool for keeping an eye the cache files.
This tool is like the collection of tools found in a unix environment but are
cross platform and stable and suitable for our needs in the test suite.
The main reason it is "needed" now is for hexdump of revision branch cache to
keep an eye on how it changes and make sure the format is stable.
It is a very generic tool that can end up being used a lot in tests, so I gave
it very generic name.
The testdir is already added to $PYTHONPATH - I think it makes sense and is
convenient to add it to $PATH too.
The following binaries are invoked from tests using full path with $TESTDIR/ -
they can now be used without specifying path:
dumbhttp.py
dummyssh
filterpyflakes.py
generate-working-copy-states.py
get-with-headers.py
hghave
histedit-helpers.sh
killdaemons.py
md5sum.py
notcapable
printenv.py
readlink.py
revlog-formatv0.py
run-tests.py
svn-safe-append.py
svnxml.py
tinyproxy.py
Before this patch, alias declaration is parsed by string base
operations: matching against "^([^(]+)\(([^)]+)\)$" and splitting by
",".
This overlooks many syntax errors like below (see the previous patch
introducing "_parsealiasdecl" for detail):
- un-closed parenthesis causes being treated as "alias symbol"
- symbol/function name aren't examined whether they are valid or not
- invalid argument list causes unexpected argument names
To parse alias declaration strictly, this patch replaces parsing
implementation by "_parsealiasdecl".
This patch tests only one typical declaration error case, because
error detection itself is already tested in the doctest of
"_parsealiasdecl".
This also removes class property "args" and "error", because these are
certainly initialized in "revsetalias.__init__".
Before this patch, any errors in the declaration of revset alias
aren't detected at all, and there is no information about error source
in the error message.
As a part of preparation for parsing alias declarations and
definitions more strictly, this patch stores full detail into
"revsetalias.error" for error source distinction.
This makes raising "Abort" and warning potential errors just use
"revsetalias.error" without any message composing.
'hg update' would hash all 'unsure' largefiles before performing the merge. It
would update the standins but not detect the very common case where the
largefile never had been changed by the user but just had been marked with an
invalid dirstate mtime to make sure any changes done by the user in the same
second would be detected. The largefile would remain in that state and would
have to be hashed again next time even though it still not had been changed.
Sad trombone.
Instead, for largefiles listed as 'unsure' or 'modified', after updating the
standin with the actual hash, mark the largefile as normal if it turns out to
not be modified relative to the revision in the parent revision. That will
prevent it from being hashed again next time.
The value of the dirstate date field cannot be used in tests and we thus have
to use debugdirstate with --nodate. It is however still very helpful to be able
to see whether the date field has been set or still is unset. The absence of
that information made it hard to debug some largefile dirstate issues.
This change _could_ make the test suite more unstable ... but that would be
places where the test suite or the code should be made more stable. (Note:
'unset' with the magic negative sizes is reliable. 'unset' for normal sizes
would probably not be reliable, but there is no such occurrences in the test
suite and it should thus be reliable.)
This output wastes more horizontal space in the --nodate output, but it also
makes things simpler that the output format always is the same. It is just a
debug command so let's keep it simple.
Previously, a backup bundle could overwrite an existing bundle and cause user
data loss. For instance, if you have A<-B<-C and strip B, it produces backup
bundle B-backup.hg. If you then hg pull -r B B-backup.hg and strip it again, it
overwrites the existing B-backup.hg and C is lost.
The fix is to add a hash of all the nodes inside that bundle to the filename.
Fixed up existing tests and added a new test in test-strip.t
Because spanset.isascending() ignored the ascending flag, the result of
"fullreposet() & x" was always sorted in ascending order.
The test case is carefully chosen to call fullreposet.__and__.
'<a .../>foo</a>' syntax is incorrect, since the first tag just "tries" to
close itself and then the actual content follows. It doesn't work, either
because web browsers know better than this or because there should be a
whitespace before /: '<a />'. So for the hgweb users the links looked
normal anyway, but now they are correct in code as well.
<p> elements can only contain inline elements, so as soon as browser encounters
a block element (e.g. block <div>) "inside" a <p>, it puts an implicit </p>.
It's better to do this explicitly.
Python 2.7.9 cleans up how it stringifies SSL errors, so we have to look only
for the important bit (certificate verify failed) rather than looking for
specific ssl module goop (which is now unstable).
As explained in a previous changeset, prioritizing heads too much behaves
pathologically when there are more heads than the sample size. To counter this,
we always inject exponential samples before reducing to the sample size limit.
This already show some benefit in the test themselves, but on a real-world example
this moves my discovery for push to pathologically headed repo from 45 rounds to
17 of them.
We should maybe ensure that at least 25% of the result sample is heads, but I
think the random sampling will be fine in practice.
The heads and exponential sample are going to end up in the same set
before any extra processing happens. We simplify the code by directly
updating a set with heads.
Changes in the order the set is built lead to small changes in the random
sampling output. But after double checking, I can confirm the input data to
the random sampling is consistent.
Before this changeset, the discovery protocol was too heads-centric. Heads of the
undiscovered set were always sent for discovery and any room remaining in the
sample were filled with exponential samples (and random ones if any room
remained).
This behaved extremely poorly when the number of heads exceeded the sample size,
because we keep just asking about the existence of heads, then their direct parent
and so on. As a result, the 'O(log(len(repo)))' discovery turns into a
'O(len(repo))' one. As a solution we take a random sample of the heads plus
exponential samples. This way we ensure some exponential sampling is achieved,
bringing back some logarithmic convergence of the discovery again.
This patch only applies this principle in one place. More places will be updated
in future patches.
One test is impacted because the random sample happen to be different. By
chance, it helps a bit in this case.
Branch name filtering in revsets was expensive. For every rev it created a
changectx and called .branch() which retrieved the branch name from the
changelog.
Instead, use the revbranchcache.
The revbranchcache is used read-only. The revset implementation with generators
and callbacks makes it hard to figure out when we are done using/updating the
cache and could write it back. It would also be 'tricky' to lock the repo for
writing from within a revset execution. Finally, the branchmap update will
usually make sure that the cache is updated before any revset can be run.
The revbranchcache is used without any locking but is short-lived and used in a
tight loop where we can assume that the changelog doesn't change ... or where
it not is relevant to us if it does.
perfrevset 'branch(mobile)' on mozilla-central.
Before:
! wall 10.989637 comb 10.970000 user 10.940000 sys 0.030000 (best of 3)
After, no cache:
! wall 7.368656 comb 7.370000 user 7.360000 sys 0.010000 (best of 3)
After, with cache:
! wall 0.528098 comb 0.530000 user 0.530000 sys 0.000000 (best of 18)
The performance improvement even without cache come from being based on
branchinfo on the changelog instead of using ctx.branch().
Some tests are added to verify that the revbranchcache works and keep an eye on
when the cache files actually are updated.
The revbranchcache is read on demand before it will be used for updating the
branch map. It is written back when the branchmap is written and it will thus
use the same locking as branchmap. The revbranchcache instance is short-lived;
it is only stored in the branchmap from .update() is invoked and until .write()
is invoked. Branchmap already assume that the repo is locked in that case.
The use of revbranchcache for branch map updates will make sure that the
revbranchcache "always" is kept up-to-date.
The perfbranchmap benchmark is somewhat bogus, especially when we can see that
the caching makes a significant difference between the realistic case of a
first run and the rare case of rerunning it with a full cache. Here are some
'base' numbers on mozilla-central:
Before:
! wall 6.912745 comb 6.910000 user 6.840000 sys 0.070000 (best of 3)
After - initial, cache is empty:
! wall 7.792569 comb 7.790000 user 7.720000 sys 0.070000 (best of 3)
After - cache is full:
! wall 0.879688 comb 0.880000 user 0.870000 sys 0.010000 (best of 4)
The overhead when running with empty cache comes from checking, missing and
updating it every time.
Most of the performance improvement comes from not having to extract the branch
info from the changelog. The last doubling of performance comes from no longer
having to convert all branch names to local encoding but reuse the few already
converted branch names.
On the hg repo:
Before:
! wall 0.715703 comb 0.710000 user 0.710000 sys 0.000000 (best of 14)
After:
! wall 0.105489 comb 0.110000 user 0.110000 sys 0.000000 (best of 87)
jsonchangeset._show() was computing the reverse status of the current
changeset. As a result, added files were showing up as removed and
removed files were showing up as adds.
There were existing tests for this code and they were flat out wrong.
The more aggressive synchronization of lfdirstate that was backed out in
4fe80f20ab06 masked the problem where lfdirstate would hold an 'R' for a
largefile that was added and then removed without a commit between. We could
just conditionally call lfdirstate.drop() or lfdirstate.remove() here, but this
also properly updates lfdirstate if the standin doesn't exist for the file
somehow (i.e. call drop instead of remove).
Without this change, the precommit status in the commit command immediately
after the test change lists the removed (and never committed) largefile as 'R'.
It can also lead to situations where the status command reports the same, long
after the commit [1].
[1] http://www.selenic.com/pipermail/mercurial-devel/2015-January/065153.html
d20af8be6a14 introduced a significant performance regression: All largefiles
were marked 'normallookup' by linear (or noop) updates and had to be rehashed
by the next command.
The previous change introduced a different solution to the problem d20af8be6a14
solved and we can thus back it out again.
d20af8be6a14 introduced a significant performance regression: All largefiles
are marked 'normallookup' in lfdirstate by linear (or noop) updates and has to
be rehashed by the next command.
To avoid such regressions, keep an eye on the dirstate content after a plain
'hg up'.
If resolving a merge conflict result in an empty changesets, we now properly
skip the changeset instead of crashing.
Original patch from Robert Collins <robertc@robertcollins.net>.
This module depends on _winreg, which is windows-only. Recent versions
of setuptools load distutils.msvc9compiler and expect it to
ImportError immediately when on non-Windows platforms, so we need to
let them do that. This breaks in an especially mystifying way, because
setuptools uses vars() on the imported module. We then throw an
exception, which vars doesn't pick up on well. For example:
In [3]: class wat(object):
...: @property
...: def __dict__(self):
...: assert False
...:
In [4]: vars(wat())
---------------------------------------------------------------------------
TypeError Traceback (most recent call last)
<ipython-input-4-2781ada5ffe6> in <module>()
----> 1 vars(wat())
TypeError: vars() argument must have __dict__ attribute
Which is similar to the problem we run into.
If an uncommitted and deleted file was forgotten, a warning would be emitted,
even though the operation was successful. See the previous patch for
'remove -A' for the exact circumstances, and details about the cause.
The bug report doesn't mention largefiles, but the given recipe doesn't fail
unless the largefiles extension is loaded. The problem only affected normal
files, whether or not any largefiles are committed, and only files that have
not been committed yet. (Files with an 'a' state are dropped from dirstate,
not marked removed.) Further, if the named normal file never existed, the
warning would be printed out twice.
The problem is that the core implementation of remove() calls repo.status(),
which eventually triggers a dirstate.walk(). When the file isn't seen in the
filesystem during the walk, the exception handling finds the file in
dirstate, so it doesn't complain. However, the largefiles implementation
called status() again with all of the original files (including the normal
ones, just dropped). This time, the exception handler doesn't find the file
in dirstate and does complain. This simply excludes the normal files from
the second repo.status() call, which the largefiles extension has no interest
is processing anyway.
The log/graphlog revset was not producing stable results since it was
iterating over a dict. Now we sort before iterating to guarantee a fixed order.
This fixes some potential flakiness in the tests.
The revset created when -f was used with a slow path (for patterns and
directories) did not actually contain any logic to enforce follow. Instead it
was depending on the passed in subset to already be limited (which was limited
to :. but not ::.). This fixes it by adding a '& ::.' to any -f log revset.
hg log -f <file> is still broken, in that it can return results that aren't
actually ancestors of the current file, but fixing that has major perf
implications, so we'll deal with it later.
When looking for untracked files that would conflict with a tracked
file in the target revision (or the remote side of a merge), we
explcitly exclude ignored files. The code was added in f1db75422e70
(merge: refactor unknown file conflict checking, 2012-02-09), but it
seems like only unknown, not ignored, files were considered since the
beginning of time.
Although ignored files are mostly build outputs and backup files, we
should still not overwrite them. Fix by simply removing the explicit
check.
We don't seem to have any tests for updating to another revision when
there are untracked files on the local side that conflict with the
those on the remote side, so let's add tests. This shows how we
overwrite untracked ignored files when updating to a revision that
tracks the file.
The state mapping also contains some magic negative values (detached
parent, ignored revision). Blindly reading the state thus lead to
unfortunate usage of the negative value as an update destination. We
now filter them out.
We do a minor alteration of the test to catch this.
Before, merging would in some cases ask "wrong" questions about
"changed/deleted" conflicts ... and even do it before the resolve phase where
they can be postponed, re"resolved" or answered in bulk operations.
Instead, check that the content of the changed file really did change.
Reading and comparing file content is expensive and should be avoided before
the resolve phase. Prompting the user is however even more expensive. Checking
the content here is thus better.
The 'f in ancestors[0]' should not be necessary but is included to be extra
safe.
Use suffix -same for cases where file changed but content is the same - that is
the case where manifestmerge doesn't detect that a file is unchanged.
(The suffix -id is already used for cases where the file didn't change - that
is the trivial case where manifestmerge detects that the file is unchanged.)
These new tests are good but the results are bad. There shouldn't be any merge
conflicts or prompts when one side didn't change.
This allow to gracefully report the failure of the bookmark push and carry on.
Before this change set. Local push would plain quit and wireprotocol would
failed in various ungraceful way.
This patch fixes a bug where hgweb would send an incomplete HTTP
response.
If an uncaught exception is raised when hgweb is processing a request,
hgweb attempts to send a generic error response and log that exception.
The server defaults to chunked transfer coding. If an uncaught exception
occurred, it was sending the error response string / chunk properly.
However, RFC 7230 Section 4.1 mandates a 0 size last chunk be sent to
indicate end of the entity body. hgweb was failing to send this last
chunk. As a result, properly written HTTP clients would assume more data
was coming and they would likely time out waiting for another chunk to
arrive.
Mercurial's own test harness was paving over the improper HTTP behavior
by not attempting to read the response body if the status code was 500.
This incorrect workaround was added in faced8f5c2af and has been removed
with this patch.
The result of 'hg rm' + 'hg rename' disagreed with the one from
'hg rename --force'. We align them on 'hg move --force' because it agrees with
what 'hg status' says after the commit.
Stopping reporting a modified file as added puts an end to the hg revert confusion in this
situation (issue4458).
However, reporting the file as modified also prevents revert from restoring the copy
source. We fix this in a later changeset.
Git diff also stop reporting the add in the middle of the chain as add. Not
sure how important (and even wrong) it is.
Because the same dictionary was used to (1) get node from parent and (2) store
annotated version, we could end up with buggy values. For example with a chain
of renames:
$ hg mv b c
$ hg mv a b
The value from 'b' would be updated as "<old-a>a", then the value of c would be
updated as "<old-b>a'. With the current dictionary sharing this ends up with:
'<new-c>' == '<old-a>aa'
This value is double-wrong as we should use '<old-b>' and a single 'a'.
We now use a read-only value for lookup. The 'test-rename.t' test is impacted
because such a chained added file is suddenly detected as such.
Changeset 5b64e22ecd8e introduced the examination of exec bit of
largefiles in "hg status --rev REV" case, but it doesn't avoid it on
the platform being unaware of exec-bit (e.g. on NTFS of Windows).
Previously, if reorder was true during the creation of a changegroup bundle,
it was possible that the manifest and filelogs would be reordered such that the
resulting bundle filelog had a linkrev that pointed to a commit that was not
the earliest instance of the filelog revision. For example:
With commits:
0<-1<---3<-4
\ /
--2<---
if 2 and 3 added the same version of a file, if the manifests of 2 and 3 have
their order reversed, but the changelog did not, it could produce a filelog with
linkrevs 0<-3 instead of 0<-2, which meant if commit 3 was stripped, it would
delete that file data from the repository and commit 2 would be corrupt (as
would any future pulls that tried to build upon that version of the file).
The fix is to make the linkrev fixup smarter. Previously it considered the first
manifest that added a file to be the first commit that added that file, which is
not true. Now, for every file revision we add to the bundle we make sure we
attach it to the earliest applicable linkrev.
"/search", which is an invalid command in hgweb, was mistakenly used for
"[show] more [revsets]" and "[show] less [revsets]" links on search page in
templates "paper" (and those which inherit paper, such as coal) before and
worked fine until 73c8d0c02c22, which made hgweb more strict about invalid
commands.
Fix inconsistent handling of plain header separation in mq patcheader - and
contrary to 05acc6157816, do it in the direction of having an empty line
between header and description. Plain patches are like mails and should thus
have an empty line between headers and body in compliance with RFC 822 3.1.
The new rebase revset didn't check for the case when there are no common
ancestors. Now it does. The new behavior should be the same as the pre-3.2
behavior. Added a test.
simplejson produces slightly different output from the built-in json
module, specifically:
* It uses 0.000 instead of 0.0000
* It likes to put a trailing space after a comma
This change works around both of those variations.
A matcher is required when enabling the subrepo option on archival.archive(),
because that calls match.narrowmatcher(), which accesses fields on the object.
It's therefore probably a bad idea to default the matcher to None on archive(),
but that's a fix for default.
This keyword remapping was introduced in 236440938a03 as part of converting
generator based iterators into list based iterators, mentioning "undesired
behavior in template" when a generator is exhausted, but doesn't say what and
introduces no tests.
The problem with the remapping was that it corrupted the output for keywords
like 'extras', 'file_copies' and 'file_copies_switch' in templates such as:
$ hg log -r 82a4f5557c6b --template "{file_copies % ' File: {file_copy}\n'}"
File: mercurial/changelog.py (mercurial/hg.py)
File: mercurial/changelog.py (mercurial/hg.py)
File: mercurial/changelog.py (mercurial/hg.py)
File: mercurial/changelog.py (mercurial/hg.py)
File: mercurial/changelog.py (mercurial/hg.py)
File: mercurial/changelog.py (mercurial/hg.py)
File: mercurial/changelog.py (mercurial/hg.py)
File: mercurial/changelog.py (mercurial/hg.py)
What was happening was that in the first call to runtemplate() inside runmap(),
'lm' mapped the keyword (e.g. file_copies) to the appropriate showxxx() method.
On each subsequent call to runtemplate() in that loop however, the keyword was
mapped to a list of the first item's pieces, e.g.:
'file_copy': ['mercurial/changelog.py', ' (', 'mercurial/hg.py', ')']
Therefore, the dict for the second and any subsequent items were not processed
through the corresponding showxxx() method, and the first item's data was
reused.
The 'extras' keyword regressed in 56b014c52204, and 'file_copies' regressed in
4e182fb53989 for other reasons. The common thread of things fixed by this seems
to be when a list of dicts are passed to the templatekw._hybrid class.
This returns the csets where matching subrepos have changed with respect to the
containing repo's first parent. The second parent shouldn't matter, because it
is either syncing up to the first parent (i.e. it hasn't changed from the
current branch's POV), or the merge changed it with respect to the first parent
(which already adds it to the set).
There's already a 'subrepo' fileset, but it is prefixed with 'set:', so there
should be no ambiguity (in code anyway). The only test I see for it is to
revert subrepos named by a glob pattern (in test-subrepo.t, line 58). Since it
doesn't return a tracked file, neither 'log "set:subrepo()"' nor
'files "set:subrepo()"' print anything. Therefore, it seems useful to have a
revset that will return something for log (and can be added to a revsetalias to
be chained with 'file' revsets.)
It might be nice to be able to filter for added, modified and removed
separately, but add/remove should be rare. It might also be nice to be able to
do a 'contains' check, in addition to this mutated check. Maybe it is possible
to get those with the existing 'adds', 'contains', 'modifies' and 'removes' by
teaching them to chase explicit paths into subrepos.
I'm not sure if this should be added to the 'modifies adds removes' line in
revset.optimize() (since it is doing an AMR check on .hgsubstate), or if it is
OK to put into 'safesymbols' (things like 'file' are on the list, and that takes
a regex, among other patterns).
As the failing revert tests in test-fileset-generated.t show,
Revert currently creates one matcher for matching files in the working
copy and another matcher for matching files in the target
revision. The two matchers are created with different contexts, which
means filesets are evaluated differently. Then the union of the sets
of files matching the matchers in the two contexts are reverted. It
doesn't seem to make sense to use two different matchers; only the
context they're applied to should be different.
It seems very likely that the user wants the filesets to be evaluated
against the working directory, which the tests
test-fileset-generated.t also assume, so let's make it so.
I willingly admit that the largefiles code was modified by trial and
error (according to tests).
The seq.py and shell echo redirected to a file ends up with platform
specific EOL, which throws off the hash when the file is committed to
the test repo on Windows. The other option is to glob the hashes, but
the ability to see if they have changed might point out problems that
would otherwise be missed.
This test is guarded by a '#if no-windows', but they can't be written out on
vfat on Linux either. con.xml can, and for testing that's good enough.
Also clean up a stray 'ls' that I mistakenly left in while writing the test.
