git patches may require copies to be handled out-of-order. For instance, take
the following sequence:
* modify a
* copy a into b
Here, we have to generate b from a before its modification. To do so,
applydiff() was scanning for copy metadata and performing the copies before
processing the other changes in-order. While smart and efficient, this approach
complicates things by handling file copies and file creations at different
places and times. While a new file must not exist before being patched a copied
file already exists before applying the first hunk.
Instead of copying the files at their final destination before patching, we
store them in a temporary file location and retrieve them when patching. The
filestore always stores file content in real files but nothing prevents adding
a cache layer. The filestore class was kept separate from fsbackend for at
least two reasons:
- This class is likely to be reused as a temporary result store for a future
repository patching call (entries just have to be extended to contain copy
sources).
- Delegating this role to backends might be more efficient in a repository
backend case: the source files are already available in the repository itself
and do not need to be copied again. It also means that third-parties backend
would have to implement two other methods. If we ever decide to merge the
filestore feature into backend, a minimalistic approach would be to compose
with filestore directly. Keep in mind this copy overhead only applies for
copy/rename sources, and may even be reduced to copy sources which have to
handled ahead of time.
Restore the previous diffstat behaviour of scaling by the maximum number of
changes to a single file. Changeset 7bb0e22a7988 modified the diffstat to be
scaled by the total number of changes. This seems to have been unintentional.
extensions that depend on other extensions (such as record) use this pattern
to check if the dependant extension is available:
try:
mq = extensions.find('mq')
except KeyError:
return
but since if an error occurs while loading an extension it leaves its entry
in the _extensions map as None, we want to raise in that situation too.
(rather than adding another check if the return value is None)
Failing to do so makes it impossible to use the memctx API to create a
changeset with a commit message or username outside of the current
encoding.encoding setting.
When trying to import a url that ends with a slash
os.path.basename() would return an empty string. So
when seeing a leading slash, remove it and infer the
patch name from the remaining url.
e.g. `hg qimport http://paste.pocoo.org/raw/xxx/`
deduced '.' to be the patch name. Now we'll deduce 'xxx'.
gitpatch objects emitted by iterhunks() are modified in place by applydiff().
Processing them earlier improves iterhunks() isolation. applydiff() modifying
them should still be fixed though.
Let ui.plain() accept an optional parameter in the form of a feature
name (as a string) to exclude from plain mode.
The result of ui.plain is now:
- False if HGPLAIN is not set or the requested feature is in HGPLAINEXCEPT
- True otherwise
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@anciens.enib.fr>
_updatedir() is no longer used by internalpatch()
The change in test-mq-missingfiles.t comes from workingbackend not considering
the missing 'b' file as changed, thus not calling addremove() on it.
When used as the default merge tool, or used as a --tool override,
the simplemerge script should not be allowed to raise a util.Abort
just because one of the files being merged is binary. Instead, return
1 and mark the file unresolved.
hg.bat expects to live in pythonxx/scripts and the python interpreter
to be in pythonxx. run-tests.py file layout is a little different and
python location must be fixed.
Suppresses output (resolved paths or "not found!") when searching a path,
similar to "grep -q".
Sample usage: hg paths -q foo || echo "there is no foo"
Just prints path names (instead of "name = result") when listing all path
definitions, like "hg bookmarks -q".
Sample usage: hg paths -q | while read i; do hg incoming "$i"; done
The verb to allow requires a direct object.
Lintian, a Debian tool to find common mistakes, reported it. I'm not a
native english speaker but I think this is correct.
These open the changelog and manifest, respectively, directly so you don't
need to specify the path.
The options have been added to debugindex, debugdata and debugrevlog.
The patch also fixes some minor usage-related bugs.
str(url) was recently changed to return only file:/. However, the
canonical way to represent absolute local paths is file:/// [1], which
is also expected by at least hgsubversion.
Relative paths are returned as file:the/relative/path.
[1] http://en.wikipedia.org/wiki/File_URI_scheme
Before this patch undo.bookmarks was created on bookmarks write and
not with other transaction-related files. There were two issues: first
is that if you have changed bookmarks few times after a transaction
happened, rollback will give you a state which can point to
non-existing revision. Second is that if you have not changed
bookmarks after a transaction, rollback will touch your state anyway.
This change also adds `localrepo._writejournal` method, which can be
used by other extensions to save their transaction-related backup in
right time.
This happens more often than expected. Say you have an svn subrepository with
python code. Python would have generated unknown .pyc files. Now, you rebase
this setup on a revision where a directory containing python code does not
exist. Subversion is first asked to remove this directory when updating, but
will not because it contains untracked items. Then it will have to bring back
the directory after the merge but will fail because it now collides with an
untracked directory.
Using --force is not very elegant and only works with svn >= 1.5 but the only
alternative I can think of is to write our own purge command for subversion.
localstr's hash method exists to prevent bogus matching on lossy local
encodings. For instance, we don't want 'caf?' to match 'café' in an
ASCII locale.
But when café can be losslessly encoded in the local charset, we can
simply use a normal string and avoid the hashing trick.
This avoids using localstr's hash method, which would prevent a match between
This is a simple patch to make hg push/hg outgoing print their remote target
path even if the operation fails. I'm not sure if the original behavior was by
design.
This patch also changes one test to reflect the changed behaviour.
After a rollback, the current bookmark might be absent from the bookmarks
file. In that case we discard it instead of displaying a traceback during
commit.
This backs out
changeset: 13158:17d1b96c0f12
user: Mads Kiilerich <mads@kiilerich.com>
date: Tue Dec 07 03:29:21 2010 +0100
summary: merge: fast-forward merge with descendant
Before named branches, the invariants were:
a) "merges" always have two parents
b) p1 is not linearly related to p2
Adding named branches made (b) problematic, so the above patch was
introduced, which fixed (b) but broke (a).
After discussion, we decided that the invariants should be:
a) "merges" always have two parents
b) p1 is not linearly related to p2 OR p1 and p2 are on different branches
This happens more often than expected. Say you have an svn subrepository with
python code. Python would have generated unknown .pyc files. Now, you rebase
this setup on a revision where a directory containing python code does not
exist. Subversion is first asked to remove this directory when updating, but
will not because it contains untracked items. Then it will have to bring back
the directory after the merge but will fail because it now collides with an
untracked directory.
Using --force is not very elegant but it is much simpler than rewriting our own
purge command for subversion.
This patch prevent mq to crash when .hg/patches/status contains Malformed lines
(without ":"). Blank lines are ignored and other malformed lines issue a
warning.
Handle parse errors in the .hgeol similarly to how parse errors in the
.hgtags file are handled: by issuing a warning. This allows the user
to revert the file using 'hg revert' or 'hg update -C'.
When matching a file against the rules in .hgeol, the eol extension's
hook should stop after the first matching rule is encountered.
Otherwise, if this rule is contradicted by other more general rule
(for example a catch-all at the end of .hgeol), some files are simply
impossible to push. Trivial example:
**.bat = CRLF
** = LF
If all matching rules were applied, a .bat file would be rejected
either because it has LFs (first rule) or because it has CRLFs (second
rule).
Converting from subversion specifying config.svn.trunk results
in storing trunk under branch named as config.svn.trunk, where `default'
brunch is expected. Submission contains patch and test.
It was suggested in IRC that people disabling the reporting of unstructed hgrc
files can masquerade as problems. This makes sure untrusted hgrc files are
always reported if --debug is used.
A bookmark can only move to a descendant on commit, pull or
unbundle. Bookmarks cannot jump between heads anymore. This fixese
issue 1502.
We explicitly use new.node(), to emphasise that we are updating the
current bookmark to the new node.
Since bookmarks are no longer merged with repo.tags() as of
8e2d23f4bd25, they don't show up in `hg id` as they used to. This adds
them back into the summary that `hg id` prints, and adds a
-B/--bookmarks flag alongside the -t/--tags and -b/--branch options.
Note this introduces a slight backwards-incompatibility: the summary
printed by `hg id` now separates bookmarks from tags with a space, as
seen below, instead of running it into the tags list.
Default summary output:
$ hg id
db815d6d32e6 tip/tag1 bm1/bm2
Output with --bookmarks:
$ hg id --bookmarks
bm1 bm2
See also 5672c9e8202d which adds bookmarks back into `hg summary`.
If the repository is not locked when clearing the dirstate, then
running test-eol.t in a loop fails sooner or later with:
ERROR: /home/mg/src/mercurial-crew/tests/test-eol.t output changed
--- /home/mg/src/mercurial-crew/tests/test-eol.t
+++ /home/mg/src/mercurial-crew/tests/test-eol.t.err
@@ -343,6 +343,7 @@
% hg status (eol activated)
M win.txt
% hg commit
+ nothing changed
% hg status
$ testmixed CRLF
However, if we cannot lock the repository, then we can also not make a
commit and so we can simply ignore a LockUnavailable error.
hg bookmark -r ancestorrev X will not mark X as the current bookmark anymore.
If you want to point a bookmark to a ancestor rev you will use hg update to
move to it. This will set the current bookmark.
This improves the misleading error message
$ hg identify
abort: there is no Mercurial repository here (.hg not found)!
to the more explicit
$ hg identify
abort: requirement 'fake' not supported!
for all commands in commands.optionalrepo, which includes the identify
and serve commands in particular.
This is for the case when a new entry in .hg/requires will be defined
in a future Mercurial release.
Previously, when rolling back a transaction, some users could be confused
between the level to which the store is rolled back, and the new parents
of the working directory.
$ hg rollback
rolling back to revision 4 (undo commit)
With this change:
$ hg rollback
repository tip rolled back to tip revision 4 (undo commit)
working directory now based on revision 2 and 1
So now the user can realize that the store has been rolled back to an older
tip, but also that the working directory may not on the tip (here we are
rolling back the merge of the heads 2 and 1)
Invalid requests could give an unhandled ErrorResponse.
Now this ErrorResponse is handled like other ErrorResponses so the client gets
an error message which also is logged on the server.
This changes the prompts on git subrepos to show only the first seven
digits of git changeset IDs (as git's command line does):
$ hg update
subrepository sources for s differ (in checked out version)
use (l)ocal source (32a3438) or (r)emote source (da5f5b1)?
We restrict : to 1. make it easer to convert bookmarks to git branches,
2. use : later for a syntax to push a local bookmark to a remote bookmark
of a different name. \0, \n, \r are fobbidden they are used to separate
bookmarks in the bookmark file.
This change breaks backward compatbility as ':' was an allowed character in
previous versions.
Consider a repository with a single subrepository. The changesets in
the main repository reference the subrepository changesets like this:
m0 -> s0
m1 -> s1
m2 -> s2
Starting from a state (m1, s0), doing 'hg update m2' in the main
repository will yield a conflict: the subrepo is at revision s0 but
the target revision says it should be at revision s2.
Before this change, Mercurial would do (m1, s0) -> (m2, s2) and thus
ignore the conflict between the working copy and the target revision.
With this change, the user is prompted to resolve the conflict by
choosing which revision he wants. This is consistent with 'hg merge',
which also prompts the user when it detects conflicts in the merged
.hgsubstate files.
The prompt looks like this:
$ hg update tip
subrepository sources for my-subrepo differ
use (l)ocal source (fc627a69481f) or (r)emote source (12a213df6fa9)?
The default behaviour is to commit subrepositories with uncommitted changes. In
my experience this is usually undesirable:
- Changes to dependencies are often debugging leftovers
- Real changes should generally be applied on the source project directly,
tested then committed. This is not always possible, subversion subrepos may
include only a small part of the source project, without the tests.
Setting ui.commitsubrepos=no will now abort commits containing such modified
subrepositories like:
$ hg --config ui.commitsubrepos=no ci -m msg
abort: uncommitted changes in subrepo sub
I ruled out the hook solution because it does not easily take --include/exclude
options in account. Also, my main concern is whether this flag could cause
problems with extensions. If there are legitimate reasons for callers to
override this behaviour (I could not find any), they might either override at ui
level, or we could add an argument to localrepo.commit() later.
v2:
- Renamed ui.commitsubs to ui.commitsubrepos
- Mention the configuration entry in hg help subrepos
Add missing calls to close() to many places where files are
opened. Relying on reference counting to catch them soon-ish is not
portable and fails in environments with a proper GC, such as PyPy.
For GUI clients its sometimes important to know which files will be ignored and
which files will be important. This allows the GUI client to skipping redoing a
'hg status' when the files are ignored but have changed. (For instance, a
typical case is that the "build" directory inside some project is ignored but
files in it frequently change.)
The patch changes the output of "hg diff --stat" when one file whose filename
has spaces has changed, making it get the full filename instead of just the
substring between the last space and the end of the filename.
It also changes the diffstat generated by "hg email -d" when one of the commit
messages starts with "diff". Because of the regex used to parse the filename,
the diffstat generated by "hg email -d" will still be not correct if a commit
message starts with "diff -r ".
Before the patch Mercurial has the following behavior:
$ echo "foobar">"file with spaces"
$ hg add "file with spaces"
$ hg diff --stat
spaces | 1 +
1 files changed, 1 insertions(+), 0 deletions(-)
$ hg diff --git --stat
file with spaces | 1 +
1 files changed, 1 insertions(+), 0 deletions(-)
After the patch:
$ echo "foobar">"file with spaces"
$ hg add "file with spaces"
$ hg diff --stat
file with spaces | 1 +
1 files changed, 1 insertions(+), 0 deletions(-)
$ hg diff --git --stat
file with spaces | 1 +
1 files changed, 1 insertions(+), 0 deletions(-)
Before the patch:
$ hg add mercurial/patch.py tests/tests-diffstat.t
$ hg commit -m "diffstat: fix parsing of filenames"
$ hg email -d --test tip
This patch series consists of 1 patches.
diffstat: fix parsing of filenames
[...]
filenames | 0
mercurial/patch.py | 6 ++++--
tests/test-diffstat.t | 17 +++++++++++++++++
3 files changed, 21 insertions(+), 2 deletions(-)
[...]
After the patch:
$ hg email -d --test tip
This patch series consists of 1 patches.
diffstat: fix parsing of filenames
[...]
mercurial/patch.py | 6 ++++--
tests/test-diffstat.t | 17 +++++++++++++++++
3 files changed, 21 insertions(+), 2 deletions(-)
[...]
We can't assume that all pushable patches early in the series have already been
applied. If a hg qselect is done while you already have patches applied, some
patches with guards may now be pushable, even though they come earlier in the
series.
So instead of checking only applied patches, explicitly check where we are in
the series against the position of the patch we want to qpush to.
Add empty repository.close() and call it in dispatch.
Remove bundlerepository.__del__(), merging it into bundlerepository.close(),
which overrides repository.close().
http://docs.python.org/reference/datamodel.html says:
"It is not guaranteed that __del__() methods are called for objects that
still exist when the interpreter exits."
