For a CVS repository checked out with "cvs co .", the prefix used to strip of
what we get from CVS was previously erroneously set to "repopath/.".
We now prevent the dot to be added.
Test folded in test-convert-cvs and simplified by Patrick Mézard
<pmezard@gmail.com>.
If the CVS repo somehow has a symbolic name that references a revision
consisting of a single number (e.g. BAD_TAG: 1), convert will fail when
attempting to find the branches, preventing the initial import from
working.
This patch skips those symbolic names--without warning.
Previously, it only checked for an mq patch if the user explicitly
passed -d/--dest. But rebasing onto an mq patch is a bad idea
regardless of how we determine the rebase destination.
They are not needed inside triple-quoted strings and they confuse the
line number computation done in i18n/hggettext. The script tries to
find the docstring in the source file. When \" in the source is turned
into just " in the docstring, the docstring can no longer be found.
- add a paragraph about default dest/source changesets
- option help: say "changeset" not "revision"
- option help: explain -b/--base better
- clarify that -a/--abort and -c/--continue are different from the
other options
This option is useful when other extensions (e.g., pbranch) want to
use rebase --collapse and append other things before committing.
This is not intended to be used from command line.
- add a paragraph about default dest/source changesets
- option help: say "changeset" not "revision"
- option help: explain -b/--base better
- clarify that -a/--abort and -c/--continue are different from the
other options
- kwcommitctx is inside the wlock of repo.commit: no lock
- _kwfwrite only needs wlock
wlock outside try block, so we don't need to import lock.release
_kwfwrite should even be safer that way, as we moved the status
call inside the try-except block.
Thanks to Benoit Boissinot for help.
The bzrlib try to import the ElementPath module but had a fallback in
case the import fails. Lazy import of this module leads to later
failure.
The bzrlib is used by the convert extension.
If the read thred is in select when the main thread is in close, the main
thread may close the socket between select and read, generating a noisy
traceback. This can be ignored if the shutdown flag is set.
hg inserve was ignoring and miscomputing the --timeout-idle option (seconds
vs. minutes).
Thanks to Jesse Glick for the bugreport and the initial patch.
Cached bookmark information would not be reset when repo.invalidate()
was called, meaning (for instanced) guis would be left with out-of-date
bookmark info. This patch fixes that by overriding bookmark_repo.invalidate()
foo [ ] 0/58
foo [> ] 1/58
foo [=> ] 2/58
...
foo [=======================================================> ] 56/58
foo [========================================================> ] 57/58
foo [=========================================================>] 58/58
The bar now has a '>' character at the end. This indicates the direction,
is consistent with the indeterminate '<=>' bar, and looks much nicer.
This has no effect in real world where progress.refresh in unlikely to be set
to zero, but is very useful in tests where all progress output is to be traced.
It failed on platforms with coarse time.time() granularity, like Windows+py25.
The goal of this patch is to add the IDs of the parents of applied MQ patches
into the patch file headers whenever qnew or qrefresh are run.
This will serve as a reminder of when the patches last applied cleanly and
will let us do more intelligent things in the future, such as:
* Resolve conflicts found when qpushing to a new location by merging
instead of simply showing rejects.
* Display better diffs of versioned MQ patches because we can tell how the
patched files have changed in the meantime.
Here are the new rules this patch introduces. They are checked in this order:
* If a patch currently has old, plain-style patch headers ("From:" and
"Date:") do not change the style or add any new headers.
* If the 'mq.plain' configuration setting is true, only plain-style
headers will be used for all MQ patches.
* qnew will initialize new patches with HG-style headers and fill in the
"# Parent" header with the appropriate parent node.
* qrefresh will refresh the "# Parent" header with the current parent of
the current patch.
Previously, the name part of an repo#name url was interpreted as a
revision, similar to using the --rev option. Now it is instead looked
up as a branch first, and if that succeeds all the heads of the branch
will be processed instead of just its tip-most head. If the branch
lookup fails, it will be assumed to be an revision as before (e.g. for
tags).
The slow path/fast path distinction has existed since mq was added to Mercurial
in 2006. The slow path was used whenever the refreshed revision wasn't the tip
We should not translate the "imported patch" message. The translated message
confuses the detection whether the user has not updated the commit message yet.
We try to avoid to translate generated commit messages.
kw_diff actually disabled restricted mode when 2 revisions were given,
because it effectively disables the extension in this case.
But the commands working with diff and patch need restricted mode
always enabled, i.e. expansion enabled when writing to the
working directory and - crucial for these commands - no expansion
when reading the filelog.
Limit was interpreted as absolute, from the topmost revision, without
counting the number of revisions matching a given file.
Which caused "glog -lN file" to show sometimes less than N csets if
the file was not modified in all of the N previous csets.
glog will now match the behavior of log.
Otherwise when processing a changeset that in fact changes no files
(perhaps due to bug in import from CVS) can get something like:
unexpected svn output:
abort: unable to cope with svn output
Bug report and patch draft by Jesse Glick <jesse.glick@sun.com>
We need to do this to avoid file/directories conflicts.
This causes patches removing a committed file and replacing it
with a directory to be "unpoppable".
The docstring is translated twice: once when used as a format string,
and once on display. The second translation fails when the first
translation introduces non-ASCII characters in the string.
The problem is that the gettext module calls unicode(message) on the
string, i.e., it decodes it to a Unicode string using the ASCII
encoding (the default encoding). By translating it into a Unicode
string here, the unicode() call becomes a noop.
* prefix messages by inotify-(client|server)
* make sure that all warning and abort messages use the same format.
* in the case where inotify.sock is an old broken symlink, say so and abort
instead of trying to overwrite the already existing link
This will mainly help us in our tests to log pids of inotify servers
started implicitely, to make sure that unkilled inotify daemons do not clutter
the output of unrelated tests.
Also desactivate the workaround introduced in 37824a274d63
Since it only changes the working directory, it does not matter whether a patch is
applied. This change makes it easier to use hg import --no-commit instead of patch.
This fixes an incompatibility with patch(1), which also uses --reverse
for reversed diffs. The --inverse flag was added in 0f0383897d54. That
name was chosen over --reverse since it was thought that --reverse
would make --rev ambiguous.
It turns out that both flags can co-exist, with the cost that --rev
can no longer be shortened to --r and --re. Since one can always use
the short -r option, this is not a real problem.
Original patch was provided by Simon Heimberg
It delegates dirstate computation to dirstate.status when dirstate is dirty:
better be slow from time to time instead of using wrong data.
This solves issue1719. As the last component, issue1810, is still not solved,
test-inotify-dirty-dirstate will fail for now. It emphasizes a regression due
to e8efd88001e7:
changeset: 9515:e8efd88001e7
user: Nicolas Dumazet <nicdumz.commits@gmail.com>
date: Sun Aug 16 11:11:37 2009 +0900
summary: inotify: use cmdutil.service instead of local daemonizing code
Ancestors of 7c01599dd30 are passing the test, when applied this patch.
Regression has to be investigated, but this patch is important since it affects
often mq operations.
Emulate the match.dir calls that are made in dirstate.walk:
* first mark the visited directories on the server side
* then extend the transmitted response to include this directory list
* and lastly call match.dir on each directory
The default list includes commands which normally print large amounts of
text. This should be more user-friendly than paging all commands by default,
which can be confusing when the pager swallows input prompts (with, e.g.,
record, merge, HTTP/SSH authentication, etc.)
The prompt function would return 'y' or 'n' untranslated. This should
therefore not be compared to _('y') and _('n'). However, it turns out
that prompt could just as well return a good old Boolean.
convert extension tries to guess the remote repository type with HTTP probes.
Unfortunately, it does not handle authentication or HTTPS handshakes, so regular
svn repositories may be excluded. Instead, when a non-404 error is retrieved,
we keep trying with the svn bindings. The drawback is missing svn bindings will
make the conversion to fail even for non-svn targets. This can be avoided with
--source.
mbox format should use time.asctime(). Unfortunately, this function writes
2-characters day of week on Windows while unix one writes a single character.
Normalize to Windows version since the other one can hardly be written with
strftime().
(issue1561)
Before this change, newancestor was used only once as a replacement
for ancestor.ancestor, but merge.update calls ancestor.ancestor
several times, so it ends up with the "wrong" ancestor (the real
ancestor, but we want the parent of the rebased changeset for all but
the first rebased changeset).
Added a new test case for this: test-rebase-newancestor.
Also, in one scenario in test-rebase-collapse, there was a spurious
conflict caused by the same issue, so that test case was fixed by
removing the now unneeded conflict resolution and the output was
adapted accordingly.
After changeset 54b67f7bd5df, ui.username will abort by default if no
username is set. It is therefore wrong to call ui.username when a
username was provided on the command line.
Most of the time, one can reverse a diff by swapping the revisions passed with
-r but it happens that if you use the global -R, and diff against the tip of
the current repo, you can't swap the revisions. One use-case for that is
reviewing changes from a bundle before unbundling. One could also pipe the
output of `hg diff` to a command line filter that reverses the diff, but that
would remove the benefit from color diffs. Therefore, having an option in
`hg diff` to reverse a diff is a good thing.
The option flag selection was tricky. GNU patch uses -R/--reverse but -R is
already used as a global option and --reverse would make --rev ambiguous.
Normally, diffs without any text insertions or deletions are reported
as having 0 lines changed by stock diffstat. Compatibility is
preserved with stock diffstat in this case, but when using --git,
binary files are marked with Bin as a means of clarification.
git diff --stat does something similar, though it also includes the
old and new file sizes.
diff/qdiff --stat invokes patch.diffstat() on the diff output.
When in interactive mode, the output's maximum width is determined by the
terminal's width.
This only happens when using --base (or no source selection options), as
rebase already aborts in this situation when using --source.
Without this change you get an abort from the underlying merge, and the
repository is in a different state than you started with (the working
dir parent is changed).
In some environments merges occur regularly but with no conflicts, and
committers find merge notifications more of a bother than a help.
By setting merge=False in [notify], merge notifications are suppressed.
This works both for incoming and for changegroup hooks.