The new method will decide between:
- cleanupnode, which calls the unsafe repair.strip
- create obsmarkers
Ideally, nobody calls "cleanupnode" directly except for "safecleanupnode".
Recently we switched rebases to run the entire rebase inside a single
transaction, which dramatically improved the speed of rebases in repos with
large working copies. Let's also move the dirstate into a single dirstateguard
to get the same benefits. This let's us avoid serializing the dirstate after
each commit.
In a large repo, rebasing 27 commits is sped up by about 20%.
I believe the test changes are because us touching the dirstate gave the
transaction something to actually rollback.
This adds an option (which defaults to False) to run entire histedits in a
single transaction. This results in 20-25% faster histedits in large repos where
transaction startup cost is expensive.
I didn't want to enable this by default because it has some unfortunate side
effects. For instance, if a pretxncommit hook throws midway through the
histedit, it will rollback the entire histedit and lose any progress the user
had made. Same if the user aborts editting a commit message. It's still worth
turning this on for large repos, but probably not for normal sized repos.
Long term, once we have inmemory merging, we could do the entire histedit in
memory, without a transaction, then we could selectively rollback just parts of
it in the event of an exception.
Tested it by running the tests with
`--extra-config-opt=histedit.singletransaction=True`. The only failure was
related to the hook rollback issue I mention above.
We only want to pop the action after the action is completed, since if the
action aborts part way through we want it to remain at the front of the list so
continue/abort will start with it.
Previously we relied on the fact that we only serialized the state file at the
beginning of the action, so the pop wasn't serialized until the next iteration
of the loop. In a future patch we will be adding a large transaction around this
area, which means if we pop the list early it might get serialized if the action
throws a user InterventionRequired error, at which point the action is not in
the list anymore. So let's only pop it once the action is really truly done.
Changeset 97936471dc8d removed the mutable default value, but did not explicitly
tested for None. Such implicit testing can introduce semantic and performance
issue. We move to an explicit testing for None as recommended by PEP8:
https://www.python.org/dev/peps/pep-0008/#programming-recommendations
Changeset 33b71926122d removed the mutable default value, but did not explicitly
tested for None. Such implicit checking can introduce semantic and performance
issue. We move to an explicit check for None as recommended by PEP8:
https://www.python.org/dev/peps/pep-0008/#programming-recommendations
This patch makes us respect pager.attend again if the extension is
enabled. It also brings back the default attend list, so e.g. summary
is not paged by default, just like it used to be before pager was
moved into core.
The named branch of the leaf changeset can be changed by updating to it,
setting the branch, and amending.
But previously, there was no good way to *just* change the branch of several
linear changes. If rebasing changes with another parent to '.', it would pick
up a pending branch change up. But when rebasing changes that have the same
parent, it would fail with 'nothing to rebase', even when the branch name was
set differently.
To fix this, allow rebasing to same parent when a branch has been set.
By including the working directory revision at the start of rebase in
the repo._rebaseset, we make sure it's not hidden when we update back
to it at the end of the rebase.
This feels like abusing the set a bit given its name (_rebaseset), but
I couldn't think of another name that's clearly better.
Now that the 'vfs' classes moved in their own module, lets use the new module
directly. We update code iteratively to help with possible bisect needs in the
future.
Now that the 'vfs' classes moved in their own module, lets use the new module
directly. We update code iteratively to help with possible bisect needs in the
future.
Now that the 'vfs' classes moved in their own module, lets use the new module
directly. We update code iteratively to help with possible bisect needs in the
future.
Now that the 'vfs' classes moved in their own module, lets use the new module
directly. We update code iteratively to help with possible bisect needs in the
future.
Now that the 'vfs' classes moved in their own module, lets use the new module
directly. We update code iteratively to help with possible bisect needs in the
future.
Previously, rebasing would open several transaction over the course of rebasing
several commits. Opening a transaction can have notable overhead (like copying
the dirstate) which can add up when rebasing many commits.
This patch adds a single large transaction around the actual commit rebase
operation, with a catch for intervention which serializes the current state if
we need to drop back to the terminal for user intervention. Amazingly, almost
all the tests seem to pass.
On large repos with large working copies, this can speed up rebasing 7 commits
by 25%. I'd expect the percentage to be a bit larger for rebasing even more
commits.
There are minor test changes because we're rolling back the entire transaction
during unexpected exceptions instead of just stopping mid-rebase, so there's no
more backup bundle. It also leave an unknown file in the working copy, since our
clean up 'hg update' doesn't delete unknown files.
Previously, if .hg/rebasestate existed but .hg/last-message.txt was missing, 'hg
rebase --abort' would say there's no rebase in progress but 'hg checkout foo'
would say 'abort: rebase in progress'. It turns out loading the collapse message
will throw a "no rebase in progress" error if the file doesn't exist, even
though .hg/rebasestate obviously indicates a rebase is in progress.
The fix is to only throw an exception if we're trying to --continue, and to just
eat the issues if we're doing --abort.
This issue is exposed by us writing the rebase state earlier in the process.
This will be used by later patches to ensure the user can appropriately 'hg
rebase --abort' if there's a crash before the first the first commit has
finished rebasing. Tests cover all of this. The only negative affect is we now
require a hg rebase --abort in a very specific exception case, as shown in the
test.
The rebaseruntime class already has the restorestatus function, so let's make it
own the store status function too. This get's rid of a lot of unnecessary
argument passing and will make a future patch cleaner that refactors storestatus
to support transactions.
Previously, rebase --abort would only call update if you were on a node that had
already been rebased. This meant that if the rebase failed during the rebase of
the first commit, the working copy would be left dirty (with a .hg/updatestate
file) and rebase --abort would not have update to clean it up.
The fix is to also perform an update if you're still on the target node or on
the original working copy node (since the working copy may be dirty, we still
need to do the update). We don't want to perform an update in all cases though
because of issue4009.
A subsequent patch makes this case much more common, since it causes the entire
rebase transaction to rollback during unexpected exceptions. This causes the
existing test-rebase-abort.t to cover this case.
This can be used to flag patches by branch or topic automatically. Flags
optionally given by --flag option are exported as {flags} template keyword,
so you can add --flag V2.
I want to move _getpatches() to _getpatchmsgs() to make sure each patch text
is tied with the corresponding revision number. This helps adding templater
support.
IIRC, the pbranch extension doesn't work with the recent Mercurial versions,
so the removal of this option wouldn't hurt.
Storing a relative path the source repository is useful when exporting
repositories over the network or when they're located on external
drives where the mountpoint isn't always fixed.
Currently, Mercurial interprets paths in `.hg/shared` relative to
$PWD. I suspect this is very much unintentional, and you have to
manually edit `.hg/shared` in order to trigger this behaviour.
However, on the off chance that someone might rely on it, I added a
new capability called 'relshared'. In addition, this makes earlier
versions of Mercurial fail with a graceful error.
I should note that I haven't tested this patch on Windows.
Now that the feature no longer lives in the extension, we document it in the
help of the core config. This include the new 'ui.color' option introduced in
the previous changesets.
As a result the color extensions can now be deprecated.
This is a documentation patch only; color is still disabled by default.
This is the last bits we needed to move out of the extensions. 'hgext/color.py'
now only contains logic to changes the default color behavior to 'auto'.
However, more cleanups are on the way and we need to document the new config
directly in core.
We've made chg utilize the common code path implemented in pager.py (by
e8fb65f5e551 and e97133c7a9dc), but the chg server does not always start
with a tty. Because of this, uisetup() of the pager extension could be
skipped on the chg server.
Kudos given to Sean Farley for dogfooding new chg and spotting this problem.
This patch makes it possible to unshelve while having missing files
in your repo as long as shelved changes don't touch those missing files.
It also makes error message better otherwise.
Before this patch, there was a single way to see multiple shelves using
`--patch --list` which show all the shelves. Doing `--patch s1 s2` returns an
error. This patch allows to show multiple shelves using `--patch` and `--stat`.
pycompat.getenv returns os.getenvb on py3 which is not available on Windows.
This patch replaces them with encoding.environ.get and checks to ensure no
new instances of os.getenv or os.setenv are introduced.
When converting a Git repository to Mercurial at Mozilla, I encountered
a scenario where I didn't want `hg convert` to automatically add the
"committer: <committer>" line to commit messages. While I can hack around
this by rewriting the Git commit before it is fed into `hg convert`,
I figured it would be a useful knob to control.
This patch introduces a config option that allows lots of control
over the committer value. I initially implemented this as a single
boolean flag to control whether to save the committer message. But
then there was feedback that it would be useful to save the committer
in extra data. While this patch doesn't implement support for saving
in extra data, it does add a mechanism for extending which actions
to take on the committer field. We should be able to easily add
actions to save in extra data.
Some of the implemented features weren't asked for. But I figured they
could be useful. If nothing else they demonstrate the extensibility
of this mechanism.
Detailed hint message is now provided when 'pull --rebase' operation detects
unclean working dir, for example:
abort: uncommitted changes
(cannot pull with rebase: please commit or shelve your changes first)
Added tests for uncommitted merge, and for subrepo support verifying that same
hint is also passed to subrepo state check.
Since chg has its own _runpager implementation, it's no longer necessary to
special-case chg in the pager extension. This will effectively enable the
new chg pager code path that runs inside runcommand.
Refuse to run 'hg pull --rebase' if there are uncommitted changes:
so that instead of going ahead with fetching changes and then suddenly aborting
the rebase, we can warn user of uncommitted changes (or unclean repo state)
right up front.
In tests, we create a 'histedit' session to verify that also an unfinished
state is detected and handled.
As discussed at [1], ui._runpager will be the new low-level API accepting a
pager command to actually run the pager. And ui.pager is the high-level API
which reads config directly from self.
This change is necessary for chgserver to override _runpager cleanly.
[1]: www.mercurial-scm.org/pipermail/mercurial-devel/2016-December/091656.html
This dictionnary is affected by the content of the config, so we should have
one for each ui config.
We rename the global dict to '_baseterminfoparams' to make the situation
clearer.
This new option control whether or not color will be used. It mirror the behavior
of '--color'. I usually avoid adding new option to '[ui]' as the section is
already filled with many option. However, I feel like 'color' is central enough
to deserves a spot in this '[ui]' section.
For now the option is not documented so it is still marked as experimental. Once
it get documented and official, we should be able to deprecate the color
extensions.
There is more cleanup to do before that documentation is written, but we need
this option early to made them. Having that option will allow for more cleanup
of the initialisation process and proper separation between color
configuration.
We now run the color initialisation as part of the standard dispatch. This is
opening the way for multiple cleanups since we now have access to the multiple 'ui'
object and we'll be able to see difference between global and local config. This
cleanup will arrive in later changesets.
As a side effect, the '--color' flag is now working without the extension.
Since we now properly initialize color for each ui idependently, we get a
warning message twice.
If we want to be able to move the initialisation in core, we need core to be
aware of that '--color' flag at all time. So we now have the definition in core. That flag
is currently unprocessed without the extensions (will be fixed soon). In
addition the default value for this flag in core is 'never'. Enabling the
extensions change that default value to 'auto'.
If 'ui.plain()' is set we should not colorize. We move that logic into the
function that determine and setup the color mode. As all other code respect
the resulting mode this will be equivalent.
Now that all ui instance carry a '_colormode' attribute, we can access and
comply to it directly in the subrepo code. The actual implementation could
probably be a bit smarter, but we stick close to the current one for the sake
of simplicity.
The decoders were already run by default for the main repo, so this seemed like
an oversight.
The extdiff extension has been using 'archive' since a80ec1ea2694 to support -S,
and a colleague noticed that after diffing, making changes, and closing it, the
line endings were wrong for the diff-tool modified files in the subrepository.
(Files in the parent repo were correct, with the same .hgeol settings.) The
editor (Visual Studio in this case) reloads the file, but doesn't notice the EOL
change. It still adds new lines with the original EOL setting, and the file
ends up inconsistent.
Without this change, the first file `cat`d in the test prints '\r (esc)' EOL,
but the second doesn't on Windows or Linux.
Now that all logics formally bared by 'colorui' have been moved to the main ui
class, that class is empty and can be dropped. As a nice side effect we can get
rid of the baroque Initialization associated to it.
There was much rejoicing.
This is similar to what we needed for 'write', we move the logic from the
extension to the core class. Beside the dispatch to 'win32print', we just apply
label to the argument.
One more step, the support for writing color is not directly in core. No
behavior change for the default case ('_colormode' = None).
Here are the details of what we have to change to the core method:
* apply to 'self.label' to input in the buffered case
* dispatch to 'win32print' when applicable
* apply to 'self.label' to input when applicable
This bring us closer to supporting color in core natively. Core already have a
'label' method that was a no-op. We update its to call the new 'colorlabel'
function. Behavior is unchanged when colormode = None.
This function is quite simple and only have one call site… located a handful of line
under the definition. We inline the function for the sake of simplicity. One of
the motivation to do that now is that it reduce the amount of new method we
will have to add to the core 'ui' class for color support.
Having all 'ui' objects aware of 'color' allows us to update the core code to
handle color. The mode will stay 'None' in the default case so that will not
introduce any changes.
This should not introduce any behavior changes when using the color extension.
In practive, the colormode will be setup at early at run time to the proper
value (from config and environment).
We do this change as this gets us closer of a state were we can have all the
mechanisms associated to color in core with the feature disabled by default.
No logic change is introduced. The previous code were using a 'dict.update' with
a complex generator. That was a bit confusing to read and not so compact.
Instead we now use an explicit loop and conditional for the sake of clarity.
This also allow for more simplification coming in the next changeset.
We won't use terminfo when curse failed to load. Before this change, we were
doing an indirect test, relying on the fact some variable ('_terminfo_params')
would be empty if curses failed to load. We update the code to be more explicit
and directly checks if we managed to load the curse module.
This is another part of moving color implementation into core. before we can
move the advance logic, we need to move the basic definition and building
bricks. This is one more step on that road.
This change adjusts and documents the new behaviour of 'roll'. It now fits nicely
with the behaviour of 'commit --amend' and the 'edit' action, by discarding the
date as well as the commit message of the second commit. Previously it used the
later date, like 'fold', but this often wasn't desirable, for example, in the
common use case of using 'roll' to add forgotten changes to a changeset
(because 'hg add' was previously forgotten or not all changes were identified
while using 'hg record').
This clarifies in the histedit documentation that the 'edit' action preserves
the date and that the 'fold' action uses the later date. The documentation was
previously silent on this issue which left users in doubt.
Before this patch, checking HG_PENDING in bookmarks.py might cause
unintentional reading unrelated '.hg/bookmarks.pending' in, because it
just examines existence of HG_PENDING environment variable.
This patch uses txnutil.trypending() to check HG_PENDING strictly.
This patch also changes share extension.
Enabling share extension (+ bookmark sharing) makes
bookmarks._getbkfile() receive repo to be shared (= "srcrepo"). On the
other hand, HG_PENDING always refers current working repo (=
"currepo"), and bookmarks.pending is written only into currepo.
Therefore, we should try to read .hg/bookmarks.pending of currepo in
at first. If it doesn't exist, we try to read .hg/bookmarks of srcrepo
in.
Even after this patch, an external hook spawned in currepo can't see
pending changes in currepo via srcrepo, even though such changes
become visible after closing transaction, because there is no easy and
cheap way to know existence of pending changes in currepo via srcrepo.
Please see https://www.mercurial-scm.org/wiki/SharedRepository, too.
BTW, this patch may cause failure of bisect in the repository of
Mercurial itself, if examination at bisecting assumes that an external
hook can see all pending changes while nested transactions across
repositories.
This invisibility issue will be fixed by subsequent patch, which
allows HG_PENDING to refer multiple repositories.
This results in the default pager-attend list being empty. Sadly, we
can't let the code be that way, because some legacy extensions depend
on hooking the pager's attend list at import time (and we'd like to
not break them), and if the list is actually *empty* that triggers
magic behavior in the extension that attends everything. Instead, we
put a long, improbable command name as the only entry in the attend
list.
New revsetlang module hosts parser, tokenizer, and miscellaneous functions
working on parsed tree. It does not include functions for evaluation such as
getset() and match().
2288 mercurial/revset.py
684 mercurial/revsetlang.py
2972 total
get*() functions are aliased since they are common in revset.py.
This makes using shelve/unshelve more consistent because
shelving can be done using name option and unshelving as
well. Author of the idea of this improvement and solution is
joshgold.
This closes the last feature gap other than the attend list from the
extension. For now, I'm leaving the attend list in the extension,
because I'm unsure it has merit in a world where commands have been
updated to take advantage of the modern API.
This moves the global flag and the --pager=yes logic into core. Only
functionality change is that users now always get a --pager flag and
can enable the pager via the flag without the extension active.
Moving the flag into core exposes a defect in the ro localization,
which will have to be corrected later.
No functionality change.
A previous version of this API had a category argument on
ui.pager(). As I migrated the commands in core, I couldn't come up
with good enough consistency in any categorization scheme so I just
scrapped the whole idea. It may be worth revisiting in the future.
Add support for the bugzilla rest api documented at
https://wiki.mozilla.org/Bugzilla:REST_API
and at
https://bugzilla.readthedocs.io/en/latest/
This backend has the following benefits:
* It supports the bugzilla api keys so hgrc does not need to contain
a user's bugzilla password
* Works with Mercurial's "hostfingerprints" support making handling
bugzilla instances with self-signed certs easier
* Does not use xmlrpc ;-)
Adds configuration item 'apikey' in [bugzilla] section.
My major concern with these patches is if the approach to HTTP access
is the right way for an extension and if hooking into request object
and the overriding the get_method to perform PUT requests was a
sensible approach.
# no-check-commit
Changeset 8b50aee1f75c removes SIGPIPE handling completely. This is wrong,
as it means that Mercurial does not exit when the pager does. Instead, raise
SignalInterrupt when SIGPIPE happens with a pager attached, to trigger the
normal exit path.
This will cause "killed!" to be printed to stderr (hence the BC warning),
but in the normal pager use case (where the pager gets both stderr and
stdout), this message is lost as we only get SIGPIPE when the pager quits.
zeroconf only knows how to deal with IPv4; I develop on a system where the only
IPv4 address is 127.0.0.1.
Teach zeroconf to ignore IPv6 addresses when looking for plausible IPv4
connectivity.
pager replaced stdout with a line buffered version to work around glibc
deciding on a buffering strategy on the first write to stdout. This is going
to make my next patch hard, as replacing stdout will make tracking time
spent blocked on it more challenging.
Move the line buffering requirement to util.py, and remove it from pager.
This means that the abuse of ui.formatted=True and pager set to cat or equivalent
no longer results in a line-buffered output to a pipe, hence (BC), although
I don't expect anyone to be affected
The default SIGPIPE handler causes Mercurial to exit immediately, without
running any Python cleanup code (except and finally blocks, atexit handlers
etc). This creates problems if you want to do something at exit.
If we need a different exit code for broken pipe from pager, then we should
code that ourselves in Python; this appears to have been cargo-culted from
the fork implementation of pager that's no longer used, where it was needed
to stop Broken Pipe errors appearing on the user's terminal.
The comment was introduced in 0a14c8556910 (rebase: ensure rebase
revision remains visible (issue4504), 2015-01-27), which mentions the
right issue in the description.
shlex.split() only accepts unicodes on Python 3. After this patch we will be
using pycompat.shlexsplit(). This patch also replaces existing occurences of
shlex.split with pycompat.shlexsplit.
Currently if our branch name contains '\' or starts with '.', shelve chooses
an illegal shelve name. This behaviour is not good as it itself is choosing
something which it won't accept further. We can raise errors if user passes
a name which is illegal.
After this patch, if '\' is contained in branch name or bookmark name, it will
be replaced by '_' while choosing a shelve name and if they starts with '.',
the first '.' is replaced by '_'.