prune: lock inside the try tss…

debugrecordpruneparents: release tr the same way than the lock Nicer code is nicer.

evolve: drop useless wlock in rewrite

divergence: actually use the hint that have been around for age Having a hint in the code is good, having it displayed is usually seens as better.

evolve: rename `tr` variable (for troubled) to `tro` tro is usually dedicated to transaction.

evolve: add a custom layer around retractboundary retract boundary is about to gain a new argument in core, we need somethign to handle this

merge with stable

evolve: point to the core mercurial tracker for bugfix As per Matt Mackall request evolution related bug should be filled again the main tracker as at some point al lof evolve wil be in core.

evolve: drop reference to 2.3 in the extension help text We are far away from 2.3 now.

readme: update the work in progress repository location

readme: update the official documentation location

evolve: deprecated --obsolete and --old-obsolete option on commit and graft Such option are unlikely to make it into core, so we deprecate them to avoid uer discovery them and getting used to them.

push: put phase in the same bundle2 than changegroup and obsmarkers Note that the usual phase push from core will be performed in all case. But the discovery should find already in sync phase at that time. The old ways sync will be properly skipped when this logic is introduced into core.

push: put obsmarkers in the same bundle2 than changeset When client is Mercurial 3.1 and server have proper evolve version and both side have bundle2 enabled, we'll includes a part containing obsolescence markers. When obsmarkers are included to the main bundle2 push, they will failed to be pushed if the changeset failed to be pushed (and reciprocally).

push: extract obsmarkers discovery in a dedicated function The need to factorise the computation of the markers to send in a dedicated function. This prepare the inclusion of obsmarkers in an atomic bundle2 push. For now, when obsmarkers are pushed using bundle2, they are still doing it on their own, in a dedicated bundle2 push.

evolve: update the tested with statement Tested with the soon to be release 3.1.

test: adapt to upstream message change A message gained parenthesis in 3.1 we reflect than in the test.

uncommit: improve error message for use with no arguments

test: use `hg log` in the `getid` function using `hg id` to retrieve full lenght node requires to use --debug. This may include debug output (regarding branch cache invalidation for example) that breaks the test. We now use `hg log` with a custom template. This is a simple, clean and robust solution.

readme: mention patchbomb usage and patch description guidelines on wiki

debian: add line to clean target to delete error files caused by failing tests

fold: improve error messages for multiple heads and roots This commit adds hints and i18n to the error messages about non-linear revisions, along with corresponding tests.

fold: reword error message for public commits This rewords the error message when attempting to fold public csets. It is somewhat clearer to not use contractions in formal writing. It also uses i18n for it. This commit also adds a new test for this error message.

fold: overhaul handling of revisions with --rev (BC) The fold command parsed the revision arguments in a very peculiar and idiosyncratic fashion: if revisions were passed without a --rev argument, then they would all be extended to ".", but if they were passed with --rev, then no extension would happen. Passing revisions both with and without --rev would abort. This is inconsistent with the way all other hg commands parse revision arguments. We have several examples of command where several revisions are passed, and the --rev option is optional for specifying those revisions (update, strip, export). This patch alters the way in which fold parses its revision arguments. No distinction is made between revisions passed with or without the --rev argument. Regardless if a single or multiple revision is specified, all will be folded together into the parent of the working directory. If the --exact argument is passed, then the parent of the working directory is ignored and the specified revisions must be a single contiguous line. The docstring and tests are modified accordingly.

fold: emit error message when folding a single revision Previously, a confusing error state would happen if the revision set to fold was a singleton.

fold: clarify message for empty revision set and abort instead (BC) I find it clearer that there are no revisions to fold because none were specified. Otherwise it seems ambiguous: maybe there are no revisions to fold because it's impossible to fold whatever was given? In analogy of other commands that are given no revisions (e.g. strip, graft), this is also now an abort instead of a mere error message, which is a BC.

evolve: do not use _ as an unused variable The _ variable is the i18n function from core Mercurial, so locally assigning to it can lead to name clashes.

merge with stable

merge with very old 2.2 branch

test-obsolete: make log templates more consistent: add {desc} to one This has the extra added benefit of making the divergence situation easier to understand.

test-obsolete: explicitly show which changesets are divergent

merge with stable

update readme (I should have a hook preventing me to push to public repo with uncommited changes)

evolve: add --tool option so we don't have to use HGMERGE

amend: add -U/--current-user option (inspired by mq)

tests: fix grammar, spelling, punctuation in comments

evolve: minor reorganisation of code responsible for evolve output This does not change any of the semantic as status is not printed when --quiet is provided. This juste make it more explicit that the output are related.

template: add a troubles keyword This keyword list all evolution troubles that affect a changeset. Evolution troubles can be "unstable", "divergent" and "bumped".

docs: update copyright statement used by sphinx

docs: add concepts guide (the set theory behind obsolescence)

docs: add guide to sharing mutable history

docs: add new user guide This has also been reviewed to death on evolve-testers. There are still a couple of short sections to write (clearly marked "TODO"), and one example to add. But (if I may be so bold) this is a gigantic improvement over the current docs, so it really should get merged. Incidentally, the figures are all SVG files created with Inkscape. They're not perfect, but they're pretty nice. Anyone who knows a better way to create technical diagrams is welcome to contribute. One wart: there's a test script that accompanies the document and largely duplicates it, but I haven't unified them. I've been concentrating on writing the best possible content, not on fiddling with tools. I suspect that unifying them will be non-trivial, but definitely worth doing.

docs: rewrite index page (including dangling links to future docs) This has been reviewed to death on the evolve-testers mailing list. I've made some minor tweaks since the last round of reviews, but the main content is the same.

amend: add -D/--current-date option, just like mq's qrefresh has Carefully designed so we can easily add -D to other changeset-creating commands (probably anything that takes -d), and so that we can add -U/--current-user as well. When that's done, the date and user options should be very similar to those in MQ.

amend: explicitly test -d option

amend: in tests, do not rely on defaults -- always use explicit -d Having a default -d option will make it hard to test -D/--current-date, which is coming shortly: presence of -d overrides -D, just like with mq's qrefresh.

evolve: fix createmarkers() wrapper to accept an iterable Docstring in mercurial.obsolete says "<relations> must be an iterable of ...", and hgsubversion takes it at its word and passes a generator. Until around Mercurial 3.0, this did not cause any problems, but now it does: in a repo with both evolve and hgsubversion enabled, pushing to Subversion when obsolete changesets are present causes a crash: [...] File "/home/gward/src/hgsubversion/hgsubversion/wrappers.py", line 346, in push obsolete.createmarkers(repo, relations) File "/home/gward/src/hg-stable/mercurial/extensions.py", line 196, in wrap return wrapper(origfn, *args, **kwargs) File "/home/gward/src/hg-evolve/hgext/evolve.py", line 332, in _createmarkers relations[idx] = (prec, sucs, meta) TypeError: 'generator' object does not support item assignment Since hgsubersion is playing by the documented rules, evolve must be the guilty party here.

merge with 4.0.0

Added tag 4.0.0 for changeset 0304fc2bab15

pkg: prepare release of version 4.0.0

pkg: drop pushexperiment from the install list

merge stable into default default is meant to be a superset of stable

merge default in to stable evolve 4.0.0 is coming.

merge back into stable The 2,7 is meant to be a subset of stable.

prepare a branch for compatibility with mercurial 2.7 to 3.0(.0) Evolve 4.0.0 is coming

test: update to output change introduced by future 3.1

simple4server: backport the bundle2 exchange

evolve: install progress support for the bundle2 pull path The custom wireprotocol command had a trick to display progress while pulling. We reinstall such logic for the new path using bundle2. Its notable that we are using the bundle2 advisory parameter for this purpose.

evolve: install progress support for the bundle2 push path The custom wireprotocol command had a trick to display progress while pushing. We reinstall such logic for the new path using bundle2.

evolve: add a pull path using bundle2 Instead of using a dedicated wireprotocol commands, we use bundle2 to transmit an obs marker parts. This aims at both testing bundle2 more and to limit the amount of special code we needs to put in simple for server to fit our needs. The massive test changes comes from the fact we can use this fast path for both remote and local push.

evolve: advertise obsmarker creation in the transation hook This will let extensions hooks on marker exchange.

evolve: add bytes and new marker data when processing part This will helps use to keep printing useful message when pulling through bundle2.

evolve: add a push pass using bundle2 Instead of using a dedicated wireprotocol commands, we use bundle2 to transmit an obs marker parts. This aims at both testing bundle2 more and to limit the amount of special code we needs to put in simple for server to fit our needs. The massive test changes comes from the fact we can use this fast path for both remote and local push.

evolve: make exchange output compact in the old pushkey case too.

evolve: compatibility with memfilectx change The future 3.1 have have more sophisticated memfilectx object. We add a compatibility layer for them.

evolve: abstract access to context.memfilectx This class is getting altered in the future Mercurial 3.1.0. This changeset prepare a compatibility layer.

import: drop version check for ``import --obsolete`` This code base does not support anything below mercurial-3.0 anyway.

evolve: bump requirement to 3.0.1 Use of bundle2 for pulling will require this version or above.

test: fix test-import condition The help text is ``"only(`` not `` only(``

merge with stable

evolve: make most line lenghts 80 characters or less There are still a few lines that are above 80 characters in length. Those have not been changed because doing so would make the code a bit less clear.

evolve: start all util.Abort messages with lower case

evolve: improve wording of a few of messages

evolve: fix a couple of typos

evolve: fix spelling error

evolve: more compact output in the wire protocol case We made the wording more similar and use less line. The writing operation is usually fast so we do not needs a specific line for it.

evolve: test exchange through wireprotocol The wireprotocol version of the exchange was only tested in the `simple4server.py` case.

evolve: align working during obsmarker push using pushkey lets use the same wording everywhere to prevent user confusion.

evolve: drop some pre-3.0 compatibility code This extension already requires Mercurial 3.0

evolve: update the errors message to request mercurial 3.0 and above And update the checking code too. This extension already requires 3.0 for some time.

readme: remove some duplicated entries Those changes where backported to 3.3.2.

readme: remove embarrassment There were… conflict marker in this file. Its over, they are gone now.

merge with 3.3.2

Added tag 3.3.2 for changeset a03ea39aca2a

prepare release 3.3.2

fold: add squash as a possible alias In order to respect the habituation of the majority of DVCS users and to make this feature more easily discoverable, the very common "squash" terminology should be added as a synonym to "fold". Since git does not have a squash command, this should be safe.

fold: add --message and --logfile options Like other commit-creating commands, this add two options for specifying the resulting commit message, with tests. If either of these options is specified, no editor is popped up asking for a commit message.

merge with stable

prune: update to successor rather than parent when prune '.'

import: --obsolete flag for automatic obsolescence marker creation A new `--obsolete` flag is added to import. When present, the new node will be marked as a successors of the one specified in the `Node` field of the imported patch. No marker are created when revision have the node expected in the patch. This improves email based work flow where implicit rebase are likely to happen and extra information are lost, changing the hash. This new behavior requires a flags, otherwise the `hg export x | hg import -` idiom would change, turning the source obsolete. (Changing from `hg graft x` to `hg rebase --dest . --rev x`) This change only takes effect when using Mercurial 3.0 and above

merge with all other change to 3.3

update 3.3.0 readme to include minimal mercurial requirement This clarify the situation.

evolve: fix grammar in a comment

README: document how to run the tests

uncommit: change 'kill' to 'prune' in error message 'kill' was renamed to 'prune' Long time ago.

readme: the forgotten update strike again

evolve: protect call to rebase within a wlock (#42, #35, #16) Without a wlock, repo.commit would blow away the dirstate's parents on OSes that have no 'os.symlink' support in python, leading evolve to produce a merge instead of a rebase. If a user ran the rebase command instead of evolve, then things would work because rebase is wrapped in a giant wlock. Unfortunately, we can't use the same idea of wrapping the evolve command in one giant wlock because that's too early in the process. If the lock did wrap the entire evolve command, then the working directory would save its current parents which, since rebase hasn't been called yet, would be just p1. Therefore, we need to obtain the lock *after* the dirstate's parents are changed but *before* the call to rebase. This way ensures that when a conflict happens the working directory correctly shows both parent changeset.

readme: update the readme with all missing entry One days one will stop forgetting to update it.

readme: rephrase the changelog entry regarding the new exchange Lets try to have it readable by human being.

readme: drop the reference to fastobs Its dead Jim.

merge with stable

Added tag 3.3.1 for changeset fc04758ea9f5

prepare release 3.3.1

fix maintainer email I'm not working at Logilab since july 2013. Using my personal address to prevent this kind of situation in the future.

fold: enable --date and --user options Like all other commit-creating commits, `hg fold` should also have these options enabled. This also allows getting consistent hashes during tests involving fold.

evolve: improves readme wording (yes I just pushed the other version…)

evolve: update changelog in README

test: update testing statement

evolve: abort when trying to move a node on top of itself This provides a clearer error message when evolve misstakenly tries to move a changeset on top of itself. Before we got an abort with a cryptic message about trying to rebase a merge changeset. A proper fix to make sure that evolve never tries to move a node on top of itself will come in later patches.

run-tests: add executable bit

rewrite: use cmdutil.logmessage instead of ad-hoc methods The logic for handling the --message and --logfile options is already in core hg in the cmdutil.logmessage function. This patch removes this duplicate logic from Evolve and just calls the core function.

fold: enable --date and --user options Like all other commit-creating commits, `hg fold` should also have these options enabled. This also allows getting consistent hashes during tests involving fold.

merge with stable

test: sort output of listkey command The pushkey namespace came as dict key. So they have no defined order. We sort them to stabilize the test.

exchange: use fixed order for obstore content The obsstore content seems unstable from one version of python to another (we most certainly use some dict) we fix the order once and for all.

Finish killing fastobs (this test failed loudly after 33fdaec24432)

Kill fastobs, as marmoute says it causes problems

tests: fix test-prune.t breakage test-prune.t was broken due to 91d28bd0e04e in upstream Mercurial, which made hidden commits visible when a local tag made the commit reachable. The fix is to delete the tag in the test.

Fix preservation of rename information on evolve (#33) Following what rebase itself does, call duplicatecopies between rebasenode and concludenode. Should fix https://bitbucket.org/marmoute/mutable-history/issue/33/evolve-command-loses-copy-metadata

merge stable into default