pull: move pulled subset into the object We compute the set of local changeset that were target of the pull. This is then used by phases logic to decide which part of the history should have it phase updated. We move this information into the object to allow extraction of phase synchronisation in its own function. I expect obsolete marker exchange to use it too in the future.

revset: added operations to spanset to duck type baseset Added more operations which are not lazy but only used so far to duck type baseset.

revset: added basic operations to spanset Added methods __add__, __sub__ and __and__ to duck type more methods in baseset

revset: added spanset class to represent revision ranges

revset: added lazyset implementation to extra revset

revset: added lazyset implementation to converted revset

revset: added lazyset implementation to closed revset

push: feed pulloperation object to _pullobsolete function We now have all necessary information in the `pulloperation` object itself.

pull: move transaction logic into the pull object Most local change that occurs during a pull are doing within a `transaction`. Currently this mean (1) adding new changeset (2) adding obsolescence markers. We want the two operations to be done in the same transaction. However we do not want to create a transaction if nothing is added to the repo. Creating an empty transaction would drop the previous transaction data and confuse tool and people who are still using rollback. So the current pull code has some logic to create and handle this transaction on demand. We are moving this logic in to the `pulloperation` object itself to simplify this lazy creation logic through all different par of the push. Note that, in the future, other part of pull (phases, bookmark) will probably want to be part of the transaction too.

pull: move obsolescence marker exchange in the exchange module The obsolescence marker exchange code was already extracted during a previous cycle. We are moving the extracted functio in this module. This function will read and write data in the `pulloperation` object and I prefer to have all core function collaborating through this object in the same place. This changeset is pure code movement only. Code change for direct consumption of the `pulloperation` object will come later.

pull: move `force` argument into pull object One more step toward a more modular pulh function.

pull: move `heads` argument into pull object One more step toward a more modular pull function.

pull: move `remote` argument into pull object One more step toward a more modular pull function.

pull: introduce a pulloperation object This object will hold all data and state gathered through the pull. This will allow us to split the long function into multiple small one. Smaller function will be easier to maintains and wrap. The idea is to blindly store all information related to the pull in this object so that each step and extension can use them if necessary. We start by putting the `repo` variable in the object. More migration in other function.

tests: lines with largefile .* file://$TESTTMP also match on windows on windows, largefile paths are written as "file:///C:/temp/...", corresponding to "file:///$TESTTMP/..." (all three slashes shown). But on posix systems they are written as "file:///tmp/..." corresponding to "file://$TESTTMP/..." (only two slashes shown). Write the glob "file:/*/" to match both versions.

debugobsolete: extract marker display in a dedicated function We want to be able to reuse and extend it from other function or extension while working on markers exchange. This changeset is pure core movement.

exchange: extract pull function from localrepo The localrepo class if far too big. Push and pull logic are extracted and reworked to better fit with the fact we exchange more than bundle now. This changeset extract the pulh code. later changeset will slowly slice it into smaller brick. The localrepo.pull method is kept for now to limit impact on user code. But it will be ultimately removed, now that the public API is hold by peer.

push: extract new common set computation from phase synchronisation Now that every necessary information is held in the `pushoperation` object, we can extract the new common set computation to it's own function. This changeset is pure code movement only.

push: move `commonheads` into the push object The phase synchronisation start by computing the new set of common head between local and remote and then do the phase synchronisation on this set. This new common set logic will eventually be used by the obsolescence markers exchange. So we are going to split the long phase synchronisation in two.

push: move discovery in its own function Now that every necessary information is held in the `pushoperation` object, we can extract the discovery logic to it's own function. This changeset is pure code movement only.

push: move outgoing check logic in its own function Now that every necessary information is held in the `pushoperation` object, we can extract the part responsible of aborting the push to it's own function. This changeset is mostly pure code movement. the exception is the fact this function returns a value to decide if changeset bundle should be pushed.

push: move `incoming` into the push object The fact that there is some unknown changes on remote one of the result of discovery. It is then used by some push validation logic. We move it in the object to be able to extract the said logic.

push: move changeset push logic in its own function Now that every necessary information is held in the `pushoperation` object, we can extract the logic pushing changeset to it's own function. This changeset is pure code movement only.

push: move `remoteheads` into the push object The heads of the remote repository are used to detect race when pushing changeset. We now store this information in `pushoperation` object to allow extraction of the changeset pushing part.

revset: added lazyset implementation to contains revset

revset: added lazyset implementation to secret revset

revset: added lazyset implementation to matching revset Performance Benchmarking: $ time hg log -qr "first(matching(0))" 0:9117c6561b0b real 0m2.213s user 0m2.149s sys 0m0.055s $ time ./hg log -qr "first(matching(0))" 0:9117c6561b0b real 0m0.177s user 0m0.137s sys 0m0.038s

revset: added lazyset implementation to _matchfiles Performance Benchmarking: $ time hg log -qr "first(file(README))" 0:9117c6561b0b real 0m2.234s user 0m2.180s sys 0m0.044s $ time ./hg log -qr "first(file(README))" 0:9117c6561b0b real 0m0.172s user 0m0.129s sys 0m0.042s

revset: added lazyset implementation to checkstatus This improves the performance of the revsets 'adds' 'modifies' and 'removes' Performance benchmarking: $ time hg log -qr "first(adds(README))" 0:9117c6561b0b real 0m2.279s user 0m2.222s sys 0m0.053s $ time ./hg log -qr "first(adds(README))" 0:9117c6561b0b real 0m0.172s user 0m0.131s sys 0m0.041s $ time hg log -qr "first(modifies(README))" 1:273ce12ad8f1 real 0m2.292s user 0m2.227s sys 0m0.061s $ time ./hg log -qr "first(modifies(README))" 1:273ce12ad8f1 real 0m0.178s user 0m0.130s sys 0m0.038s $ time hg log -qr "first(removes(README))" 2379:e90cff87f871 real 0m2.297s user 0m2.235s sys 0m0.058s $ time ./hg log -qr "first(removes(README))" 2379:e90cff87f871 real 0m0.975s user 0m0.797s sys 0m0.056s

revset: added lazyset implementation to public revset Performance Benchmarking: $ time hg log -qr "first(public())" ... real 0m1.184s user 0m1.051s sys 0m0.130s $ time ./hg log -qr "first(public())" ... real 0m0.548s user 0m0.427s sys 0m0.118s