localrepo: iteratively derive local repository type
This commit implements the dynamic local repository type derivation
that was explained in the recent commit
bfeab472e3c0 "localrepo: create new function for instantiating a local
repo object."
Instead of a static localrepository class/type which must be customized
after construction, we now dynamically construct a type by building up
base classes/types to represent specific repository interfaces.
Conceptually, the end state is similar to what was happening when
various extensions would monkeypatch the __class__ of newly-constructed
repo instances. However, the approach is inverted. Instead of making
the instance then customizing it, we do the customization up front
by influencing the behavior of the type then we instantiate that
custom type.
This approach gives us much more flexibility. For example, we can
use completely separate classes for implementing different aspects
of the repository. For example, we could have one class representing
revlog-based file storage and another representing non-revlog based
file storage. When then choose which implementation to use based on
the presence of repo requirements.
A concern with this approach is that it creates a lot more types
and complexity and that complexity adds overhead. Yes, it is true that
this approach will result in more types being created. Yes, this is
more complicated than traditional "instantiate a static type." However,
I believe the alternatives to supporting alternate storage backends
are just as complicated. (Before I arrived at this solution, I had
patches storing factory functions on local repo instances for e.g.
constructing a file storage instance. We ended up having a handful
of these. And this was logically identical to assigning custom
methods. Since we were logically changing the type of the instance,
I figured it would be better to just use specialized types instead
of introducing levels of abstraction at run-time.)
On the performance front, I don't believe that having N base classes
has any significant performance overhead compared to just a single base
class. Intuition says that Python will need to iterate the base classes
to find an attribute. However, CPython caches method lookups: as long as
the __class__ or MRO isn't changing, method attribute lookup should be
constant time after first access. And non-method attributes are stored
in __dict__, of which there is only 1 per object, so the number of
base classes for __dict__ is irrelevant.
Anyway, this commit splits up the monolithic completelocalrepository
interface into sub-interfaces: 1 for file storage and 1 representing
everything else.
We've taught ``makelocalrepository()`` to call a series of factory
functions which will produce types implementing specific interfaces.
It then calls type() to create a new type from the built-up list of
base types.
This commit should be considered a start and not the end state. I
suspect we'll hit a number of problems as we start to implement
alternate storage backends:
* Passing custom arguments to __init__ and setting custom attributes
on __dict__.
* Customizing the set of interfaces that are needed. e.g. the
"readonly" intent could translate to not requesting an interface
providing methods related to writing.
* More ergonomic way for extensions to insert themselves so their
callbacks aren't unconditionally called.
* Wanting to modify vfs instances, other arguments passed to __init__.
That being said, this code is usable in its current state and I'm
convinced future commits will demonstrate the value in this approach.
Differential Revision: https://phab.mercurial-scm.org/D4642
--- a/mercurial/localrepo.py Tue Sep 18 15:15:24 2018 -0700
+++ b/mercurial/localrepo.py Tue Sep 18 15:29:42 2018 -0700
@@ -380,11 +380,32 @@
"""Create a local repository object.
Given arguments needed to construct a local repository, this function
- derives a type suitable for representing that repository and returns an
- instance of it.
+ performs various early repository loading functionality (such as
+ reading the ``.hg/requires`` and ``.hg/hgrc`` files), validates that
+ the repository can be opened, derives a type suitable for representing
+ that repository, and returns an instance of it.
The returned object conforms to the ``repository.completelocalrepository``
interface.
+
+ The repository type is derived by calling a series of factory functions
+ for each aspect/interface of the final repository. These are defined by
+ ``REPO_INTERFACES``.
+
+ Each factory function is called to produce a type implementing a specific
+ interface. The cumulative list of returned types will be combined into a
+ new type and that type will be instantiated to represent the local
+ repository.
+
+ The factory functions each receive various state that may be consulted
+ as part of deriving a type.
+
+ Extensions should wrap these factory functions to customize repository type
+ creation. Note that an extension's wrapped function may be called even if
+ that extension is not loaded for the repo being constructed. Extensions
+ should check if their ``__name__`` appears in the
+ ``extensionmodulenames`` set passed to the factory function and no-op if
+ not.
"""
ui = baseui.copy()
# Prevent copying repo configuration.
@@ -430,6 +451,9 @@
else:
extensions.loadall(ui)
+ # Set of module names of extensions loaded for this repository.
+ extensionmodulenames = {m.__name__ for n, m in extensions.extensions(ui)}
+
supportedrequirements = gathersupportedrequirements(ui)
# We first validate the requirements are known.
@@ -490,7 +514,46 @@
cachevfs = vfsmod.vfs(cachepath, cacheaudited=True)
cachevfs.createmode = store.createmode
- return localrepository(
+ # Now resolve the type for the repository object. We do this by repeatedly
+ # calling a factory function to produces types for specific aspects of the
+ # repo's operation. The aggregate returned types are used as base classes
+ # for a dynamically-derived type, which will represent our new repository.
+
+ bases = []
+ extrastate = {}
+
+ for iface, fn in REPO_INTERFACES:
+ # We pass all potentially useful state to give extensions tons of
+ # flexibility.
+ typ = fn(ui=ui,
+ intents=intents,
+ requirements=requirements,
+ wdirvfs=wdirvfs,
+ hgvfs=hgvfs,
+ store=store,
+ storevfs=storevfs,
+ storeoptions=storevfs.options,
+ cachevfs=cachevfs,
+ extensionmodulenames=extensionmodulenames,
+ extrastate=extrastate,
+ baseclasses=bases)
+
+ if not isinstance(typ, type):
+ raise error.ProgrammingError('unable to construct type for %s' %
+ iface)
+
+ bases.append(typ)
+
+ # type() allows you to use characters in type names that wouldn't be
+ # recognized as Python symbols in source code. We abuse that to add
+ # rich information about our constructed repo.
+ name = pycompat.sysstr(b'derivedrepo:%s<%s>' % (
+ wdirvfs.base,
+ b','.join(sorted(requirements))))
+
+ cls = type(name, tuple(bases), {})
+
+ return cls(
baseui=baseui,
ui=ui,
origroot=path,
@@ -666,8 +729,47 @@
return options
-@interfaceutil.implementer(repository.completelocalrepository)
+def makemain(**kwargs):
+ """Produce a type conforming to ``ilocalrepositorymain``."""
+ return localrepository
+
+@interfaceutil.implementer(repository.ilocalrepositoryfilestorage)
+class revlogfilestorage(object):
+ """File storage when using revlogs."""
+
+ def file(self, path):
+ if path[0] == b'/':
+ path = path[1:]
+
+ return filelog.filelog(self.svfs, path)
+
+def makefilestorage(requirements, **kwargs):
+ """Produce a type conforming to ``ilocalrepositoryfilestorage``."""
+ return revlogfilestorage
+
+# List of repository interfaces and factory functions for them. Each
+# will be called in order during ``makelocalrepository()`` to iteratively
+# derive the final type for a local repository instance.
+REPO_INTERFACES = [
+ (repository.ilocalrepositorymain, makemain),
+ (repository.ilocalrepositoryfilestorage, makefilestorage),
+]
+
+@interfaceutil.implementer(repository.ilocalrepositorymain)
class localrepository(object):
+ """Main class for representing local repositories.
+
+ All local repositories are instances of this class.
+
+ Constructed on its own, instances of this class are not usable as
+ repository objects. To obtain a usable repository object, call
+ ``hg.repository()``, ``localrepo.instance()``, or
+ ``localrepo.makelocalrepository()``. The latter is the lowest-level.
+ ``instance()`` adds support for creating new repositories.
+ ``hg.repository()`` adds more extension integration, including calling
+ ``reposetup()``. Generally speaking, ``hg.repository()`` should be
+ used.
+ """
# obsolete experimental requirements:
# - manifestv2: An experimental new manifest format that allowed
@@ -1325,11 +1427,6 @@
def wjoin(self, f, *insidef):
return self.vfs.reljoin(self.root, f, *insidef)
- def file(self, f):
- if f[0] == '/':
- f = f[1:]
- return filelog.filelog(self.svfs, f)
-
def setparents(self, p1, p2=nullid):
with self.dirstate.parentchange():
copies = self.dirstate.setparents(p1, p2)
--- a/mercurial/repository.py Tue Sep 18 15:15:24 2018 -0700
+++ b/mercurial/repository.py Tue Sep 18 15:29:42 2018 -0700
@@ -1222,8 +1222,21 @@
Raises ``error.LookupError`` if the node is not known.
"""
-class completelocalrepository(interfaceutil.Interface):
- """Monolithic interface for local repositories.
+class ilocalrepositoryfilestorage(interfaceutil.Interface):
+ """Local repository sub-interface providing access to tracked file storage.
+
+ This interface defines how a repository accesses storage for a single
+ tracked file path.
+ """
+
+ def file(f):
+ """Obtain a filelog for a tracked path.
+
+ The returned type conforms to the ``ifilestorage`` interface.
+ """
+
+class ilocalrepositorymain(interfaceutil.Interface):
+ """Main interface for local repositories.
This currently captures the reality of things - not how things should be.
"""
@@ -1439,12 +1452,6 @@
def wjoin(f, *insidef):
"""Calls self.vfs.reljoin(self.root, f, *insidef)"""
- def file(f):
- """Obtain a filelog for a tracked path.
-
- The returned type conforms to the ``ifilestorage`` interface.
- """
-
def setparents(p1, p2):
"""Set the parent nodes of the working directory."""
@@ -1572,3 +1579,7 @@
def savecommitmessage(text):
pass
+
+class completelocalrepository(ilocalrepositorymain,
+ ilocalrepositoryfilestorage):
+ """Complete interface for a local repository."""
--- a/mercurial/statichttprepo.py Tue Sep 18 15:15:24 2018 -0700
+++ b/mercurial/statichttprepo.py Tue Sep 18 15:29:42 2018 -0700
@@ -134,7 +134,8 @@
def canpush(self):
return False
-class statichttprepository(localrepo.localrepository):
+class statichttprepository(localrepo.localrepository,
+ localrepo.revlogfilestorage):
supported = localrepo.localrepository._basesupported
def __init__(self, ui, path):
--- a/tests/test-check-interfaces.py Tue Sep 18 15:15:24 2018 -0700
+++ b/tests/test-check-interfaces.py Tue Sep 18 15:29:42 2018 -0700
@@ -142,8 +142,10 @@
ziverify.verifyClass(repository.ipeerbase, unionrepo.unionpeer)
checkzobject(unionrepo.unionpeer(dummyrepo()))
- ziverify.verifyClass(repository.completelocalrepository,
+ ziverify.verifyClass(repository.ilocalrepositorymain,
localrepo.localrepository)
+ ziverify.verifyClass(repository.ilocalrepositoryfilestorage,
+ localrepo.revlogfilestorage)
repo = localrepo.makelocalrepository(ui, rootdir)
checkzobject(repo)