Mercurial > hg
view mercurial/pycompat.py @ 35569:964212780daf
rust: implementation of `hg`
This commit provides a mostly-working implementation of the
`hg` script in Rust along with scaffolding to support Rust in
the repository.
If you are familiar with Rust, the contents of the added rust/
directory should be pretty straightforward. We create an "hgcli"
package that implements a binary application to run Mercurial.
The output of this package is an "hg" binary.
Our Rust `hg` (henceforth "rhg") essentially is a port of the existing
`hg` Python script. The main difference is the creation of the embedded
CPython interpreter is handled by the binary itself instead of relying
on the shebang. In that sense, rhg is more similar to the "exe wrapper"
we currently use on Windows. However, unlike the exe wrapper, rhg does
not call the `hg` Python script. Instead, it uses the CPython APIs to
import mercurial modules and call appropriate functions. The amount of
code here is surprisingly small.
It is my intent to replace the existing C-based exe wrapper with rhg.
Preferably in the next Mercurial release. This should be achievable -
at least for some Mercurial distributions. The future/timeline for
rhg on other platforms is less clear. We already ship a hg.exe on
Windows. So if we get the quirks with Rust worked out, shipping a
Rust-based hg.exe should hopefully not be too contentious.
Now onto the implementation.
We're using python27-sys and the cpython crates for talking to the
CPython API. We currently don't use too much functionality of the
cpython crate and could have probably cut it out. However, it does
provide a reasonable abstraction over unsafe {} CPython function
calls. While we still have our fair share of those, at least we're
not dealing with too much refcounting, error checking, etc. So I
think the use of the cpython crate is justified. Plus, there is
not-yet-implemented functionality that could benefit from cpython. I
see our use of this crate only increasing.
The cpython and python27-sys crates are not without their issues.
The cpython crate didn't seem to account for the embedding use case
in its design. Instead, it seems to assume that you are building
a Python extension. It is making some questionable decisions around
certain CPython APIs. For example, it insists that
PyEval_ThreadsInitialized() is called and that the Python code
likely isn't the main thread in the underlying application. It
is also missing some functionality that is important for embedded
use cases (such as exporting the path to the Python interpreter
from its build script). After spending several hours trying to
wrangle python27-sys and cpython, I gave up and forked the project
on GitHub. Our Cargo.toml tracks this fork. I'm optimistic that
the upstream project will accept our contributions and we can
eventually unfork.
There is a non-trivial amount of code in our custom Cargo build
script. Our build.rs (which is called as part of building the hgcli
crate):
* Validates that the Python interpreter that was detected by the
python27-sys crate provides a shared library (we only support
shared library linking at this time - although this restriction
could be loosened).
* Validates that the Python is built with UCS-4 support. This ensures
maximum Unicode compatibility.
* Exports variables to the crate build allowing the built crate to e.g.
find the path to the Python interpreter.
The produced rhg should be considered alpha quality. There are several
known deficiencies. Many of these are documented with inline TODOs.
Probably the biggest limitation of rhg is that it assumes it is
running from the ./rust/target/<target> directory of a source
distribution. So, rhg is currently not very practical for real-world
use. But, if you can `cargo build` it, running the binary *should*
yield a working Mercurial CLI.
In order to support using rhg with the test harness, we needed to hack
up run-tests.py so the path to Mercurial's Python files is set properly.
The change is extremely hacky and is only intended to be a stop-gap
until the test harness gains first-class support for installing rhg.
This will likely occur after we support running rhg outside the
source directory.
Despite its officially alpha quality, rhg copes extremely well with
the test harness (at least on Linux). Using
`run-tests.py --with-hg ../rust/target/debug/hg`, I only encounter
the following failures:
* test-run-tests.t -- Warnings emitted about using an unexpected
Mercurial library. This is due to the hacky nature of setting the
Python directory when run-tests.py detected rhg.
* test-devel-warnings.t -- Expected stack trace missing frame for `hg`
(This is expected since we no longer have an `hg` script!)
* test-convert.t -- Test running `$PYTHON "$BINDIR"/hg`, which obviously
assumes `hg` is a Python script.
* test-merge-tools.t -- Same assumption about `hg` being executable with
Python.
* test-http-bad-server.t -- Seeing exit code 255 instead of 1 around
line 358.
* test-blackbox.t -- Exit code 255 instead of 1.
* test-basic.t -- Exit code 255 instead of 1.
It certainly looks like we have a bug around exit code handling. I
don't think it is severe enough to hold up review and landing of this
initial implementation. Perfect is the enemy of good.
Differential Revision: https://phab.mercurial-scm.org/D1581
author | Gregory Szorc <gregory.szorc@gmail.com> |
---|---|
date | Wed, 10 Jan 2018 08:53:22 -0800 |
parents | e66d6e938d2d |
children | 1a31111e6239 |
line wrap: on
line source
# pycompat.py - portability shim for python 3 # # This software may be used and distributed according to the terms of the # GNU General Public License version 2 or any later version. """Mercurial portability shim for python 3. This contains aliases to hide python version-specific details from the core. """ from __future__ import absolute_import import getopt import os import shlex import sys ispy3 = (sys.version_info[0] >= 3) ispypy = (r'__pypy__' in sys.builtin_module_names) if not ispy3: import cookielib import cPickle as pickle import httplib import Queue as _queue import SocketServer as socketserver import xmlrpclib else: import http.cookiejar as cookielib import http.client as httplib import pickle import queue as _queue import socketserver import xmlrpc.client as xmlrpclib empty = _queue.Empty queue = _queue.Queue def identity(a): return a if ispy3: import builtins import functools import io import struct fsencode = os.fsencode fsdecode = os.fsdecode oslinesep = os.linesep.encode('ascii') osname = os.name.encode('ascii') ospathsep = os.pathsep.encode('ascii') ossep = os.sep.encode('ascii') osaltsep = os.altsep if osaltsep: osaltsep = osaltsep.encode('ascii') # os.getcwd() on Python 3 returns string, but it has os.getcwdb() which # returns bytes. getcwd = os.getcwdb sysplatform = sys.platform.encode('ascii') sysexecutable = sys.executable if sysexecutable: sysexecutable = os.fsencode(sysexecutable) stringio = io.BytesIO maplist = lambda *args: list(map(*args)) ziplist = lambda *args: list(zip(*args)) rawinput = input # TODO: .buffer might not exist if std streams were replaced; we'll need # a silly wrapper to make a bytes stream backed by a unicode one. stdin = sys.stdin.buffer stdout = sys.stdout.buffer stderr = sys.stderr.buffer # Since Python 3 converts argv to wchar_t type by Py_DecodeLocale() on Unix, # we can use os.fsencode() to get back bytes argv. # # https://hg.python.org/cpython/file/v3.5.1/Programs/python.c#l55 # # TODO: On Windows, the native argv is wchar_t, so we'll need a different # workaround to simulate the Python 2 (i.e. ANSI Win32 API) behavior. if getattr(sys, 'argv', None) is not None: sysargv = list(map(os.fsencode, sys.argv)) bytechr = struct.Struct('>B').pack class bytestr(bytes): """A bytes which mostly acts as a Python 2 str >>> bytestr(), bytestr(bytearray(b'foo')), bytestr(u'ascii'), bytestr(1) (b'', b'foo', b'ascii', b'1') >>> s = bytestr(b'foo') >>> assert s is bytestr(s) __bytes__() should be called if provided: >>> class bytesable(object): ... def __bytes__(self): ... return b'bytes' >>> bytestr(bytesable()) b'bytes' There's no implicit conversion from non-ascii str as its encoding is unknown: >>> bytestr(chr(0x80)) # doctest: +ELLIPSIS Traceback (most recent call last): ... UnicodeEncodeError: ... Comparison between bytestr and bytes should work: >>> assert bytestr(b'foo') == b'foo' >>> assert b'foo' == bytestr(b'foo') >>> assert b'f' in bytestr(b'foo') >>> assert bytestr(b'f') in b'foo' Sliced elements should be bytes, not integer: >>> s[1], s[:2] (b'o', b'fo') >>> list(s), list(reversed(s)) ([b'f', b'o', b'o'], [b'o', b'o', b'f']) As bytestr type isn't propagated across operations, you need to cast bytes to bytestr explicitly: >>> s = bytestr(b'foo').upper() >>> t = bytestr(s) >>> s[0], t[0] (70, b'F') Be careful to not pass a bytestr object to a function which expects bytearray-like behavior. >>> t = bytes(t) # cast to bytes >>> assert type(t) is bytes """ def __new__(cls, s=b''): if isinstance(s, bytestr): return s if (not isinstance(s, (bytes, bytearray)) and not hasattr(s, u'__bytes__')): # hasattr-py3-only s = str(s).encode(u'ascii') return bytes.__new__(cls, s) def __getitem__(self, key): s = bytes.__getitem__(self, key) if not isinstance(s, bytes): s = bytechr(s) return s def __iter__(self): return iterbytestr(bytes.__iter__(self)) def iterbytestr(s): """Iterate bytes as if it were a str object of Python 2""" return map(bytechr, s) def sysbytes(s): """Convert an internal str (e.g. keyword, __doc__) back to bytes This never raises UnicodeEncodeError, but only ASCII characters can be round-trip by sysstr(sysbytes(s)). """ return s.encode(u'utf-8') def sysstr(s): """Return a keyword str to be passed to Python functions such as getattr() and str.encode() This never raises UnicodeDecodeError. Non-ascii characters are considered invalid and mapped to arbitrary but unique code points such that 'sysstr(a) != sysstr(b)' for all 'a != b'. """ if isinstance(s, builtins.str): return s return s.decode(u'latin-1') def strurl(url): """Converts a bytes url back to str""" return url.decode(u'ascii') def bytesurl(url): """Converts a str url to bytes by encoding in ascii""" return url.encode(u'ascii') def raisewithtb(exc, tb): """Raise exception with the given traceback""" raise exc.with_traceback(tb) def getdoc(obj): """Get docstring as bytes; may be None so gettext() won't confuse it with _('')""" doc = getattr(obj, u'__doc__', None) if doc is None: return doc return sysbytes(doc) def _wrapattrfunc(f): @functools.wraps(f) def w(object, name, *args): return f(object, sysstr(name), *args) return w # these wrappers are automagically imported by hgloader delattr = _wrapattrfunc(builtins.delattr) getattr = _wrapattrfunc(builtins.getattr) hasattr = _wrapattrfunc(builtins.hasattr) setattr = _wrapattrfunc(builtins.setattr) xrange = builtins.range unicode = str def open(name, mode='r', buffering=-1): return builtins.open(name, sysstr(mode), buffering) def _getoptbwrapper(orig, args, shortlist, namelist): """ Takes bytes arguments, converts them to unicode, pass them to getopt.getopt(), convert the returned values back to bytes and then return them for Python 3 compatibility as getopt.getopt() don't accepts bytes on Python 3. """ args = [a.decode('latin-1') for a in args] shortlist = shortlist.decode('latin-1') namelist = [a.decode('latin-1') for a in namelist] opts, args = orig(args, shortlist, namelist) opts = [(a[0].encode('latin-1'), a[1].encode('latin-1')) for a in opts] args = [a.encode('latin-1') for a in args] return opts, args def strkwargs(dic): """ Converts the keys of a python dictonary to str i.e. unicodes so that they can be passed as keyword arguments as dictonaries with bytes keys can't be passed as keyword arguments to functions on Python 3. """ dic = dict((k.decode('latin-1'), v) for k, v in dic.iteritems()) return dic def byteskwargs(dic): """ Converts keys of python dictonaries to bytes as they were converted to str to pass that dictonary as a keyword argument on Python 3. """ dic = dict((k.encode('latin-1'), v) for k, v in dic.iteritems()) return dic # TODO: handle shlex.shlex(). def shlexsplit(s): """ Takes bytes argument, convert it to str i.e. unicodes, pass that into shlex.split(), convert the returned value to bytes and return that for Python 3 compatibility as shelx.split() don't accept bytes on Python 3. """ ret = shlex.split(s.decode('latin-1')) return [a.encode('latin-1') for a in ret] else: import cStringIO bytechr = chr bytestr = str iterbytestr = iter sysbytes = identity sysstr = identity strurl = identity bytesurl = identity # this can't be parsed on Python 3 exec('def raisewithtb(exc, tb):\n' ' raise exc, None, tb\n') def fsencode(filename): """ Partial backport from os.py in Python 3, which only accepts bytes. In Python 2, our paths should only ever be bytes, a unicode path indicates a bug. """ if isinstance(filename, str): return filename else: raise TypeError( "expect str, not %s" % type(filename).__name__) # In Python 2, fsdecode() has a very chance to receive bytes. So it's # better not to touch Python 2 part as it's already working fine. fsdecode = identity def getdoc(obj): return getattr(obj, '__doc__', None) def _getoptbwrapper(orig, args, shortlist, namelist): return orig(args, shortlist, namelist) strkwargs = identity byteskwargs = identity oslinesep = os.linesep osname = os.name ospathsep = os.pathsep ossep = os.sep osaltsep = os.altsep stdin = sys.stdin stdout = sys.stdout stderr = sys.stderr if getattr(sys, 'argv', None) is not None: sysargv = sys.argv sysplatform = sys.platform getcwd = os.getcwd sysexecutable = sys.executable shlexsplit = shlex.split stringio = cStringIO.StringIO maplist = map ziplist = zip rawinput = raw_input isjython = sysplatform.startswith('java') isdarwin = sysplatform == 'darwin' isposix = osname == 'posix' iswindows = osname == 'nt' def getoptb(args, shortlist, namelist): return _getoptbwrapper(getopt.getopt, args, shortlist, namelist) def gnugetoptb(args, shortlist, namelist): return _getoptbwrapper(getopt.gnu_getopt, args, shortlist, namelist)