contrib/byteify-strings.py

rust-discovery: using from Python code

As previously done in other topics, the Rust version is used if it's been
built.

The version fully in Rust of the partialdiscovery class has the performance
advantage over the Python version (actually using the Rust MissingAncestor) if
the undecided set is big enough. Otherwise no sampling occurs, and the
discovery is reasonably fast anyway.

Note: it's hard to predict the size of the initial undecided set, it can
depend on the kind of topological changes between the local and remote graphs.
The point of the Rust version is to make the bad cases acceptable.

More specifically, the performance advantages are:

- faster sampling, especially takefullsample()
- much faster addmissings() in almost all cases (see commit message in
  grandparent of the present changeset)
- no conversion cost of the undecided set at the interface between Rust and
  Python

== Measurements with big undecided sets

For an extreme example, discovery between mozilla-try and mozilla-unified
(over one million undecided revisions, same case as in dbd0fcca6dfc), we
get roughly a x2.5/x3 better performance:

Growing sample size (5% starting with 200): time goes down from
210 to 72 seconds.
Constant sample size of 200: time down from 1853 to 659 seconds.

With a sample size computed from number of roots and heads of the
undecided set (`respectsize` is `False`), here are perfdiscovery results:

Before ! wall 9.358729 comb 9.360000 user 9.310000 sys 0.050000 (median of 50)
After  ! wall 3.793819 comb 3.790000 user 3.750000 sys 0.040000 (median of 50)

In that later case, the sample sizes are routinely in the hundreds of
thousands of revisions. While still faster, the Rust iteration in
addmissings has less of an advantage than with smaller sample sizes, but
one sees addcommons becoming faster, probably a consequence of not having
to copy big sets back and forth.

This example is not a goal in itself, but it showcases several different
areas in which the process can become slow, due to different factors, and
how this full Rust version can help.

== Measurements with small undecided sets

In cases the undecided set is small enough than no sampling occurs,
the Rust version has a disadvantage at init if `targetheads` is really big
(some time is lost in the translation to Rust data structures),
and that is compensated by the faster `addmissings()`.

On a private repository with over one million commits, we still get a minor
improvement, of 6.8%:

  Before ! wall 0.593585 comb 0.590000 user 0.550000 sys 0.040000 (median of 50)
  After  ! wall 0.553035 comb 0.550000 user 0.520000 sys 0.030000 (median of 50)

What's interesting in that case is the first addinfo() at 180ms for Rust and
233ms for Python+C, mostly due to add_missings and the children cache
computation being done in less than 0.2ms on the Rust side vs over 40ms on the
Python side.

The worst case we have on hand is with mozilla-try, prepared with
discovery-helper.sh for 10 heads and depth 10, time goes up 2.2% on the median.
In this case `targetheads` is really huge with 165842 server heads.

  Before ! wall 0.823884 comb 0.810000 user 0.790000 sys 0.020000 (median of 50)
  After  ! wall 0.842607 comb 0.840000 user 0.800000 sys 0.040000 (median of 50)

If that would be considered a problem, more adjustments can be made, which are
prematurate at this stage: cooking special variants of methods of the inner
MissingAncestors object, retrieving local heads directly from Rust to avoid
the cost of conversion. Effort would probably be better spent at this point
improving the surroundings if needed.

Here's another data point with a smaller repository, pypy, where performance
is almost identical

 Before ! wall 0.015121 comb 0.030000 user 0.020000 sys 0.010000 (median of 186)
 After  ! wall 0.015009 comb 0.010000 user 0.010000 sys 0.000000 (median of 184)

Differential Revision: https://phab.mercurial-scm.org/D6430

#!/usr/bin/env python3
#
# byteify-strings.py - transform string literals to be Python 3 safe
#
# Copyright 2015 Gregory Szorc <[e-mail]>
#
# This software may be used and distributed according to the terms of the
# GNU General Public License version 2 or any later version.

from __future__ import absolute_import, print_function

import argparse
import contextlib
import errno
import os
import sys
import tempfile
import token
import tokenize

def adjusttokenpos(t, ofs):
    """Adjust start/end column of the given token"""
    return t._replace(start=(t.start[0], t.start[1] + ofs),
                      end=(t.end[0], t.end[1] + ofs))

def replacetokens(tokens, opts):
    """Transform a stream of tokens from raw to Python 3.

    Returns a generator of possibly rewritten tokens.

    The input token list may be mutated as part of processing. However,
    its changes do not necessarily match the output token stream.
    """
    sysstrtokens = set()

    # The following utility functions access the tokens list and i index of
    # the for i, t enumerate(tokens) loop below
    def _isop(j, *o):
        """Assert that tokens[j] is an OP with one of the given values"""
        try:
            return tokens[j].type == token.OP and tokens[j].string in o
        except IndexError:
            return False

    def _findargnofcall(n):
        """Find arg n of a call expression (start at 0)

        Returns index of the first token of that argument, or None if
        there is not that many arguments.

        Assumes that token[i + 1] is '('.

        """
        nested = 0
        for j in range(i + 2, len(tokens)):
            if _isop(j, ')', ']', '}'):
                # end of call, tuple, subscription or dict / set
                nested -= 1
                if nested < 0:
                    return None
            elif n == 0:
                # this is the starting position of arg
                return j
            elif _isop(j, '(', '[', '{'):
                nested += 1
            elif _isop(j, ',') and nested == 0:
                n -= 1

        return None

    def _ensuresysstr(j):
        """Make sure the token at j is a system string

        Remember the given token so the string transformer won't add
        the byte prefix.

        Ignores tokens that are not strings. Assumes bounds checking has
        already been done.

        """
        k = j
        currtoken = tokens[k]
        while currtoken.type in (token.STRING, token.NEWLINE, tokenize.NL):
            k += 1
            if (
                currtoken.type == token.STRING
                and currtoken.string.startswith(("'", '"'))
            ):
                sysstrtokens.add(currtoken)
            try:
                currtoken = tokens[k]
            except IndexError:
                break

    def _isitemaccess(j):
        """Assert the next tokens form an item access on `tokens[j]` and that
        `tokens[j]` is a name.
        """
        try:
            return (
                tokens[j].type == token.NAME
                and _isop(j + 1, '[')
                and tokens[j + 2].type == token.STRING
                and _isop(j + 3, ']')
            )
        except IndexError:
            return False

    def _ismethodcall(j, *methodnames):
        """Assert the next tokens form a call to `methodname` with a string
        as first argument on `tokens[j]` and that `tokens[j]` is a name.
        """
        try:
            return (
                tokens[j].type == token.NAME
                and _isop(j + 1, '.')
                and tokens[j + 2].type == token.NAME
                and tokens[j + 2].string in methodnames
                and _isop(j + 3, '(')
                and tokens[j + 4].type == token.STRING
            )
        except IndexError:
            return False

    coldelta = 0  # column increment for new opening parens
    coloffset = -1  # column offset for the current line (-1: TBD)
    parens = [(0, 0, 0, -1)]  # stack of (line, end-column, column-offset, type)
    ignorenextline = False  # don't transform the next line
    insideignoreblock = False # don't transform until turned off
    for i, t in enumerate(tokens):
        # Compute the column offset for the current line, such that
        # the current line will be aligned to the last opening paren
        # as before.
        if coloffset < 0:
            lastparen = parens[-1]
            if t.start[1] == lastparen[1]:
                coloffset = lastparen[2]
            elif (
                t.start[1] + 1 == lastparen[1]
                and lastparen[3] not in (token.NEWLINE, tokenize.NL)
            ):
                # fix misaligned indent of s/util.Abort/error.Abort/
                coloffset = lastparen[2] + (lastparen[1] - t.start[1])
            else:
                coloffset = 0

        # Reset per-line attributes at EOL.
        if t.type in (token.NEWLINE, tokenize.NL):
            yield adjusttokenpos(t, coloffset)
            coldelta = 0
            coloffset = -1
            if not insideignoreblock:
                ignorenextline = (
                    tokens[i - 1].type == token.COMMENT
                    and tokens[i - 1].string == "# no-py3-transform"
                )
            continue

        if t.type == token.COMMENT:
            if t.string == "# py3-transform: off":
                insideignoreblock = True
            if t.string == "# py3-transform: on":
                insideignoreblock = False

        if ignorenextline or insideignoreblock:
            yield adjusttokenpos(t, coloffset)
            continue

        # Remember the last paren position.
        if _isop(i, '(', '[', '{'):
            parens.append(t.end + (coloffset + coldelta, tokens[i + 1].type))
        elif _isop(i, ')', ']', '}'):
            parens.pop()

        # Convert most string literals to byte literals. String literals
        # in Python 2 are bytes. String literals in Python 3 are unicode.
        # Most strings in Mercurial are bytes and unicode strings are rare.
        # Rather than rewrite all string literals to use ``b''`` to indicate
        # byte strings, we apply this token transformer to insert the ``b``
        # prefix nearly everywhere.
        if t.type == token.STRING and t not in sysstrtokens:
            s = t.string

            # Preserve docstrings as string literals. This is inconsistent
            # with regular unprefixed strings. However, the
            # "from __future__" parsing (which allows a module docstring to
            # exist before it) doesn't properly handle the docstring if it
            # is b''' prefixed, leading to a SyntaxError. We leave all
            # docstrings as unprefixed to avoid this. This means Mercurial
            # components touching docstrings need to handle unicode,
            # unfortunately.
            if s[0:3] in ("'''", '"""'):
                # If it's assigned to something, it's not a docstring
                if not _isop(i - 1, '='):
                    yield adjusttokenpos(t, coloffset)
                    continue

            # If the first character isn't a quote, it is likely a string
            # prefixing character (such as 'b', 'u', or 'r'. Ignore.
            if s[0] not in ("'", '"'):
                yield adjusttokenpos(t, coloffset)
                continue

            # String literal. Prefix to make a b'' string.
            yield adjusttokenpos(t._replace(string='b%s' % t.string),
                                 coloffset)
            coldelta += 1
            continue

        # This looks like a function call.
        if t.type == token.NAME and _isop(i + 1, '('):
            fn = t.string

            # *attr() builtins don't accept byte strings to 2nd argument.
            if fn in (
                'getattr', 'setattr', 'hasattr', 'safehasattr', 'wrapfunction',
                'wrapclass', 'addattr'
            ) and (opts['allow-attr-methods'] or not _isop(i - 1, '.')):
                arg1idx = _findargnofcall(1)
                if arg1idx is not None:
                    _ensuresysstr(arg1idx)

            # .encode() and .decode() on str/bytes/unicode don't accept
            # byte strings on Python 3.
            elif fn in ('encode', 'decode') and _isop(i - 1, '.'):
                for argn in range(2):
                    argidx = _findargnofcall(argn)
                    if argidx is not None:
                        _ensuresysstr(argidx)

            # It changes iteritems/values to items/values as they are not
            # present in Python 3 world.
            elif opts['dictiter'] and fn in ('iteritems', 'itervalues'):
                yield adjusttokenpos(t._replace(string=fn[4:]), coloffset)
                continue

        if t.type == token.NAME and t.string in opts['treat-as-kwargs']:
            if _isitemaccess(i):
                _ensuresysstr(i + 2)
            if _ismethodcall(i, 'get', 'pop', 'setdefault', 'popitem'):
                _ensuresysstr(i + 4)

        # Looks like "if __name__ == '__main__'".
        if (t.type == token.NAME and t.string == '__name__'
            and _isop(i + 1, '==')):
            _ensuresysstr(i + 2)

        # Emit unmodified token.
        yield adjusttokenpos(t, coloffset)

def process(fin, fout, opts):
    tokens = tokenize.tokenize(fin.readline)
    tokens = replacetokens(list(tokens), opts)
    fout.write(tokenize.untokenize(tokens))

def tryunlink(fname):
    try:
        os.unlink(fname)
    except OSError as err:
        if err.errno != errno.ENOENT:
            raise

@contextlib.contextmanager
def editinplace(fname):
    n = os.path.basename(fname)
    d = os.path.dirname(fname)
    fp = tempfile.NamedTemporaryFile(prefix='.%s-' % n, suffix='~', dir=d,
                                     delete=False)
    try:
        yield fp
        fp.close()
        if os.name == 'nt':
            tryunlink(fname)
        os.rename(fp.name, fname)
    finally:
        fp.close()
        tryunlink(fp.name)

def main():
    ap = argparse.ArgumentParser()
    ap.add_argument('--version', action='version',
                    version='Byteify strings 1.0')
    ap.add_argument('-i', '--inplace', action='store_true', default=False,
                    help='edit files in place')
    ap.add_argument('--dictiter', action='store_true', default=False,
                    help='rewrite iteritems() and itervalues()'),
    ap.add_argument('--allow-attr-methods', action='store_true',
                    default=False,
                    help='also handle attr*() when they are methods'),
    ap.add_argument('--treat-as-kwargs', nargs="+", default=[],
                    help="ignore kwargs-like objects"),
    ap.add_argument('files', metavar='FILE', nargs='+', help='source file')
    args = ap.parse_args()
    opts = {
        'dictiter': args.dictiter,
        'treat-as-kwargs': set(args.treat_as_kwargs),
        'allow-attr-methods': args.allow_attr_methods,
    }
    for fname in args.files:
        if args.inplace:
            with editinplace(fname) as fout:
                with open(fname, 'rb') as fin:
                    process(fin, fout, opts)
        else:
            with open(fname, 'rb') as fin:
                fout = sys.stdout.buffer
                process(fin, fout, opts)

if __name__ == '__main__':
    if sys.version_info.major < 3:
        print('This script must be run under Python 3.')
        sys.exit(3)
    main()