hgext/infinitepush/store.py
author Manuel Jacob <me@manueljacob.de>
Mon, 11 Jul 2022 01:51:20 +0200
branchstable
changeset 49378 094a5fa3cf52
parent 48946 642e31cb55f0
permissions -rw-r--r--
procutil: make stream detection in make_line_buffered more correct and strict In make_line_buffered(), we don’t want to wrap the stream if we know that lines get flushed to the underlying raw stream already. Previously, the heuristic was too optimistic. It assumed that any stream which is not an instance of io.BufferedIOBase doesn’t need wrapping. However, there are buffered streams that aren’t instances of io.BufferedIOBase, like Mercurial’s own winstdout. The new logic is different in two ways: First, only for the check, if unwraps any combination of WriteAllWrapper and winstdout. Second, it skips wrapping the stream only if it is an instance of io.RawIOBase (or already wrapped). If it is an instance of io.BufferedIOBase, it gets wrapped. In any other case, the function raises an exception. This ensures that, if an unknown stream is passed or we add another wrapper in the future, we don’t wrap the stream if it’s already line buffered or not wrap the stream if it’s not line buffered. In fact, this was already helpful during development of this change. Without it, I possibly would have forgot that WriteAllWrapper needs to be ignored for the check, leading to unnecessary wrapping if stdout is unbuffered. The alternative would have been to always wrap unknown streams. However, I don’t think that anyone would benefit from being less strict. We can expect streams from the standard library to be subclassing either io.RawIOBase or io.BufferedIOBase, so running Mercurial in the standard way should not regress by this change. Py2exe might replace sys.stdout and sys.stderr, but that currently breaks Mercurial anyway and also these streams don’t claim to be interactive, so this function is not called for them.

# This software may be used and distributed according to the terms of the
# GNU General Public License version 2 or any later version.

# based on bundleheads extension by Gregory Szorc <gps@mozilla.com>


import abc
import os
import subprocess

from mercurial.node import hex
from mercurial.pycompat import open
from mercurial import pycompat
from mercurial.utils import (
    hashutil,
    procutil,
)


class BundleWriteException(Exception):
    pass


class BundleReadException(Exception):
    pass


class abstractbundlestore:  # pytype: disable=ignored-metaclass
    """Defines the interface for bundle stores.

    A bundle store is an entity that stores raw bundle data. It is a simple
    key-value store. However, the keys are chosen by the store. The keys can
    be any Python object understood by the corresponding bundle index (see
    ``abstractbundleindex`` below).
    """

    __metaclass__ = abc.ABCMeta

    @abc.abstractmethod
    def write(self, data):
        """Write bundle data to the store.

        This function receives the raw data to be written as a str.
        Throws BundleWriteException
        The key of the written data MUST be returned.
        """

    @abc.abstractmethod
    def read(self, key):
        """Obtain bundle data for a key.

        Returns None if the bundle isn't known.
        Throws BundleReadException
        The returned object should be a file object supporting read()
        and close().
        """


class filebundlestore:
    """bundle store in filesystem

    meant for storing bundles somewhere on disk and on network filesystems
    """

    def __init__(self, ui, repo):
        self.ui = ui
        self.repo = repo
        self.storepath = ui.configpath(b'scratchbranch', b'storepath')
        if not self.storepath:
            self.storepath = self.repo.vfs.join(
                b"scratchbranches", b"filebundlestore"
            )
        if not os.path.exists(self.storepath):
            os.makedirs(self.storepath)

    def _dirpath(self, hashvalue):
        """First two bytes of the hash are the name of the upper
        level directory, next two bytes are the name of the
        next level directory"""
        return os.path.join(self.storepath, hashvalue[0:2], hashvalue[2:4])

    def _filepath(self, filename):
        return os.path.join(self._dirpath(filename), filename)

    def write(self, data):
        filename = hex(hashutil.sha1(data).digest())
        dirpath = self._dirpath(filename)

        if not os.path.exists(dirpath):
            os.makedirs(dirpath)

        with open(self._filepath(filename), b'wb') as f:
            f.write(data)

        return filename

    def read(self, key):
        try:
            with open(self._filepath(key), b'rb') as f:
                return f.read()
        except IOError:
            return None


def format_placeholders_args(args, filename=None, handle=None):
    """Formats `args` with Infinitepush replacements.

    Hack to get `str.format()`-ed strings working in a BC way with
    bytes.
    """
    formatted_args = []
    for arg in args:
        if filename and arg == b'{filename}':
            formatted_args.append(filename)
        elif handle and arg == b'{handle}':
            formatted_args.append(handle)
        else:
            formatted_args.append(arg)
    return formatted_args


class externalbundlestore(abstractbundlestore):
    def __init__(self, put_binary, put_args, get_binary, get_args):
        """
        `put_binary` - path to binary file which uploads bundle to external
            storage and prints key to stdout
        `put_args` - format string with additional args to `put_binary`
                     {filename} replacement field can be used.
        `get_binary` - path to binary file which accepts filename and key
            (in that order), downloads bundle from store and saves it to file
        `get_args` - format string with additional args to `get_binary`.
                     {filename} and {handle} replacement field can be used.
        """

        self.put_args = put_args
        self.get_args = get_args
        self.put_binary = put_binary
        self.get_binary = get_binary

    def _call_binary(self, args):
        p = subprocess.Popen(
            pycompat.rapply(procutil.tonativestr, args),
            stdout=subprocess.PIPE,
            stderr=subprocess.PIPE,
            close_fds=True,
        )
        stdout, stderr = p.communicate()
        returncode = p.returncode
        return returncode, stdout, stderr

    def write(self, data):
        # Won't work on windows because you can't open file second time without
        # closing it
        # TODO: rewrite without str.format() and replace NamedTemporaryFile()
        # with pycompat.namedtempfile()
        with pycompat.namedtempfile() as temp:
            temp.write(data)
            temp.flush()
            temp.seek(0)
            formatted_args = format_placeholders_args(
                self.put_args, filename=temp.name
            )
            returncode, stdout, stderr = self._call_binary(
                [self.put_binary] + formatted_args
            )

            if returncode != 0:
                raise BundleWriteException(
                    b'Failed to upload to external store: %s' % stderr
                )
            stdout_lines = stdout.splitlines()
            if len(stdout_lines) == 1:
                return stdout_lines[0]
            else:
                raise BundleWriteException(
                    b'Bad output from %s: %s' % (self.put_binary, stdout)
                )

    def read(self, handle):
        # Won't work on windows because you can't open file second time without
        # closing it
        with pycompat.namedtempfile() as temp:
            formatted_args = format_placeholders_args(
                self.get_args, filename=temp.name, handle=handle
            )
            returncode, stdout, stderr = self._call_binary(
                [self.get_binary] + formatted_args
            )

            if returncode != 0:
                raise BundleReadException(
                    b'Failed to download from external store: %s' % stderr
                )
            return temp.read()