comparison mercurial/encoding.py @ 48892:fa2b1a46d92e

encoding: remove Python 2 support code

Differential Revision: https://phab.mercurial-scm.org/D12295

48891:4eae533354ae

asciiupper = charencode.asciiupper
_jsonescapeu8fast = charencode.jsonescapeu8fast

_sysstr = pycompat.sysstr

if pycompat.ispy3:
    unichr = chr

# These unicode characters are ignored by HFS+ (Apple Technote 1150,
# "Unicode Subtleties"), so we need to ignore them in some places for
# sanity.
_ignore = [

    return s


# encoding.environ is provided read-only, which may not be used to modify
# the process environment
_nativeenviron = not pycompat.ispy3 or os.supports_bytes_environ
if not pycompat.ispy3:
    environ = os.environ  # re-exports
elif _nativeenviron:
    environ = os.environb  # re-exports
else:
    # preferred encoding isn't known yet; use utf-8 to avoid unicode error
    # and recreate it once encoding is settled
    environ = {

    b'ANSI_X3.4-1968': b'ascii',
}
# cp65001 is a Windows variant of utf-8, which isn't supported on Python 2.
# No idea if it should be rewritten to the canonical name 'utf-8' on Python 3.
# https://bugs.python.org/issue13216
if pycompat.iswindows and not pycompat.ispy3:
    _encodingrewrites[b'cp65001'] = b'utf-8'

try:
    encoding = environ.get(b"HGENCODING")
    if not encoding:

# converter functions between native str and byte string. use these if the
# character encoding is not aware (e.g. exception message) or is known to
# be locale dependent (e.g. date formatting.)
if pycompat.ispy3:
    strtolocal = unitolocal
    strfromlocal = unifromlocal
    strmethod = unimethod
else:

    def strtolocal(s):
        # type: (str) -> bytes
        return s  # pytype: disable=bad-return-type

    def strfromlocal(s):
        # type: (bytes) -> str
        return s  # pytype: disable=bad-return-type

    strmethod = pycompat.identity


def lower(s):
    # type: (bytes) -> bytes
    """best-effort encoding-aware case-folding of local string s"""

if not _nativeenviron:
    # now encoding and helper functions are available, recreate the environ
    # dict to be exported to other modules
    if pycompat.iswindows and pycompat.ispy3:

        class WindowsEnviron(dict):
            """`os.environ` normalizes environment variables to uppercase on windows"""

            def get(self, key, default=None):

        environ[tolocal(k.encode('utf-8'))] = tolocal(v.encode('utf-8'))


DRIVE_RE = re.compile(b'^[a-z]:')

if pycompat.ispy3:
    # os.getcwd() on Python 3 returns string, but it has os.getcwdb() which
    # returns bytes.
    if pycompat.iswindows:
        # Python 3 on Windows issues a DeprecationWarning about using the bytes
        # API when os.getcwdb() is called.
        #
        # Additionally, py3.8+ uppercases the drive letter when calling
        # os.path.realpath(), which is used on ``repo.root``.  Since those
        # strings are compared in various places as simple strings, also call
        # realpath here.  See https://bugs.python.org/issue40368
        #
        # However this is not reliable, so lets explicitly make this drive
        # letter upper case.
        #
        # note: we should consider dropping realpath here since it seems to
        # change the semantic of `getcwd`.

        def getcwd():
            cwd = os.getcwd()  # re-exports
            cwd = os.path.realpath(cwd)
            cwd = strtolocal(cwd)
            if DRIVE_RE.match(cwd):
                cwd = cwd[0:1].upper() + cwd[1:]
            return cwd

    else:
        getcwd = os.getcwdb  # re-exports
else:
    getcwd = os.getcwd  # re-exports

# How to treat ambiguous-width characters. Set to 'wide' to treat as wide.
_wide = _sysstr(
    environ.get(b"HGENCODINGAMBIGUOUS", b"narrow") == b"wide"
    and b"WFA"

    return charencodepure.jsonescapeu8fallback(u8chars, paranoid)


# We need to decode/encode U+DCxx codes transparently since invalid UTF-8
# bytes are mapped to that range.
if pycompat.ispy3:
    _utf8strict = r'surrogatepass'
else:
    _utf8strict = r'strict'

_utf8len = [0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 2, 2, 3, 4]


def getutf8char(s, pos):

48892:fa2b1a46d92e

asciiupper = charencode.asciiupper
_jsonescapeu8fast = charencode.jsonescapeu8fast

_sysstr = pycompat.sysstr

unichr = chr

# These unicode characters are ignored by HFS+ (Apple Technote 1150,
# "Unicode Subtleties"), so we need to ignore them in some places for
# sanity.
_ignore = [

    return s


# encoding.environ is provided read-only, which may not be used to modify
# the process environment
_nativeenviron = os.supports_bytes_environ
if _nativeenviron:
    environ = os.environb  # re-exports
else:
    # preferred encoding isn't known yet; use utf-8 to avoid unicode error
    # and recreate it once encoding is settled
    environ = {

    b'ANSI_X3.4-1968': b'ascii',
}
# cp65001 is a Windows variant of utf-8, which isn't supported on Python 2.
# No idea if it should be rewritten to the canonical name 'utf-8' on Python 3.
# https://bugs.python.org/issue13216
if pycompat.iswindows:
    _encodingrewrites[b'cp65001'] = b'utf-8'

try:
    encoding = environ.get(b"HGENCODING")
    if not encoding:

# converter functions between native str and byte string. use these if the
# character encoding is not aware (e.g. exception message) or is known to
# be locale dependent (e.g. date formatting.)
strtolocal = unitolocal
strfromlocal = unifromlocal
strmethod = unimethod


def lower(s):
    # type: (bytes) -> bytes
    """best-effort encoding-aware case-folding of local string s"""

if not _nativeenviron:
    # now encoding and helper functions are available, recreate the environ
    # dict to be exported to other modules
    if pycompat.iswindows:

        class WindowsEnviron(dict):
            """`os.environ` normalizes environment variables to uppercase on windows"""

            def get(self, key, default=None):

        environ[tolocal(k.encode('utf-8'))] = tolocal(v.encode('utf-8'))


DRIVE_RE = re.compile(b'^[a-z]:')

# os.getcwd() on Python 3 returns string, but it has os.getcwdb() which
# returns bytes.
if pycompat.iswindows:
    # Python 3 on Windows issues a DeprecationWarning about using the bytes
    # API when os.getcwdb() is called.
    #
    # Additionally, py3.8+ uppercases the drive letter when calling
    # os.path.realpath(), which is used on ``repo.root``.  Since those
    # strings are compared in various places as simple strings, also call
    # realpath here.  See https://bugs.python.org/issue40368
    #
    # However this is not reliable, so lets explicitly make this drive
    # letter upper case.
    #
    # note: we should consider dropping realpath here since it seems to
    # change the semantic of `getcwd`.

    def getcwd():
        cwd = os.getcwd()  # re-exports
        cwd = os.path.realpath(cwd)
        cwd = strtolocal(cwd)
        if DRIVE_RE.match(cwd):
            cwd = cwd[0:1].upper() + cwd[1:]
        return cwd


else:
    getcwd = os.getcwdb  # re-exports

# How to treat ambiguous-width characters. Set to 'wide' to treat as wide.
_wide = _sysstr(
    environ.get(b"HGENCODINGAMBIGUOUS", b"narrow") == b"wide"
    and b"WFA"

    return charencodepure.jsonescapeu8fallback(u8chars, paranoid)


# We need to decode/encode U+DCxx codes transparently since invalid UTF-8
# bytes are mapped to that range.
_utf8strict = r'surrogatepass'

_utf8len = [0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 2, 2, 3, 4]


def getutf8char(s, pos):