py3: replace pycompat.getenv with encoding.environ.get

pycompat.getenv returns os.getenvb on py3 which is not available on Windows.
This patch replaces them with encoding.environ.get and checks to ensure no
new instances of os.getenv or os.setenv are introduced.

contrib/check-code.py

@@ -1,691 +1,692
 #!/usr/bin/env python
 #
 # check-code - a style and portability checker for Mercurial
 #
 # Copyright 2010 Matt Mackall <[e-mail]>
 #
 # This software may be used and distributed according to the terms of the
 # GNU General Public License version 2 or any later version.
 
 """style and portability checker for Mercurial
 
 when a rule triggers wrong, do one of the following (prefer one from top):
  * do the work-around the rule suggests
  * doublecheck that it is a false match
  * improve the rule pattern
  * add an ignore pattern to the rule (3rd arg) which matches your good line
    (you can append a short comment and match this, like: #re-raises)
  * change the pattern to a warning and list the exception in test-check-code-hg
  * ONLY use no--check-code for skipping entire files from external sources
 """
 
 from __future__ import absolute_import, print_function
 import glob
 import keyword
 import optparse
 import os
 import re
 import sys
 if sys.version_info[0] < 3:
     opentext = open
 else:
     def opentext(f):
         return open(f, encoding='ascii')
 try:
     xrange
 except NameError:
     xrange = range
 try:
     import re2
 except ImportError:
     re2 = None
 
 def compilere(pat, multiline=False):
     if multiline:
         pat = '(?m)' + pat
     if re2:
         try:
             return re2.compile(pat)
         except re2.error:
             pass
     return re.compile(pat)
 
 # check "rules depending on implementation of repquote()" in each
 # patterns (especially pypats), before changing around repquote()
 _repquotefixedmap = {' ': ' ', '\n': '\n', '.': 'p', ':': 'q',
                      '%': '%', '\\': 'b', '*': 'A', '+': 'P', '-': 'M'}
 def _repquoteencodechr(i):
     if i > 255:
         return 'u'
     c = chr(i)
     if c in _repquotefixedmap:
         return _repquotefixedmap[c]
     if c.isalpha():
         return 'x'
     if c.isdigit():
         return 'n'
     return 'o'
 _repquotett = ''.join(_repquoteencodechr(i) for i in xrange(256))
 
 def repquote(m):
     t = m.group('text')
     t = t.translate(_repquotett)
     return m.group('quote') + t + m.group('quote')
 
 def reppython(m):
     comment = m.group('comment')
     if comment:
         l = len(comment.rstrip())
         return "#" * l + comment[l:]
     return repquote(m)
 
 def repcomment(m):
     return m.group(1) + "#" * len(m.group(2))
 
 def repccomment(m):
     t = re.sub(r"((?<=\n) )|\S", "x", m.group(2))
     return m.group(1) + t + "*/"
 
 def repcallspaces(m):
     t = re.sub(r"\n\s+", "\n", m.group(2))
     return m.group(1) + t
 
 def repinclude(m):
     return m.group(1) + "<foo>"
 
 def rephere(m):
     t = re.sub(r"\S", "x", m.group(2))
     return m.group(1) + t
 
 
 testpats = [
   [
     (r'pushd|popd', "don't use 'pushd' or 'popd', use 'cd'"),
     (r'\W\$?\(\([^\)\n]*\)\)', "don't use (()) or $(()), use 'expr'"),
     (r'grep.*-q', "don't use 'grep -q', redirect to /dev/null"),
     (r'(?<!hg )grep.* -a', "don't use 'grep -a', use in-line python"),
     (r'sed.*-i', "don't use 'sed -i', use a temporary file"),
     (r'\becho\b.*\\n', "don't use 'echo \\n', use printf"),
     (r'echo -n', "don't use 'echo -n', use printf"),
     (r'(^|\|\s*)\bwc\b[^|]*$\n(?!.*\(re\))', "filter wc output"),
     (r'head -c', "don't use 'head -c', use 'dd'"),
     (r'tail -n', "don't use the '-n' option to tail, just use '-<num>'"),
     (r'sha1sum', "don't use sha1sum, use $TESTDIR/md5sum.py"),
     (r'ls.*-\w*R', "don't use 'ls -R', use 'find'"),
     (r'printf.*[^\\]\\([1-9]|0\d)', r"don't use 'printf \NNN', use Python"),
     (r'printf.*[^\\]\\x', "don't use printf \\x, use Python"),
     (r'\$\(.*\)', "don't use $(expr), use `expr`"),
     (r'rm -rf \*', "don't use naked rm -rf, target a directory"),
     (r'(^|\|\s*)grep (-\w\s+)*[^|]*[(|]\w',
      "use egrep for extended grep syntax"),
     (r'/bin/', "don't use explicit paths for tools"),
     (r'[^\n]\Z', "no trailing newline"),
     (r'export .*=', "don't export and assign at once"),
     (r'^source\b', "don't use 'source', use '.'"),
     (r'touch -d', "don't use 'touch -d', use 'touch -t' instead"),
     (r'\bls +[^|\n-]+ +-', "options to 'ls' must come before filenames"),
     (r'[^>\n]>\s*\$HGRCPATH', "don't overwrite $HGRCPATH, append to it"),
     (r'^stop\(\)', "don't use 'stop' as a shell function name"),
     (r'(\[|\btest\b).*-e ', "don't use 'test -e', use 'test -f'"),
     (r'\[\[\s+[^\]]*\]\]', "don't use '[[ ]]', use '[ ]'"),
     (r'^alias\b.*=', "don't use alias, use a function"),
     (r'if\s*!', "don't use '!' to negate exit status"),
     (r'/dev/u?random', "don't use entropy, use /dev/zero"),
     (r'do\s*true;\s*done', "don't use true as loop body, use sleep 0"),
     (r'^( *)\t', "don't use tabs to indent"),
     (r'sed (-e )?\'(\d+|/[^/]*/)i(?!\\\n)',
      "put a backslash-escaped newline after sed 'i' command"),
     (r'^diff *-\w*[uU].*$\n(^  \$ |^$)', "prefix diff -u/-U with cmp"),
     (r'^\s+(if)? diff *-\w*[uU]', "prefix diff -u/-U with cmp"),
     (r'seq ', "don't use 'seq', use $TESTDIR/seq.py"),
     (r'\butil\.Abort\b', "directly use error.Abort"),
     (r'\|&', "don't use |&, use 2>&1"),
     (r'\w =  +\w', "only one space after = allowed"),
     (r'\bsed\b.*[^\\]\\n', "don't use 'sed ... \\n', use a \\ and a newline"),
     (r'env.*-u', "don't use 'env -u VAR', use 'unset VAR'"),
     (r'cp.* -r ', "don't use 'cp -r', use 'cp -R'"),
   ],
   # warnings
   [
     (r'^function', "don't use 'function', use old style"),
     (r'^diff.*-\w*N', "don't use 'diff -N'"),
     (r'\$PWD|\${PWD}', "don't use $PWD, use `pwd`"),
     (r'^([^"\'\n]|("[^"\n]*")|(\'[^\'\n]*\'))*\^', "^ must be quoted"),
     (r'kill (`|\$\()', "don't use kill, use killdaemons.py")
   ]
 ]
 
 testfilters = [
     (r"( *)(#([^\n]*\S)?)", repcomment),
     (r"<<(\S+)((.|\n)*?\n\1)", rephere),
 ]
 
 winglobmsg = "use (glob) to match Windows paths too"
 uprefix = r"^  \$ "
 utestpats = [
   [
     (r'^(\S.*||  [$>] \S.*)[ \t]\n', "trailing whitespace on non-output"),
     (uprefix + r'.*\|\s*sed[^|>\n]*\n',
      "use regex test output patterns instead of sed"),
     (uprefix + r'(true|exit 0)', "explicit zero exit unnecessary"),
     (uprefix + r'.*(?<!\[)\$\?', "explicit exit code checks unnecessary"),
     (uprefix + r'.*\|\| echo.*(fail|error)',
      "explicit exit code checks unnecessary"),
     (uprefix + r'set -e', "don't use set -e"),
     (uprefix + r'(\s|fi\b|done\b)', "use > for continued lines"),
     (uprefix + r'.*:\.\S*/', "x:.y in a path does not work on msys, rewrite "
      "as x://.y, or see `hg log -k msys` for alternatives", r'-\S+:\.|' #-Rxxx
      '# no-msys'), # in test-pull.t which is skipped on windows
     (r'^  saved backup bundle to \$TESTTMP.*\.hg$', winglobmsg),
     (r'^  changeset .* references (corrupted|missing) \$TESTTMP/.*[^)]$',
      winglobmsg),
     (r'^  pulling from \$TESTTMP/.*[^)]$', winglobmsg,
      '\$TESTTMP/unix-repo$'), # in test-issue1802.t which skipped on windows
     (r'^  reverting (?!subrepo ).*/.*[^)]$', winglobmsg),
     (r'^  cloning subrepo \S+/.*[^)]$', winglobmsg),
     (r'^  pushing to \$TESTTMP/.*[^)]$', winglobmsg),
     (r'^  pushing subrepo \S+/\S+ to.*[^)]$', winglobmsg),
     (r'^  moving \S+/.*[^)]$', winglobmsg),
     (r'^  no changes made to subrepo since.*/.*[^)]$', winglobmsg),
     (r'^  .*: largefile \S+ not available from file:.*/.*[^)]$', winglobmsg),
     (r'^  .*file://\$TESTTMP',
      'write "file:/*/$TESTTMP" + (glob) to match on windows too'),
     (r'^  [^$>].*27\.0\.0\.1.*[^)]$',
      'use (glob) to match localhost IP on hosts without 127.0.0.1 too'),
     (r'^  (cat|find): .*: No such file or directory',
      'use test -f to test for file existence'),
     (r'^  diff -[^ -]*p',
      "don't use (external) diff with -p for portability"),
     (r'^  [-+][-+][-+] .* [-+]0000 \(glob\)',
      "glob timezone field in diff output for portability"),
     (r'^  @@ -[0-9]+ [+][0-9]+,[0-9]+ @@',
      "use '@@ -N* +N,n @@ (glob)' style chunk header for portability"),
     (r'^  @@ -[0-9]+,[0-9]+ [+][0-9]+ @@',
      "use '@@ -N,n +N* @@ (glob)' style chunk header for portability"),
     (r'^  @@ -[0-9]+ [+][0-9]+ @@',
      "use '@@ -N* +N* @@ (glob)' style chunk header for portability"),
     (uprefix + r'hg( +-[^ ]+( +[^ ]+)?)* +extdiff'
      r'( +(-[^ po-]+|--(?!program|option)[^ ]+|[^-][^ ]*))*$',
      "use $RUNTESTDIR/pdiff via extdiff (or -o/-p for false-positives)"),
   ],
   # warnings
   [
     (r'^  (?!.*127\.0\.0\.1)[^*?/\n]* \(glob\)$',
      "glob match with no glob string (?, *, /, and 127.0.0.1)"),
   ]
 ]
 
 for i in [0, 1]:
     for tp in testpats[i]:
         p = tp[0]
         m = tp[1]
         if p.startswith(r'^'):
             p = r"^  [$>] (%s)" % p[1:]
         else:
             p = r"^  [$>] .*(%s)" % p
         utestpats[i].append((p, m) + tp[2:])
 
 utestfilters = [
     (r"<<(\S+)((.|\n)*?\n  > \1)", rephere),
     (r"( +)(#([^\n]*\S)?)", repcomment),
 ]
 
 pypats = [
   [
     (r'^\s*def\s*\w+\s*\(.*,\s*\(',
      "tuple parameter unpacking not available in Python 3+"),
     (r'lambda\s*\(.*,.*\)',
      "tuple parameter unpacking not available in Python 3+"),
     (r'(?<!def)\s+(cmp)\(', "cmp is not available in Python 3+"),
     (r'\breduce\s*\(.*', "reduce is not available in Python 3+"),
     (r'\bdict\(.*=', 'dict() is different in Py2 and 3 and is slower than {}',
      'dict-from-generator'),
     (r'\.has_key\b', "dict.has_key is not available in Python 3+"),
     (r'\s<>\s', '<> operator is not available in Python 3+, use !='),
     (r'^\s*\t', "don't use tabs"),
     (r'\S;\s*\n', "semicolon"),
     (r'[^_]_\([ \t\n]*(?:"[^"]+"[ \t\n+]*)+%', "don't use % inside _()"),
     (r"[^_]_\([ \t\n]*(?:'[^']+'[ \t\n+]*)+%", "don't use % inside _()"),
     (r'(\w|\)),\w', "missing whitespace after ,"),
     (r'(\w|\))[+/*\-<>]\w', "missing whitespace in expression"),
     (r'^\s+(\w|\.)+=\w[^,()\n]*$', "missing whitespace in assignment"),
     (r'\w\s=\s\s+\w', "gratuitous whitespace after ="),
     (r'.{81}', "line too long"),
     (r'[^\n]\Z', "no trailing newline"),
     (r'(\S[ \t]+|^[ \t]+)\n', "trailing whitespace"),
 #    (r'^\s+[^_ \n][^_. \n]+_[^_\n]+\s*=',
 #     "don't use underbars in identifiers"),
     (r'^\s+(self\.)?[A-za-z][a-z0-9]+[A-Z]\w* = ',
      "don't use camelcase in identifiers"),
     (r'^\s*(if|while|def|class|except|try)\s[^[\n]*:\s*[^\\n]#\s]+',
      "linebreak after :"),
     (r'class\s[^( \n]+:', "old-style class, use class foo(object)",
      r'#.*old-style'),
     (r'class\s[^( \n]+\(\):',
      "class foo() creates old style object, use class foo(object)",
      r'#.*old-style'),
     (r'\b(%s)\(' % '|'.join(k for k in keyword.kwlist
                             if k not in ('print', 'exec')),
      "Python keyword is not a function"),
     (r',]', "unneeded trailing ',' in list"),
 #    (r'class\s[A-Z][^\(]*\((?!Exception)',
 #     "don't capitalize non-exception classes"),
 #    (r'in range\(', "use xrange"),
 #    (r'^\s*print\s+', "avoid using print in core and extensions"),
     (r'[\x80-\xff]', "non-ASCII character literal"),
     (r'("\')\.format\(', "str.format() has no bytes counterpart, use %"),
     (r'^\s*(%s)\s\s' % '|'.join(keyword.kwlist),
      "gratuitous whitespace after Python keyword"),
     (r'([\(\[][ \t]\S)|(\S[ \t][\)\]])', "gratuitous whitespace in () or []"),
 #    (r'\s\s=', "gratuitous whitespace before ="),
     (r'[^>< ](\+=|-=|!=|<>|<=|>=|<<=|>>=|%=)\S',
      "missing whitespace around operator"),
     (r'[^>< ](\+=|-=|!=|<>|<=|>=|<<=|>>=|%=)\s',
      "missing whitespace around operator"),
     (r'\s(\+=|-=|!=|<>|<=|>=|<<=|>>=|%=)\S',
      "missing whitespace around operator"),
     (r'[^^+=*/!<>&| %-](\s=|=\s)[^= ]',
      "wrong whitespace around ="),
     (r'\([^()]*( =[^=]|[^<>!=]= )',
      "no whitespace around = for named parameters"),
     (r'raise Exception', "don't raise generic exceptions"),
     (r'raise [^,(]+, (\([^\)]+\)|[^,\(\)]+)$',
      "don't use old-style two-argument raise, use Exception(message)"),
     (r' is\s+(not\s+)?["\'0-9-]', "object comparison with literal"),
     (r' [=!]=\s+(True|False|None)',
      "comparison with singleton, use 'is' or 'is not' instead"),
     (r'^\s*(while|if) [01]:',
      "use True/False for constant Boolean expression"),
     (r'(?:(?<!def)\s+|\()hasattr\(',
      'hasattr(foo, bar) is broken, use util.safehasattr(foo, bar) instead'),
     (r'opener\([^)]*\).read\(',
      "use opener.read() instead"),
     (r'opener\([^)]*\).write\(',
      "use opener.write() instead"),
     (r'[\s\(](open|file)\([^)]*\)\.read\(',
      "use util.readfile() instead"),
     (r'[\s\(](open|file)\([^)]*\)\.write\(',
      "use util.writefile() instead"),
     (r'^[\s\(]*(open(er)?|file)\([^)]*\)',
      "always assign an opened file to a variable, and close it afterwards"),
     (r'[\s\(](open|file)\([^)]*\)\.',
      "always assign an opened file to a variable, and close it afterwards"),
     (r'(?i)descend[e]nt', "the proper spelling is descendAnt"),
     (r'\.debug\(\_', "don't mark debug messages for translation"),
     (r'\.strip\(\)\.split\(\)', "no need to strip before splitting"),
     (r'^\s*except\s*:', "naked except clause", r'#.*re-raises'),
     (r'^\s*except\s([^\(,]+|\([^\)]+\))\s*,',
      'legacy exception syntax; use "as" instead of ","'),
     (r':\n(    )*( ){1,3}[^ ]', "must indent 4 spaces"),
     (r'release\(.*wlock, .*lock\)', "wrong lock release order"),
     (r'\b__bool__\b', "__bool__ should be __nonzero__ in Python 2"),
     (r'os\.path\.join\(.*, *(""|\'\')\)',
      "use pathutil.normasprefix(path) instead of os.path.join(path, '')"),
     (r'\s0[0-7]+\b', 'legacy octal syntax; use "0o" prefix instead of "0"'),
     # XXX only catch mutable arguments on the first line of the definition
     (r'def.*[( ]\w+=\{\}', "don't use mutable default arguments"),
     (r'\butil\.Abort\b', "directly use error.Abort"),
     (r'^@(\w*\.)?cachefunc', "module-level @cachefunc is risky, please avoid"),
     (r'^import Queue', "don't use Queue, use util.queue + util.empty"),
     (r'^import cStringIO', "don't use cStringIO.StringIO, use util.stringio"),
     (r'^import urllib', "don't use urllib, use util.urlreq/util.urlerr"),
     (r'^import SocketServer', "don't use SockerServer, use util.socketserver"),
     (r'^import urlparse', "don't use urlparse, use util.urlparse"),
     (r'^import xmlrpclib', "don't use xmlrpclib, use util.xmlrpclib"),
     (r'^import cPickle', "don't use cPickle, use util.pickle"),
     (r'^import pickle', "don't use pickle, use util.pickle"),
     (r'^import httplib', "don't use httplib, use util.httplib"),
     (r'^import BaseHTTPServer', "use util.httpserver instead"),
     (r'\.next\(\)', "don't use .next(), use next(...)"),
 
     # rules depending on implementation of repquote()
     (r' x+[xpqo%APM][\'"]\n\s+[\'"]x',
      'string join across lines with no space'),
     (r'''(?x)ui\.(status|progress|write|note|warn)\(
          [ \t\n#]*
          (?# any strings/comments might precede a string, which
            # contains translatable message)
          ((['"]|\'\'\'|""")[ \npq%bAPMxno]*(['"]|\'\'\'|""")[ \t\n#]+)*
          (?# sequence consisting of below might precede translatable message
            # - formatting string: "% 10s", "%05d", "% -3.2f", "%*s", "%%" ...
            # - escaped character: "\\", "\n", "\0" ...
            # - character other than '%', 'b' as '\', and 'x' as alphabet)
          (['"]|\'\'\'|""")
          ((%([ n]?[PM]?([np]+|A))?x)|%%|b[bnx]|[ \nnpqAPMo])*x
          (?# this regexp can't use [^...] style,
            # because _preparepats forcibly adds "\n" into [^...],
            # even though this regexp wants match it against "\n")''',
      "missing _() in ui message (use () to hide false-positives)"),
   ],
   # warnings
   [
     # rules depending on implementation of repquote()
     (r'(^| )pp +xxxxqq[ \n][^\n]', "add two newlines after '.. note::'"),
   ]
 ]
 
 pyfilters = [
     (r"""(?msx)(?P<comment>\#.*?$)|
          ((?P<quote>('''|\"\"\"|(?<!')'(?!')|(?<!")"(?!")))
           (?P<text>(([^\\]|\\.)*?))
           (?P=quote))""", reppython),
 ]
 
 txtfilters = []
 
 txtpats = [
   [
     ('\s$', 'trailing whitespace'),
     ('.. note::[ \n][^\n]', 'add two newlines after note::')
   ],
   []
 ]
 
 cpats = [
   [
     (r'//', "don't use //-style comments"),
     (r'^  ', "don't use spaces to indent"),
     (r'\S\t', "don't use tabs except for indent"),
     (r'(\S[ \t]+|^[ \t]+)\n', "trailing whitespace"),
     (r'.{81}', "line too long"),
     (r'(while|if|do|for)\(', "use space after while/if/do/for"),
     (r'return\(', "return is not a function"),
     (r' ;', "no space before ;"),
     (r'[^;] \)', "no space before )"),
     (r'[)][{]', "space between ) and {"),
     (r'\w+\* \w+', "use int *foo, not int* foo"),
     (r'\W\([^\)]+\) \w+', "use (int)foo, not (int) foo"),
     (r'\w+ (\+\+|--)', "use foo++, not foo ++"),
     (r'\w,\w', "missing whitespace after ,"),
     (r'^[^#]\w[+/*]\w', "missing whitespace in expression"),
     (r'\w\s=\s\s+\w', "gratuitous whitespace after ="),
     (r'^#\s+\w', "use #foo, not # foo"),
     (r'[^\n]\Z', "no trailing newline"),
     (r'^\s*#import\b', "use only #include in standard C code"),
     (r'strcpy\(', "don't use strcpy, use strlcpy or memcpy"),
     (r'strcat\(', "don't use strcat"),
 
     # rules depending on implementation of repquote()
   ],
   # warnings
   [
     # rules depending on implementation of repquote()
   ]
 ]
 
 cfilters = [
     (r'(/\*)(((\*(?!/))|[^*])*)\*/', repccomment),
     (r'''(?P<quote>(?<!")")(?P<text>([^"]|\\")+)"(?!")''', repquote),
     (r'''(#\s*include\s+<)([^>]+)>''', repinclude),
     (r'(\()([^)]+\))', repcallspaces),
 ]
 
 inutilpats = [
   [
     (r'\bui\.', "don't use ui in util"),
   ],
   # warnings
   []
 ]
 
 inrevlogpats = [
   [
     (r'\brepo\.', "don't use repo in revlog"),
   ],
   # warnings
   []
 ]
 
 webtemplatefilters = []
 
 webtemplatepats = [
   [],
   [
     (r'{desc(\|(?!websub|firstline)[^\|]*)+}',
      'follow desc keyword with either firstline or websub'),
   ]
 ]
 
 allfilesfilters = []
 
 allfilespats = [
   [
     (r'(http|https)://[a-zA-Z0-9./]*selenic.com/',
      'use mercurial-scm.org domain URL'),
   ],
   # warnings
   [],
 ]
 
 py3pats = [
   [
     (r'os\.environ', "use encoding.environ instead (py3)"),
     (r'os\.name', "use pycompat.osname instead (py3)"),
     (r'os\.getcwd', "use pycompat.getcwd instead (py3)"),
     (r'os\.sep', "use pycompat.ossep instead (py3)"),
     (r'os\.pathsep', "use pycompat.ospathsep instead (py3)"),
     (r'os\.altsep', "use pycompat.osaltsep instead (py3)"),
-    (r'os\.getenv', "use pycompat.osgetenv instead (py3)"),
     (r'sys\.platform', "use pycompat.sysplatform instead (py3)"),
     (r'getopt\.getopt', "use pycompat.getoptb instead (py3)"),
+    (r'os\.getenv', "use encoding.environ.get instead"),
+    (r'os\.setenv', "modifying the environ dict is not preferred"),
   ],
   # warnings
   [],
 ]
 
 checks = [
     ('python', r'.*\.(py|cgi)$', r'^#!.*python', pyfilters, pypats),
     ('python 3', r'.*(hgext|mercurial).*(?<!pycompat)\.py', '',
             pyfilters, py3pats),
     ('test script', r'(.*/)?test-[^.~]*$', '', testfilters, testpats),
     ('c', r'.*\.[ch]$', '', cfilters, cpats),
     ('unified test', r'.*\.t$', '', utestfilters, utestpats),
     ('layering violation repo in revlog', r'mercurial/revlog\.py', '',
      pyfilters, inrevlogpats),
     ('layering violation ui in util', r'mercurial/util\.py', '', pyfilters,
      inutilpats),
     ('txt', r'.*\.txt$', '', txtfilters, txtpats),
     ('web template', r'mercurial/templates/.*\.tmpl', '',
      webtemplatefilters, webtemplatepats),
     ('all except for .po', r'.*(?<!\.po)$', '',
      allfilesfilters, allfilespats),
 ]
 
 def _preparepats():
     for c in checks:
         failandwarn = c[-1]
         for pats in failandwarn:
             for i, pseq in enumerate(pats):
                 # fix-up regexes for multi-line searches
                 p = pseq[0]
                 # \s doesn't match \n
                 p = re.sub(r'(?<!\\)\\s', r'[ \\t]', p)
                 # [^...] doesn't match newline
                 p = re.sub(r'(?<!\\)\[\^', r'[^\\n', p)
 
                 pats[i] = (re.compile(p, re.MULTILINE),) + pseq[1:]
         filters = c[3]
         for i, flt in enumerate(filters):
             filters[i] = re.compile(flt[0]), flt[1]
 
 class norepeatlogger(object):
     def __init__(self):
         self._lastseen = None
 
     def log(self, fname, lineno, line, msg, blame):
         """print error related a to given line of a given file.
 
         The faulty line will also be printed but only once in the case
         of multiple errors.
 
         :fname: filename
         :lineno: line number
         :line: actual content of the line
         :msg: error message
         """
         msgid = fname, lineno, line
         if msgid != self._lastseen:
             if blame:
                 print("%s:%d (%s):" % (fname, lineno, blame))
             else:
                 print("%s:%d:" % (fname, lineno))
             print(" > %s" % line)
             self._lastseen = msgid
         print(" " + msg)
 
 _defaultlogger = norepeatlogger()
 
 def getblame(f):
     lines = []
     for l in os.popen('hg annotate -un %s' % f):
         start, line = l.split(':', 1)
         user, rev = start.split()
         lines.append((line[1:-1], user, rev))
     return lines
 
 def checkfile(f, logfunc=_defaultlogger.log, maxerr=None, warnings=False,
               blame=False, debug=False, lineno=True):
     """checks style and portability of a given file
 
     :f: filepath
     :logfunc: function used to report error
               logfunc(filename, linenumber, linecontent, errormessage)
     :maxerr: number of error to display before aborting.
              Set to false (default) to report all errors
 
     return True if no error is found, False otherwise.
     """
     blamecache = None
     result = True
 
     try:
         with opentext(f) as fp:
             try:
                 pre = post = fp.read()
             except UnicodeDecodeError as e:
                 print("%s while reading %s" % (e, f))
                 return result
     except IOError as e:
         print("Skipping %s, %s" % (f, str(e).split(':', 1)[0]))
         return result
 
     for name, match, magic, filters, pats in checks:
         post = pre # discard filtering result of previous check
         if debug:
             print(name, f)
         fc = 0
         if not (re.match(match, f) or (magic and re.search(magic, pre))):
             if debug:
                 print("Skipping %s for %s it doesn't match %s" % (
                        name, match, f))
             continue
         if "no-" "check-code" in pre:
             # If you're looking at this line, it's because a file has:
             # no- check- code
             # but the reason to output skipping is to make life for
             # tests easier. So, instead of writing it with a normal
             # spelling, we write it with the expected spelling from
             # tests/test-check-code.t
             print("Skipping %s it has no-che?k-code (glob)" % f)
             return "Skip" # skip checking this file
         for p, r in filters:
             post = re.sub(p, r, post)
         nerrs = len(pats[0]) # nerr elements are errors
         if warnings:
             pats = pats[0] + pats[1]
         else:
             pats = pats[0]
         # print post # uncomment to show filtered version
 
         if debug:
             print("Checking %s for %s" % (name, f))
 
         prelines = None
         errors = []
         for i, pat in enumerate(pats):
             if len(pat) == 3:
                 p, msg, ignore = pat
             else:
                 p, msg = pat
                 ignore = None
             if i >= nerrs:
                 msg = "warning: " + msg
 
             pos = 0
             n = 0
             for m in p.finditer(post):
                 if prelines is None:
                     prelines = pre.splitlines()
                     postlines = post.splitlines(True)
 
                 start = m.start()
                 while n < len(postlines):
                     step = len(postlines[n])
                     if pos + step > start:
                         break
                     pos += step
                     n += 1
                 l = prelines[n]
 
                 if ignore and re.search(ignore, l, re.MULTILINE):
                     if debug:
                         print("Skipping %s for %s:%s (ignore pattern)" % (
                             name, f, n))
                     continue
                 bd = ""
                 if blame:
                     bd = 'working directory'
                     if not blamecache:
                         blamecache = getblame(f)
                     if n < len(blamecache):
                         bl, bu, br = blamecache[n]
                         if bl == l:
                             bd = '%s@%s' % (bu, br)
 
                 errors.append((f, lineno and n + 1, l, msg, bd))
                 result = False
 
         errors.sort()
         for e in errors:
             logfunc(*e)
             fc += 1
             if maxerr and fc >= maxerr:
                 print(" (too many errors, giving up)")
                 break
 
     return result
 
 def main():
     parser = optparse.OptionParser("%prog [options] [files]")
     parser.add_option("-w", "--warnings", action="store_true",
                       help="include warning-level checks")
     parser.add_option("-p", "--per-file", type="int",
                       help="max warnings per file")
     parser.add_option("-b", "--blame", action="store_true",
                       help="use annotate to generate blame info")
     parser.add_option("", "--debug", action="store_true",
                       help="show debug information")
     parser.add_option("", "--nolineno", action="store_false",
                       dest='lineno', help="don't show line numbers")
 
     parser.set_defaults(per_file=15, warnings=False, blame=False, debug=False,
                         lineno=True)
     (options, args) = parser.parse_args()
 
     if len(args) == 0:
         check = glob.glob("*")
     else:
         check = args
 
     _preparepats()
 
     ret = 0
     for f in check:
         if not checkfile(f, maxerr=options.per_file, warnings=options.warnings,
                          blame=options.blame, debug=options.debug,
                          lineno=options.lineno):
             ret = 1
     return ret
 
 if __name__ == "__main__":
     sys.exit(main())

hgext/largefiles/lfutil.py

@@ -1,665 +1,666
 # Copyright 2009-2010 Gregory P. Ward
 # Copyright 2009-2010 Intelerad Medical Systems Incorporated
 # Copyright 2010-2011 Fog Creek Software
 # Copyright 2010-2011 Unity Technologies
 #
 # This software may be used and distributed according to the terms of the
 # GNU General Public License version 2 or any later version.
 
 '''largefiles utility code: must not import other modules in this package.'''
 from __future__ import absolute_import
 
 import copy
 import hashlib
 import os
 import platform
 import stat
 
 from mercurial.i18n import _
 
 from mercurial import (
     dirstate,
+    encoding,
     error,
     httpconnection,
     match as matchmod,
     node,
     pycompat,
     scmutil,
     util,
 )
 
 shortname = '.hglf'
 shortnameslash = shortname + '/'
 longname = 'largefiles'
 
 # -- Private worker functions ------------------------------------------
 
 def getminsize(ui, assumelfiles, opt, default=10):
     lfsize = opt
     if not lfsize and assumelfiles:
         lfsize = ui.config(longname, 'minsize', default=default)
     if lfsize:
         try:
             lfsize = float(lfsize)
         except ValueError:
             raise error.Abort(_('largefiles: size must be number (not %s)\n')
                              % lfsize)
     if lfsize is None:
         raise error.Abort(_('minimum size for largefiles must be specified'))
     return lfsize
 
 def link(src, dest):
     """Try to create hardlink - if that fails, efficiently make a copy."""
     util.makedirs(os.path.dirname(dest))
     try:
         util.oslink(src, dest)
     except OSError:
         # if hardlinks fail, fallback on atomic copy
         with open(src, 'rb') as srcf:
             with util.atomictempfile(dest) as dstf:
                 for chunk in util.filechunkiter(srcf):
                     dstf.write(chunk)
         os.chmod(dest, os.stat(src).st_mode)
 
 def usercachepath(ui, hash):
     '''Return the correct location in the "global" largefiles cache for a file
     with the given hash.
     This cache is used for sharing of largefiles across repositories - both
     to preserve download bandwidth and storage space.'''
     return os.path.join(_usercachedir(ui), hash)
 
 def _usercachedir(ui):
     '''Return the location of the "global" largefiles cache.'''
     path = ui.configpath(longname, 'usercache', None)
     if path:
         return path
     if pycompat.osname == 'nt':
-        appdata = pycompat.osgetenv('LOCALAPPDATA',\
-                        pycompat.osgetenv('APPDATA'))
+        appdata = encoding.environ.get('LOCALAPPDATA',\
+                        encoding.environ.get('APPDATA'))
         if appdata:
             return os.path.join(appdata, longname)
     elif platform.system() == 'Darwin':
-        home = pycompat.osgetenv('HOME')
+        home = encoding.environ.get('HOME')
         if home:
             return os.path.join(home, 'Library', 'Caches', longname)
     elif pycompat.osname == 'posix':
-        path = pycompat.osgetenv('XDG_CACHE_HOME')
+        path = encoding.environ.get('XDG_CACHE_HOME')
         if path:
             return os.path.join(path, longname)
-        home = pycompat.osgetenv('HOME')
+        home = encoding.environ.get('HOME')
         if home:
             return os.path.join(home, '.cache', longname)
     else:
         raise error.Abort(_('unknown operating system: %s\n')
                           % pycompat.osname)
     raise error.Abort(_('unknown %s usercache location') % longname)
 
 def inusercache(ui, hash):
     path = usercachepath(ui, hash)
     return os.path.exists(path)
 
 def findfile(repo, hash):
     '''Return store path of the largefile with the specified hash.
     As a side effect, the file might be linked from user cache.
     Return None if the file can't be found locally.'''
     path, exists = findstorepath(repo, hash)
     if exists:
         repo.ui.note(_('found %s in store\n') % hash)
         return path
     elif inusercache(repo.ui, hash):
         repo.ui.note(_('found %s in system cache\n') % hash)
         path = storepath(repo, hash)
         link(usercachepath(repo.ui, hash), path)
         return path
     return None
 
 class largefilesdirstate(dirstate.dirstate):
     def __getitem__(self, key):
         return super(largefilesdirstate, self).__getitem__(unixpath(key))
     def normal(self, f):
         return super(largefilesdirstate, self).normal(unixpath(f))
     def remove(self, f):
         return super(largefilesdirstate, self).remove(unixpath(f))
     def add(self, f):
         return super(largefilesdirstate, self).add(unixpath(f))
     def drop(self, f):
         return super(largefilesdirstate, self).drop(unixpath(f))
     def forget(self, f):
         return super(largefilesdirstate, self).forget(unixpath(f))
     def normallookup(self, f):
         return super(largefilesdirstate, self).normallookup(unixpath(f))
     def _ignore(self, f):
         return False
     def write(self, tr=False):
         # (1) disable PENDING mode always
         #     (lfdirstate isn't yet managed as a part of the transaction)
         # (2) avoid develwarn 'use dirstate.write with ....'
         super(largefilesdirstate, self).write(None)
 
 def openlfdirstate(ui, repo, create=True):
     '''
     Return a dirstate object that tracks largefiles: i.e. its root is
     the repo root, but it is saved in .hg/largefiles/dirstate.
     '''
     vfs = repo.vfs
     lfstoredir = longname
     opener = scmutil.opener(vfs.join(lfstoredir))
     lfdirstate = largefilesdirstate(opener, ui, repo.root,
                                      repo.dirstate._validate)
 
     # If the largefiles dirstate does not exist, populate and create
     # it. This ensures that we create it on the first meaningful
     # largefiles operation in a new clone.
     if create and not vfs.exists(vfs.join(lfstoredir, 'dirstate')):
         matcher = getstandinmatcher(repo)
         standins = repo.dirstate.walk(matcher, [], False, False)
 
         if len(standins) > 0:
             vfs.makedirs(lfstoredir)
 
         for standin in standins:
             lfile = splitstandin(standin)
             lfdirstate.normallookup(lfile)
     return lfdirstate
 
 def lfdirstatestatus(lfdirstate, repo):
     wctx = repo['.']
     match = matchmod.always(repo.root, repo.getcwd())
     unsure, s = lfdirstate.status(match, [], False, False, False)
     modified, clean = s.modified, s.clean
     for lfile in unsure:
         try:
             fctx = wctx[standin(lfile)]
         except LookupError:
             fctx = None
         if not fctx or fctx.data().strip() != hashfile(repo.wjoin(lfile)):
             modified.append(lfile)
         else:
             clean.append(lfile)
             lfdirstate.normal(lfile)
     return s
 
 def listlfiles(repo, rev=None, matcher=None):
     '''return a list of largefiles in the working copy or the
     specified changeset'''
 
     if matcher is None:
         matcher = getstandinmatcher(repo)
 
     # ignore unknown files in working directory
     return [splitstandin(f)
             for f in repo[rev].walk(matcher)
             if rev is not None or repo.dirstate[f] != '?']
 
 def instore(repo, hash, forcelocal=False):
     '''Return true if a largefile with the given hash exists in the store'''
     return os.path.exists(storepath(repo, hash, forcelocal))
 
 def storepath(repo, hash, forcelocal=False):
     '''Return the correct location in the repository largefiles store for a
     file with the given hash.'''
     if not forcelocal and repo.shared():
         return repo.vfs.reljoin(repo.sharedpath, longname, hash)
     return repo.join(longname, hash)
 
 def findstorepath(repo, hash):
     '''Search through the local store path(s) to find the file for the given
     hash.  If the file is not found, its path in the primary store is returned.
     The return value is a tuple of (path, exists(path)).
     '''
     # For shared repos, the primary store is in the share source.  But for
     # backward compatibility, force a lookup in the local store if it wasn't
     # found in the share source.
     path = storepath(repo, hash, False)
 
     if instore(repo, hash):
         return (path, True)
     elif repo.shared() and instore(repo, hash, True):
         return storepath(repo, hash, True), True
 
     return (path, False)
 
 def copyfromcache(repo, hash, filename):
     '''Copy the specified largefile from the repo or system cache to
     filename in the repository. Return true on success or false if the
     file was not found in either cache (which should not happened:
     this is meant to be called only after ensuring that the needed
     largefile exists in the cache).'''
     wvfs = repo.wvfs
     path = findfile(repo, hash)
     if path is None:
         return False
     wvfs.makedirs(wvfs.dirname(wvfs.join(filename)))
     # The write may fail before the file is fully written, but we
     # don't use atomic writes in the working copy.
     with open(path, 'rb') as srcfd:
         with wvfs(filename, 'wb') as destfd:
             gothash = copyandhash(
                 util.filechunkiter(srcfd), destfd)
     if gothash != hash:
         repo.ui.warn(_('%s: data corruption in %s with hash %s\n')
                      % (filename, path, gothash))
         wvfs.unlink(filename)
         return False
     return True
 
 def copytostore(repo, rev, file, uploaded=False):
     wvfs = repo.wvfs
     hash = readstandin(repo, file, rev)
     if instore(repo, hash):
         return
     if wvfs.exists(file):
         copytostoreabsolute(repo, wvfs.join(file), hash)
     else:
         repo.ui.warn(_("%s: largefile %s not available from local store\n") %
                      (file, hash))
 
 def copyalltostore(repo, node):
     '''Copy all largefiles in a given revision to the store'''
 
     ctx = repo[node]
     for filename in ctx.files():
         if isstandin(filename) and filename in ctx.manifest():
             realfile = splitstandin(filename)
             copytostore(repo, ctx.node(), realfile)
 
 def copytostoreabsolute(repo, file, hash):
     if inusercache(repo.ui, hash):
         link(usercachepath(repo.ui, hash), storepath(repo, hash))
     else:
         util.makedirs(os.path.dirname(storepath(repo, hash)))
         with open(file, 'rb') as srcf:
             with util.atomictempfile(storepath(repo, hash),
                                      createmode=repo.store.createmode) as dstf:
                 for chunk in util.filechunkiter(srcf):
                     dstf.write(chunk)
         linktousercache(repo, hash)
 
 def linktousercache(repo, hash):
     '''Link / copy the largefile with the specified hash from the store
     to the cache.'''
     path = usercachepath(repo.ui, hash)
     link(storepath(repo, hash), path)
 
 def getstandinmatcher(repo, rmatcher=None):
     '''Return a match object that applies rmatcher to the standin directory'''
     wvfs = repo.wvfs
     standindir = shortname
 
     # no warnings about missing files or directories
     badfn = lambda f, msg: None
 
     if rmatcher and not rmatcher.always():
         pats = [wvfs.join(standindir, pat) for pat in rmatcher.files()]
         if not pats:
             pats = [wvfs.join(standindir)]
         match = scmutil.match(repo[None], pats, badfn=badfn)
         # if pats is empty, it would incorrectly always match, so clear _always
         match._always = False
     else:
         # no patterns: relative to repo root
         match = scmutil.match(repo[None], [wvfs.join(standindir)], badfn=badfn)
     return match
 
 def composestandinmatcher(repo, rmatcher):
     '''Return a matcher that accepts standins corresponding to the
     files accepted by rmatcher. Pass the list of files in the matcher
     as the paths specified by the user.'''
     smatcher = getstandinmatcher(repo, rmatcher)
     isstandin = smatcher.matchfn
     def composedmatchfn(f):
         return isstandin(f) and rmatcher.matchfn(splitstandin(f))
     smatcher.matchfn = composedmatchfn
 
     return smatcher
 
 def standin(filename):
     '''Return the repo-relative path to the standin for the specified big
     file.'''
     # Notes:
     # 1) Some callers want an absolute path, but for instance addlargefiles
     #    needs it repo-relative so it can be passed to repo[None].add().  So
     #    leave it up to the caller to use repo.wjoin() to get an absolute path.
     # 2) Join with '/' because that's what dirstate always uses, even on
     #    Windows. Change existing separator to '/' first in case we are
     #    passed filenames from an external source (like the command line).
     return shortnameslash + util.pconvert(filename)
 
 def isstandin(filename):
     '''Return true if filename is a big file standin. filename must be
     in Mercurial's internal form (slash-separated).'''
     return filename.startswith(shortnameslash)
 
 def splitstandin(filename):
     # Split on / because that's what dirstate always uses, even on Windows.
     # Change local separator to / first just in case we are passed filenames
     # from an external source (like the command line).
     bits = util.pconvert(filename).split('/', 1)
     if len(bits) == 2 and bits[0] == shortname:
         return bits[1]
     else:
         return None
 
 def updatestandin(repo, standin):
     file = repo.wjoin(splitstandin(standin))
     if repo.wvfs.exists(splitstandin(standin)):
         hash = hashfile(file)
         executable = getexecutable(file)
         writestandin(repo, standin, hash, executable)
     else:
         raise error.Abort(_('%s: file not found!') % splitstandin(standin))
 
 def readstandin(repo, filename, node=None):
     '''read hex hash from standin for filename at given node, or working
     directory if no node is given'''
     return repo[node][standin(filename)].data().strip()
 
 def writestandin(repo, standin, hash, executable):
     '''write hash to <repo.root>/<standin>'''
     repo.wwrite(standin, hash + '\n', executable and 'x' or '')
 
 def copyandhash(instream, outfile):
     '''Read bytes from instream (iterable) and write them to outfile,
     computing the SHA-1 hash of the data along the way. Return the hash.'''
     hasher = hashlib.sha1('')
     for data in instream:
         hasher.update(data)
         outfile.write(data)
     return hasher.hexdigest()
 
 def hashrepofile(repo, file):
     return hashfile(repo.wjoin(file))
 
 def hashfile(file):
     if not os.path.exists(file):
         return ''
     hasher = hashlib.sha1('')
     with open(file, 'rb') as fd:
         for data in util.filechunkiter(fd):
             hasher.update(data)
     return hasher.hexdigest()
 
 def getexecutable(filename):
     mode = os.stat(filename).st_mode
     return ((mode & stat.S_IXUSR) and
             (mode & stat.S_IXGRP) and
             (mode & stat.S_IXOTH))
 
 def urljoin(first, second, *arg):
     def join(left, right):
         if not left.endswith('/'):
             left += '/'
         if right.startswith('/'):
             right = right[1:]
         return left + right
 
     url = join(first, second)
     for a in arg:
         url = join(url, a)
     return url
 
 def hexsha1(data):
     """hexsha1 returns the hex-encoded sha1 sum of the data in the file-like
     object data"""
     h = hashlib.sha1()
     for chunk in util.filechunkiter(data):
         h.update(chunk)
     return h.hexdigest()
 
 def httpsendfile(ui, filename):
     return httpconnection.httpsendfile(ui, filename, 'rb')
 
 def unixpath(path):
     '''Return a version of path normalized for use with the lfdirstate.'''
     return util.pconvert(os.path.normpath(path))
 
 def islfilesrepo(repo):
     '''Return true if the repo is a largefile repo.'''
     if ('largefiles' in repo.requirements and
             any(shortnameslash in f[0] for f in repo.store.datafiles())):
         return True
 
     return any(openlfdirstate(repo.ui, repo, False))
 
 class storeprotonotcapable(Exception):
     def __init__(self, storetypes):
         self.storetypes = storetypes
 
 def getstandinsstate(repo):
     standins = []
     matcher = getstandinmatcher(repo)
     for standin in repo.dirstate.walk(matcher, [], False, False):
         lfile = splitstandin(standin)
         try:
             hash = readstandin(repo, lfile)
         except IOError:
             hash = None
         standins.append((lfile, hash))
     return standins
 
 def synclfdirstate(repo, lfdirstate, lfile, normallookup):
     lfstandin = standin(lfile)
     if lfstandin in repo.dirstate:
         stat = repo.dirstate._map[lfstandin]
         state, mtime = stat[0], stat[3]
     else:
         state, mtime = '?', -1
     if state == 'n':
         if (normallookup or mtime < 0 or
             not repo.wvfs.exists(lfile)):
             # state 'n' doesn't ensure 'clean' in this case
             lfdirstate.normallookup(lfile)
         else:
             lfdirstate.normal(lfile)
     elif state == 'm':
         lfdirstate.normallookup(lfile)
     elif state == 'r':
         lfdirstate.remove(lfile)
     elif state == 'a':
         lfdirstate.add(lfile)
     elif state == '?':
         lfdirstate.drop(lfile)
 
 def markcommitted(orig, ctx, node):
     repo = ctx.repo()
 
     orig(node)
 
     # ATTENTION: "ctx.files()" may differ from "repo[node].files()"
     # because files coming from the 2nd parent are omitted in the latter.
     #
     # The former should be used to get targets of "synclfdirstate",
     # because such files:
     # - are marked as "a" by "patch.patch()" (e.g. via transplant), and
     # - have to be marked as "n" after commit, but
     # - aren't listed in "repo[node].files()"
 
     lfdirstate = openlfdirstate(repo.ui, repo)
     for f in ctx.files():
         if isstandin(f):
             lfile = splitstandin(f)
             synclfdirstate(repo, lfdirstate, lfile, False)
     lfdirstate.write()
 
     # As part of committing, copy all of the largefiles into the cache.
     copyalltostore(repo, node)
 
 def getlfilestoupdate(oldstandins, newstandins):
     changedstandins = set(oldstandins).symmetric_difference(set(newstandins))
     filelist = []
     for f in changedstandins:
         if f[0] not in filelist:
             filelist.append(f[0])
     return filelist
 
 def getlfilestoupload(repo, missing, addfunc):
     for i, n in enumerate(missing):
         repo.ui.progress(_('finding outgoing largefiles'), i,
             unit=_('revisions'), total=len(missing))
         parents = [p for p in repo[n].parents() if p != node.nullid]
 
         oldlfstatus = repo.lfstatus
         repo.lfstatus = False
         try:
             ctx = repo[n]
         finally:
             repo.lfstatus = oldlfstatus
 
         files = set(ctx.files())
         if len(parents) == 2:
             mc = ctx.manifest()
             mp1 = ctx.parents()[0].manifest()
             mp2 = ctx.parents()[1].manifest()
             for f in mp1:
                 if f not in mc:
                     files.add(f)
             for f in mp2:
                 if f not in mc:
                     files.add(f)
             for f in mc:
                 if mc[f] != mp1.get(f, None) or mc[f] != mp2.get(f, None):
                     files.add(f)
         for fn in files:
             if isstandin(fn) and fn in ctx:
                 addfunc(fn, ctx[fn].data().strip())
     repo.ui.progress(_('finding outgoing largefiles'), None)
 
 def updatestandinsbymatch(repo, match):
     '''Update standins in the working directory according to specified match
 
     This returns (possibly modified) ``match`` object to be used for
     subsequent commit process.
     '''
 
     ui = repo.ui
 
     # Case 1: user calls commit with no specific files or
     # include/exclude patterns: refresh and commit all files that
     # are "dirty".
     if match is None or match.always():
         # Spend a bit of time here to get a list of files we know
         # are modified so we can compare only against those.
         # It can cost a lot of time (several seconds)
         # otherwise to update all standins if the largefiles are
         # large.
         lfdirstate = openlfdirstate(ui, repo)
         dirtymatch = matchmod.always(repo.root, repo.getcwd())
         unsure, s = lfdirstate.status(dirtymatch, [], False, False,
                                       False)
         modifiedfiles = unsure + s.modified + s.added + s.removed
         lfiles = listlfiles(repo)
         # this only loops through largefiles that exist (not
         # removed/renamed)
         for lfile in lfiles:
             if lfile in modifiedfiles:
                 if repo.wvfs.exists(standin(lfile)):
                     # this handles the case where a rebase is being
                     # performed and the working copy is not updated
                     # yet.
                     if repo.wvfs.exists(lfile):
                         updatestandin(repo,
                             standin(lfile))
 
         return match
 
     lfiles = listlfiles(repo)
     match._files = repo._subdirlfs(match.files(), lfiles)
 
     # Case 2: user calls commit with specified patterns: refresh
     # any matching big files.
     smatcher = composestandinmatcher(repo, match)
     standins = repo.dirstate.walk(smatcher, [], False, False)
 
     # No matching big files: get out of the way and pass control to
     # the usual commit() method.
     if not standins:
         return match
 
     # Refresh all matching big files.  It's possible that the
     # commit will end up failing, in which case the big files will
     # stay refreshed.  No harm done: the user modified them and
     # asked to commit them, so sooner or later we're going to
     # refresh the standins.  Might as well leave them refreshed.
     lfdirstate = openlfdirstate(ui, repo)
     for fstandin in standins:
         lfile = splitstandin(fstandin)
         if lfdirstate[lfile] != 'r':
             updatestandin(repo, fstandin)
 
     # Cook up a new matcher that only matches regular files or
     # standins corresponding to the big files requested by the
     # user.  Have to modify _files to prevent commit() from
     # complaining "not tracked" for big files.
     match = copy.copy(match)
     origmatchfn = match.matchfn
 
     # Check both the list of largefiles and the list of
     # standins because if a largefile was removed, it
     # won't be in the list of largefiles at this point
     match._files += sorted(standins)
 
     actualfiles = []
     for f in match._files:
         fstandin = standin(f)
 
         # For largefiles, only one of the normal and standin should be
         # committed (except if one of them is a remove).  In the case of a
         # standin removal, drop the normal file if it is unknown to dirstate.
         # Thus, skip plain largefile names but keep the standin.
         if f in lfiles or fstandin in standins:
             if repo.dirstate[fstandin] != 'r':
                 if repo.dirstate[f] != 'r':
                     continue
             elif repo.dirstate[f] == '?':
                 continue
 
         actualfiles.append(f)
     match._files = actualfiles
 
     def matchfn(f):
         if origmatchfn(f):
             return f not in lfiles
         else:
             return f in standins
 
     match.matchfn = matchfn
 
     return match
 
 class automatedcommithook(object):
     '''Stateful hook to update standins at the 1st commit of resuming
 
     For efficiency, updating standins in the working directory should
     be avoided while automated committing (like rebase, transplant and
     so on), because they should be updated before committing.
 
     But the 1st commit of resuming automated committing (e.g. ``rebase
     --continue``) should update them, because largefiles may be
     modified manually.
     '''
     def __init__(self, resuming):
         self.resuming = resuming
 
     def __call__(self, repo, match):
         if self.resuming:
             self.resuming = False # avoids updating at subsequent commits
             return updatestandinsbymatch(repo, match)
         else:
             return match
 
 def getstatuswriter(ui, repo, forcibly=None):
     '''Return the function to write largefiles specific status out
 
     If ``forcibly`` is ``None``, this returns the last element of
     ``repo._lfstatuswriters`` as "default" writer function.
 
     Otherwise, this returns the function to always write out (or
     ignore if ``not forcibly``) status.
     '''
     if forcibly is None and util.safehasattr(repo, '_largefilesenabled'):
         return repo._lfstatuswriters[-1]
     else:
         if forcibly:
             return ui.status # forcibly WRITE OUT
         else:
             return lambda *msg, **opts: None # forcibly IGNORE

mercurial/profiling.py

@@ -1,176 +1,176
 # profiling.py - profiling functions
 #
 # Copyright 2016 Gregory Szorc <[e-mail]>
 #
 # This software may be used and distributed according to the terms of the
 # GNU General Public License version 2 or any later version.
 
 from __future__ import absolute_import, print_function
 
 import contextlib
 import time
 
 from .i18n import _
 from . import (
+    encoding,
     error,
-    pycompat,
     util,
 )
 
 @contextlib.contextmanager
 def lsprofile(ui, fp):
     format = ui.config('profiling', 'format', default='text')
     field = ui.config('profiling', 'sort', default='inlinetime')
     limit = ui.configint('profiling', 'limit', default=30)
     climit = ui.configint('profiling', 'nested', default=0)
 
     if format not in ['text', 'kcachegrind']:
         ui.warn(_("unrecognized profiling format '%s'"
                     " - Ignored\n") % format)
         format = 'text'
 
     try:
         from . import lsprof
     except ImportError:
         raise error.Abort(_(
             'lsprof not available - install from '
             'http://codespeak.net/svn/user/arigo/hack/misc/lsprof/'))
     p = lsprof.Profiler()
     p.enable(subcalls=True)
     try:
         yield
     finally:
         p.disable()
 
         if format == 'kcachegrind':
             from . import lsprofcalltree
             calltree = lsprofcalltree.KCacheGrind(p)
             calltree.output(fp)
         else:
             # format == 'text'
             stats = lsprof.Stats(p.getstats())
             stats.sort(field)
             stats.pprint(limit=limit, file=fp, climit=climit)
 
 @contextlib.contextmanager
 def flameprofile(ui, fp):
     try:
         from flamegraph import flamegraph
     except ImportError:
         raise error.Abort(_(
             'flamegraph not available - install from '
             'https://github.com/evanhempel/python-flamegraph'))
     # developer config: profiling.freq
     freq = ui.configint('profiling', 'freq', default=1000)
     filter_ = None
     collapse_recursion = True
     thread = flamegraph.ProfileThread(fp, 1.0 / freq,
                                       filter_, collapse_recursion)
     start_time = time.clock()
     try:
         thread.start()
         yield
     finally:
         thread.stop()
         thread.join()
         print('Collected %d stack frames (%d unique) in %2.2f seconds.' % (
             time.clock() - start_time, thread.num_frames(),
             thread.num_frames(unique=True)))
 
 @contextlib.contextmanager
 def statprofile(ui, fp):
     from . import statprof
 
     freq = ui.configint('profiling', 'freq', default=1000)
     if freq > 0:
         # Cannot reset when profiler is already active. So silently no-op.
         if statprof.state.profile_level == 0:
             statprof.reset(freq)
     else:
         ui.warn(_("invalid sampling frequency '%s' - ignoring\n") % freq)
 
     statprof.start(mechanism='thread')
 
     try:
         yield
     finally:
         data = statprof.stop()
 
         profformat = ui.config('profiling', 'statformat', 'hotpath')
 
         formats = {
             'byline': statprof.DisplayFormats.ByLine,
             'bymethod': statprof.DisplayFormats.ByMethod,
             'hotpath': statprof.DisplayFormats.Hotpath,
             'json': statprof.DisplayFormats.Json,
         }
 
         if profformat in formats:
             displayformat = formats[profformat]
         else:
             ui.warn(_('unknown profiler output format: %s\n') % profformat)
             displayformat = statprof.DisplayFormats.Hotpath
 
         statprof.display(fp, data=data, format=displayformat)
 
 @contextlib.contextmanager
 def profile(ui):
     """Start profiling.
 
     Profiling is active when the context manager is active. When the context
     manager exits, profiling results will be written to the configured output.
     """
-    profiler = pycompat.osgetenv('HGPROF')
+    profiler = encoding.environ.get('HGPROF')
     if profiler is None:
         profiler = ui.config('profiling', 'type', default='stat')
     if profiler not in ('ls', 'stat', 'flame'):
         ui.warn(_("unrecognized profiler '%s' - ignored\n") % profiler)
         profiler = 'stat'
 
     output = ui.config('profiling', 'output')
 
     if output == 'blackbox':
         fp = util.stringio()
     elif output:
         path = ui.expandpath(output)
         fp = open(path, 'wb')
     else:
         fp = ui.ferr
 
     try:
         if profiler == 'ls':
             proffn = lsprofile
         elif profiler == 'flame':
             proffn = flameprofile
         else:
             proffn = statprofile
 
         with proffn(ui, fp):
             yield
 
     finally:
         if output:
             if output == 'blackbox':
                 val = 'Profile:\n%s' % fp.getvalue()
                 # ui.log treats the input as a format string,
                 # so we need to escape any % signs.
                 val = val.replace('%', '%%')
                 ui.log('profile', val)
             fp.close()
 
 @contextlib.contextmanager
 def maybeprofile(ui):
     """Profile if enabled, else do nothing.
 
     This context manager can be used to optionally profile if profiling
     is enabled. Otherwise, it does nothing.
 
     The purpose of this context manager is to make calling code simpler:
     just use a single code path for calling into code you may want to profile
     and this function determines whether to start profiling.
     """
     if ui.configbool('profiling', 'enabled'):
         with profile(ui):
             yield
     else:
         yield

mercurial/pycompat.py

@@ -1,293 +1,291
 # pycompat.py - portability shim for python 3
 #
 # This software may be used and distributed according to the terms of the
 # GNU General Public License version 2 or any later version.
 
 """Mercurial portability shim for python 3.
 
 This contains aliases to hide python version-specific details from the core.
 """
 
 from __future__ import absolute_import
 
 import getopt
 import os
 import shlex
 import sys
 
 ispy3 = (sys.version_info[0] >= 3)
 
 if not ispy3:
     import cPickle as pickle
     import cStringIO as io
     import httplib
     import Queue as _queue
     import SocketServer as socketserver
     import urlparse
     urlunquote = urlparse.unquote
     import xmlrpclib
 else:
     import http.client as httplib
     import io
     import pickle
     import queue as _queue
     import socketserver
     import urllib.parse as urlparse
     urlunquote = urlparse.unquote_to_bytes
     import xmlrpc.client as xmlrpclib
 
 if ispy3:
     import builtins
     import functools
     fsencode = os.fsencode
     fsdecode = os.fsdecode
     # A bytes version of os.name.
     osname = os.name.encode('ascii')
     ospathsep = os.pathsep.encode('ascii')
     ossep = os.sep.encode('ascii')
     osaltsep = os.altsep
-    osgetenv = os.getenvb
     if osaltsep:
         osaltsep = osaltsep.encode('ascii')
     # os.getcwd() on Python 3 returns string, but it has os.getcwdb() which
     # returns bytes.
     getcwd = os.getcwdb
     sysplatform = sys.platform.encode('ascii')
     sysexecutable = sys.executable
     if sysexecutable:
         sysexecutable = os.fsencode(sysexecutable)
 
     # TODO: .buffer might not exist if std streams were replaced; we'll need
     # a silly wrapper to make a bytes stream backed by a unicode one.
     stdin = sys.stdin.buffer
     stdout = sys.stdout.buffer
     stderr = sys.stderr.buffer
 
     # Since Python 3 converts argv to wchar_t type by Py_DecodeLocale() on Unix,
     # we can use os.fsencode() to get back bytes argv.
     #
     # https://hg.python.org/cpython/file/v3.5.1/Programs/python.c#l55
     #
     # TODO: On Windows, the native argv is wchar_t, so we'll need a different
     # workaround to simulate the Python 2 (i.e. ANSI Win32 API) behavior.
     sysargv = list(map(os.fsencode, sys.argv))
 
     def sysstr(s):
         """Return a keyword str to be passed to Python functions such as
         getattr() and str.encode()
 
         This never raises UnicodeDecodeError. Non-ascii characters are
         considered invalid and mapped to arbitrary but unique code points
         such that 'sysstr(a) != sysstr(b)' for all 'a != b'.
         """
         if isinstance(s, builtins.str):
             return s
         return s.decode(u'latin-1')
 
     def _wrapattrfunc(f):
         @functools.wraps(f)
         def w(object, name, *args):
             return f(object, sysstr(name), *args)
         return w
 
     # these wrappers are automagically imported by hgloader
     delattr = _wrapattrfunc(builtins.delattr)
     getattr = _wrapattrfunc(builtins.getattr)
     hasattr = _wrapattrfunc(builtins.hasattr)
     setattr = _wrapattrfunc(builtins.setattr)
     xrange = builtins.range
 
     # getopt.getopt() on Python 3 deals with unicodes internally so we cannot
     # pass bytes there. Passing unicodes will result in unicodes as return
     # values which we need to convert again to bytes.
     def getoptb(args, shortlist, namelist):
         args = [a.decode('latin-1') for a in args]
         shortlist = shortlist.decode('latin-1')
         namelist = [a.decode('latin-1') for a in namelist]
         opts, args = getopt.getopt(args, shortlist, namelist)
         opts = [(a[0].encode('latin-1'), a[1].encode('latin-1'))
                 for a in opts]
         args = [a.encode('latin-1') for a in args]
         return opts, args
 
     # keys of keyword arguments in Python need to be strings which are unicodes
     # Python 3. This function takes keyword arguments, convert the keys to str.
     def strkwargs(dic):
         dic = dict((k.decode('latin-1'), v) for k, v in dic.iteritems())
         return dic
 
     # keys of keyword arguments need to be unicode while passing into
     # a function. This function helps us to convert those keys back to bytes
     # again as we need to deal with bytes.
     def byteskwargs(dic):
         dic = dict((k.encode('latin-1'), v) for k, v in dic.iteritems())
         return dic
 
     # shlex.split() accepts unicodes on Python 3. This function takes bytes
     # argument, convert it into unicodes, pass into shlex.split(), convert the
     # returned value to bytes and return that.
     # TODO: handle shlex.shlex().
     def shlexsplit(s):
         ret = shlex.split(s.decode('latin-1'))
         return [a.encode('latin-1') for a in ret]
 
 else:
     def sysstr(s):
         return s
 
     # Partial backport from os.py in Python 3, which only accepts bytes.
     # In Python 2, our paths should only ever be bytes, a unicode path
     # indicates a bug.
     def fsencode(filename):
         if isinstance(filename, str):
             return filename
         else:
             raise TypeError(
                 "expect str, not %s" % type(filename).__name__)
 
     # In Python 2, fsdecode() has a very chance to receive bytes. So it's
     # better not to touch Python 2 part as it's already working fine.
     def fsdecode(filename):
         return filename
 
     def getoptb(args, shortlist, namelist):
         return getopt.getopt(args, shortlist, namelist)
 
     def strkwargs(dic):
         return dic
 
     def byteskwargs(dic):
         return dic
 
     osname = os.name
     ospathsep = os.pathsep
     ossep = os.sep
     osaltsep = os.altsep
     stdin = sys.stdin
     stdout = sys.stdout
     stderr = sys.stderr
     sysargv = sys.argv
     sysplatform = sys.platform
     getcwd = os.getcwd
-    osgetenv = os.getenv
     sysexecutable = sys.executable
     shlexsplit = shlex.split
 
 stringio = io.StringIO
 empty = _queue.Empty
 queue = _queue.Queue
 
 class _pycompatstub(object):
     def __init__(self):
         self._aliases = {}
 
     def _registeraliases(self, origin, items):
         """Add items that will be populated at the first access"""
         items = map(sysstr, items)
         self._aliases.update(
             (item.replace(sysstr('_'), sysstr('')).lower(), (origin, item))
             for item in items)
 
     def __getattr__(self, name):
         try:
             origin, item = self._aliases[name]
         except KeyError:
             raise AttributeError(name)
         self.__dict__[name] = obj = getattr(origin, item)
         return obj
 
 httpserver = _pycompatstub()
 urlreq = _pycompatstub()
 urlerr = _pycompatstub()
 if not ispy3:
     import BaseHTTPServer
     import CGIHTTPServer
     import SimpleHTTPServer
     import urllib2
     import urllib
     urlreq._registeraliases(urllib, (
         "addclosehook",
         "addinfourl",
         "ftpwrapper",
         "pathname2url",
         "quote",
         "splitattr",
         "splitpasswd",
         "splitport",
         "splituser",
         "unquote",
         "url2pathname",
         "urlencode",
     ))
     urlreq._registeraliases(urllib2, (
         "AbstractHTTPHandler",
         "BaseHandler",
         "build_opener",
         "FileHandler",
         "FTPHandler",
         "HTTPBasicAuthHandler",
         "HTTPDigestAuthHandler",
         "HTTPHandler",
         "HTTPPasswordMgrWithDefaultRealm",
         "HTTPSHandler",
         "install_opener",
         "ProxyHandler",
         "Request",
         "urlopen",
     ))
     urlerr._registeraliases(urllib2, (
         "HTTPError",
         "URLError",
     ))
     httpserver._registeraliases(BaseHTTPServer, (
         "HTTPServer",
         "BaseHTTPRequestHandler",
     ))
     httpserver._registeraliases(SimpleHTTPServer, (
         "SimpleHTTPRequestHandler",
     ))
     httpserver._registeraliases(CGIHTTPServer, (
         "CGIHTTPRequestHandler",
     ))
 
 else:
     import urllib.request
     urlreq._registeraliases(urllib.request, (
         "AbstractHTTPHandler",
         "addclosehook",
         "addinfourl",
         "BaseHandler",
         "build_opener",
         "FileHandler",
         "FTPHandler",
         "ftpwrapper",
         "HTTPHandler",
         "HTTPSHandler",
         "install_opener",
         "pathname2url",
         "HTTPBasicAuthHandler",
         "HTTPDigestAuthHandler",
         "HTTPPasswordMgrWithDefaultRealm",
         "ProxyHandler",
         "quote",
         "Request",
         "splitattr",
         "splitpasswd",
         "splitport",
         "splituser",
         "unquote",
         "url2pathname",
         "urlopen",
     ))
     import urllib.error
     urlerr._registeraliases(urllib.error, (
         "HTTPError",
         "URLError",
     ))
     import http.server
     httpserver._registeraliases(http.server, (
         "HTTPServer",
         "BaseHTTPRequestHandler",
         "SimpleHTTPRequestHandler",
         "CGIHTTPRequestHandler",
     ))

mercurial/url.py

@@ -1,486 +1,486
 # url.py - HTTP handling for mercurial
 #
 # Copyright 2005, 2006, 2007, 2008 Matt Mackall <[e-mail]>
 # Copyright 2006, 2007 Alexis S. L. Carvalho <[e-mail]>
 # Copyright 2006 Vadim Gelfer <[e-mail]>
 #
 # This software may be used and distributed according to the terms of the
 # GNU General Public License version 2 or any later version.
 
 from __future__ import absolute_import
 
 import base64
 import os
 import socket
 
 from .i18n import _
 from . import (
+    encoding,
     error,
     httpconnection as httpconnectionmod,
     keepalive,
-    pycompat,
     sslutil,
     util,
 )
 
 httplib = util.httplib
 stringio = util.stringio
 urlerr = util.urlerr
 urlreq = util.urlreq
 
 class passwordmgr(object):
     def __init__(self, ui, passwddb):
         self.ui = ui
         self.passwddb = passwddb
 
     def add_password(self, realm, uri, user, passwd):
         return self.passwddb.add_password(realm, uri, user, passwd)
 
     def find_user_password(self, realm, authuri):
         authinfo = self.passwddb.find_user_password(realm, authuri)
         user, passwd = authinfo
         if user and passwd:
             self._writedebug(user, passwd)
             return (user, passwd)
 
         if not user or not passwd:
             res = httpconnectionmod.readauthforuri(self.ui, authuri, user)
             if res:
                 group, auth = res
                 user, passwd = auth.get('username'), auth.get('password')
                 self.ui.debug("using auth.%s.* for authentication\n" % group)
         if not user or not passwd:
             u = util.url(authuri)
             u.query = None
             if not self.ui.interactive():
                 raise error.Abort(_('http authorization required for %s') %
                                  util.hidepassword(str(u)))
 
             self.ui.write(_("http authorization required for %s\n") %
                           util.hidepassword(str(u)))
             self.ui.write(_("realm: %s\n") % realm)
             if user:
                 self.ui.write(_("user: %s\n") % user)
             else:
                 user = self.ui.prompt(_("user:"), default=None)
 
             if not passwd:
                 passwd = self.ui.getpass()
 
         self.passwddb.add_password(realm, authuri, user, passwd)
         self._writedebug(user, passwd)
         return (user, passwd)
 
     def _writedebug(self, user, passwd):
         msg = _('http auth: user %s, password %s\n')
         self.ui.debug(msg % (user, passwd and '*' * len(passwd) or 'not set'))
 
     def find_stored_password(self, authuri):
         return self.passwddb.find_user_password(None, authuri)
 
 class proxyhandler(urlreq.proxyhandler):
     def __init__(self, ui):
         proxyurl = (ui.config("http_proxy", "host") or
-                        pycompat.osgetenv('http_proxy'))
+                        encoding.environ.get('http_proxy'))
         # XXX proxyauthinfo = None
 
         if proxyurl:
             # proxy can be proper url or host[:port]
             if not (proxyurl.startswith('http:') or
                     proxyurl.startswith('https:')):
                 proxyurl = 'http://' + proxyurl + '/'
             proxy = util.url(proxyurl)
             if not proxy.user:
                 proxy.user = ui.config("http_proxy", "user")
                 proxy.passwd = ui.config("http_proxy", "passwd")
 
             # see if we should use a proxy for this url
             no_list = ["localhost", "127.0.0.1"]
             no_list.extend([p.lower() for
                             p in ui.configlist("http_proxy", "no")])
             no_list.extend([p.strip().lower() for
-                            p in pycompat.osgetenv("no_proxy", '').split(',')
+                            p in encoding.environ.get("no_proxy", '').split(',')
                             if p.strip()])
             # "http_proxy.always" config is for running tests on localhost
             if ui.configbool("http_proxy", "always"):
                 self.no_list = []
             else:
                 self.no_list = no_list
 
             proxyurl = str(proxy)
             proxies = {'http': proxyurl, 'https': proxyurl}
             ui.debug('proxying through http://%s:%s\n' %
                       (proxy.host, proxy.port))
         else:
             proxies = {}
 
         urlreq.proxyhandler.__init__(self, proxies)
         self.ui = ui
 
     def proxy_open(self, req, proxy, type_):
         host = req.get_host().split(':')[0]
         for e in self.no_list:
             if host == e:
                 return None
             if e.startswith('*.') and host.endswith(e[2:]):
                 return None
             if e.startswith('.') and host.endswith(e[1:]):
                 return None
 
         return urlreq.proxyhandler.proxy_open(self, req, proxy, type_)
 
 def _gen_sendfile(orgsend):
     def _sendfile(self, data):
         # send a file
         if isinstance(data, httpconnectionmod.httpsendfile):
             # if auth required, some data sent twice, so rewind here
             data.seek(0)
             for chunk in util.filechunkiter(data):
                 orgsend(self, chunk)
         else:
             orgsend(self, data)
     return _sendfile
 
 has_https = util.safehasattr(urlreq, 'httpshandler')
 
 class httpconnection(keepalive.HTTPConnection):
     # must be able to send big bundle as stream.
     send = _gen_sendfile(keepalive.HTTPConnection.send)
 
     def getresponse(self):
         proxyres = getattr(self, 'proxyres', None)
         if proxyres:
             if proxyres.will_close:
                 self.close()
             self.proxyres = None
             return proxyres
         return keepalive.HTTPConnection.getresponse(self)
 
 # general transaction handler to support different ways to handle
 # HTTPS proxying before and after Python 2.6.3.
 def _generic_start_transaction(handler, h, req):
     tunnel_host = getattr(req, '_tunnel_host', None)
     if tunnel_host:
         if tunnel_host[:7] not in ['http://', 'https:/']:
             tunnel_host = 'https://' + tunnel_host
         new_tunnel = True
     else:
         tunnel_host = req.get_selector()
         new_tunnel = False
 
     if new_tunnel or tunnel_host == req.get_full_url(): # has proxy
         u = util.url(tunnel_host)
         if new_tunnel or u.scheme == 'https': # only use CONNECT for HTTPS
             h.realhostport = ':'.join([u.host, (u.port or '443')])
             h.headers = req.headers.copy()
             h.headers.update(handler.parent.addheaders)
             return
 
     h.realhostport = None
     h.headers = None
 
 def _generic_proxytunnel(self):
     proxyheaders = dict(
             [(x, self.headers[x]) for x in self.headers
              if x.lower().startswith('proxy-')])
     self.send('CONNECT %s HTTP/1.0\r\n' % self.realhostport)
     for header in proxyheaders.iteritems():
         self.send('%s: %s\r\n' % header)
     self.send('\r\n')
 
     # majority of the following code is duplicated from
     # httplib.HTTPConnection as there are no adequate places to
     # override functions to provide the needed functionality
     res = self.response_class(self.sock,
                               strict=self.strict,
                               method=self._method)
 
     while True:
         version, status, reason = res._read_status()
         if status != httplib.CONTINUE:
             break
         # skip lines that are all whitespace
         list(iter(lambda: res.fp.readline().strip(), ''))
     res.status = status
     res.reason = reason.strip()
 
     if res.status == 200:
         # skip lines until we find a blank line
         list(iter(res.fp.readline, '\r\n'))
         return True
 
     if version == 'HTTP/1.0':
         res.version = 10
     elif version.startswith('HTTP/1.'):
         res.version = 11
     elif version == 'HTTP/0.9':
         res.version = 9
     else:
         raise httplib.UnknownProtocol(version)
 
     if res.version == 9:
         res.length = None
         res.chunked = 0
         res.will_close = 1
         res.msg = httplib.HTTPMessage(stringio())
         return False
 
     res.msg = httplib.HTTPMessage(res.fp)
     res.msg.fp = None
 
     # are we using the chunked-style of transfer encoding?
     trenc = res.msg.getheader('transfer-encoding')
     if trenc and trenc.lower() == "chunked":
         res.chunked = 1
         res.chunk_left = None
     else:
         res.chunked = 0
 
     # will the connection close at the end of the response?
     res.will_close = res._check_close()
 
     # do we have a Content-Length?
     # NOTE: RFC 2616, section 4.4, #3 says we ignore this if
     # transfer-encoding is "chunked"
     length = res.msg.getheader('content-length')
     if length and not res.chunked:
         try:
             res.length = int(length)
         except ValueError:
             res.length = None
         else:
             if res.length < 0:  # ignore nonsensical negative lengths
                 res.length = None
     else:
         res.length = None
 
     # does the body have a fixed length? (of zero)
     if (status == httplib.NO_CONTENT or status == httplib.NOT_MODIFIED or
         100 <= status < 200 or # 1xx codes
         res._method == 'HEAD'):
         res.length = 0
 
     # if the connection remains open, and we aren't using chunked, and
     # a content-length was not provided, then assume that the connection
     # WILL close.
     if (not res.will_close and
        not res.chunked and
        res.length is None):
         res.will_close = 1
 
     self.proxyres = res
 
     return False
 
 class httphandler(keepalive.HTTPHandler):
     def http_open(self, req):
         return self.do_open(httpconnection, req)
 
     def _start_transaction(self, h, req):
         _generic_start_transaction(self, h, req)
         return keepalive.HTTPHandler._start_transaction(self, h, req)
 
 if has_https:
     class httpsconnection(httplib.HTTPConnection):
         response_class = keepalive.HTTPResponse
         default_port = httplib.HTTPS_PORT
         # must be able to send big bundle as stream.
         send = _gen_sendfile(keepalive.safesend)
         getresponse = keepalive.wrapgetresponse(httplib.HTTPConnection)
 
         def __init__(self, host, port=None, key_file=None, cert_file=None,
                      *args, **kwargs):
             httplib.HTTPConnection.__init__(self, host, port, *args, **kwargs)
             self.key_file = key_file
             self.cert_file = cert_file
 
         def connect(self):
             self.sock = socket.create_connection((self.host, self.port))
 
             host = self.host
             if self.realhostport: # use CONNECT proxy
                 _generic_proxytunnel(self)
                 host = self.realhostport.rsplit(':', 1)[0]
             self.sock = sslutil.wrapsocket(
                 self.sock, self.key_file, self.cert_file, ui=self.ui,
                 serverhostname=host)
             sslutil.validatesocket(self.sock)
 
     class httpshandler(keepalive.KeepAliveHandler, urlreq.httpshandler):
         def __init__(self, ui):
             keepalive.KeepAliveHandler.__init__(self)
             urlreq.httpshandler.__init__(self)
             self.ui = ui
             self.pwmgr = passwordmgr(self.ui,
                                      self.ui.httppasswordmgrdb)
 
         def _start_transaction(self, h, req):
             _generic_start_transaction(self, h, req)
             return keepalive.KeepAliveHandler._start_transaction(self, h, req)
 
         def https_open(self, req):
             # req.get_full_url() does not contain credentials and we may
             # need them to match the certificates.
             url = req.get_full_url()
             user, password = self.pwmgr.find_stored_password(url)
             res = httpconnectionmod.readauthforuri(self.ui, url, user)
             if res:
                 group, auth = res
                 self.auth = auth
                 self.ui.debug("using auth.%s.* for authentication\n" % group)
             else:
                 self.auth = None
             return self.do_open(self._makeconnection, req)
 
         def _makeconnection(self, host, port=None, *args, **kwargs):
             keyfile = None
             certfile = None
 
             if len(args) >= 1: # key_file
                 keyfile = args[0]
             if len(args) >= 2: # cert_file
                 certfile = args[1]
             args = args[2:]
 
             # if the user has specified different key/cert files in
             # hgrc, we prefer these
             if self.auth and 'key' in self.auth and 'cert' in self.auth:
                 keyfile = self.auth['key']
                 certfile = self.auth['cert']
 
             conn = httpsconnection(host, port, keyfile, certfile, *args,
                                    **kwargs)
             conn.ui = self.ui
             return conn
 
 class httpdigestauthhandler(urlreq.httpdigestauthhandler):
     def __init__(self, *args, **kwargs):
         urlreq.httpdigestauthhandler.__init__(self, *args, **kwargs)
         self.retried_req = None
 
     def reset_retry_count(self):
         # Python 2.6.5 will call this on 401 or 407 errors and thus loop
         # forever. We disable reset_retry_count completely and reset in
         # http_error_auth_reqed instead.
         pass
 
     def http_error_auth_reqed(self, auth_header, host, req, headers):
         # Reset the retry counter once for each request.
         if req is not self.retried_req:
             self.retried_req = req
             self.retried = 0
         return urlreq.httpdigestauthhandler.http_error_auth_reqed(
                     self, auth_header, host, req, headers)
 
 class httpbasicauthhandler(urlreq.httpbasicauthhandler):
     def __init__(self, *args, **kwargs):
         self.auth = None
         urlreq.httpbasicauthhandler.__init__(self, *args, **kwargs)
         self.retried_req = None
 
     def http_request(self, request):
         if self.auth:
             request.add_unredirected_header(self.auth_header, self.auth)
 
         return request
 
     def https_request(self, request):
         if self.auth:
             request.add_unredirected_header(self.auth_header, self.auth)
 
         return request
 
     def reset_retry_count(self):
         # Python 2.6.5 will call this on 401 or 407 errors and thus loop
         # forever. We disable reset_retry_count completely and reset in
         # http_error_auth_reqed instead.
         pass
 
     def http_error_auth_reqed(self, auth_header, host, req, headers):
         # Reset the retry counter once for each request.
         if req is not self.retried_req:
             self.retried_req = req
             self.retried = 0
         return urlreq.httpbasicauthhandler.http_error_auth_reqed(
                         self, auth_header, host, req, headers)
 
     def retry_http_basic_auth(self, host, req, realm):
         user, pw = self.passwd.find_user_password(realm, req.get_full_url())
         if pw is not None:
             raw = "%s:%s" % (user, pw)
             auth = 'Basic %s' % base64.b64encode(raw).strip()
             if req.get_header(self.auth_header, None) == auth:
                 return None
             self.auth = auth
             req.add_unredirected_header(self.auth_header, auth)
             return self.parent.open(req)
         else:
             return None
 
 handlerfuncs = []
 
 def opener(ui, authinfo=None):
     '''
     construct an opener suitable for urllib2
     authinfo will be added to the password manager
     '''
     # experimental config: ui.usehttp2
     if ui.configbool('ui', 'usehttp2', False):
         handlers = [
             httpconnectionmod.http2handler(
                 ui,
                 passwordmgr(ui, ui.httppasswordmgrdb))
         ]
     else:
         handlers = [httphandler()]
         if has_https:
             handlers.append(httpshandler(ui))
 
     handlers.append(proxyhandler(ui))
 
     passmgr = passwordmgr(ui, ui.httppasswordmgrdb)
     if authinfo is not None:
         realm, uris, user, passwd = authinfo
         saveduser, savedpass = passmgr.find_stored_password(uris[0])
         if user != saveduser or passwd:
             passmgr.add_password(realm, uris, user, passwd)
         ui.debug('http auth: user %s, password %s\n' %
                  (user, passwd and '*' * len(passwd) or 'not set'))
 
     handlers.extend((httpbasicauthhandler(passmgr),
                      httpdigestauthhandler(passmgr)))
     handlers.extend([h(ui, passmgr) for h in handlerfuncs])
     opener = urlreq.buildopener(*handlers)
 
     # The user agent should should *NOT* be used by servers for e.g.
     # protocol detection or feature negotiation: there are other
     # facilities for that.
     #
     # "mercurial/proto-1.0" was the original user agent string and
     # exists for backwards compatibility reasons.
     #
     # The "(Mercurial %s)" string contains the distribution
     # name and version. Other client implementations should choose their
     # own distribution name. Since servers should not be using the user
     # agent string for anything, clients should be able to define whatever
     # user agent they deem appropriate.
     agent = 'mercurial/proto-1.0 (Mercurial %s)' % util.version()
     opener.addheaders = [('User-agent', agent)]
 
     # This header should only be needed by wire protocol requests. But it has
     # been sent on all requests since forever. We keep sending it for backwards
     # compatibility reasons. Modern versions of the wire protocol use
     # X-HgProto-<N> for advertising client support.
     opener.addheaders.append(('Accept', 'application/mercurial-0.1'))
     return opener
 
 def open(ui, url_, data=None):
     u = util.url(url_)
     if u.scheme:
         u.scheme = u.scheme.lower()
         url_, authinfo = u.authinfo()
     else:
         path = util.normpath(os.path.abspath(url_))
         url_ = 'file://' + urlreq.pathname2url(path)
         authinfo = None
     return opener(ui, authinfo).open(url_, data)