hgext/highlight/highlight.py
author Augie Fackler <augie@google.com>
Sun, 06 Oct 2019 09:48:39 -0400
changeset 43077 687b865b95ad
parent 43076 2372284d9457
child 43117 8ff1ecfadcd1
permissions -rw-r--r--
formatting: byteify all mercurial/ and hgext/ string literals Done with python3.7 contrib/byteify-strings.py -i $(hg files 'set:mercurial/**.py - mercurial/thirdparty/** + hgext/**.py - hgext/fsmonitor/pywatchman/** - mercurial/__init__.py') black -l 80 -t py33 -S $(hg files 'set:**.py - mercurial/thirdparty/** - "contrib/python-zstandard/**" - hgext/fsmonitor/pywatchman/**') # skip-blame mass-reformatting only Differential Revision: https://phab.mercurial-scm.org/D6972
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
8251
7fc30044b514 highlight: add copyright and license header
Martin Geisler <mg@lazybytes.net>
parents: 7948
diff changeset
     1
# highlight.py - highlight extension implementation file
7fc30044b514 highlight: add copyright and license header
Martin Geisler <mg@lazybytes.net>
parents: 7948
diff changeset
     2
#
7fc30044b514 highlight: add copyright and license header
Martin Geisler <mg@lazybytes.net>
parents: 7948
diff changeset
     3
#  Copyright 2007-2009 Adam Hupp <adam@hupp.org> and others
7fc30044b514 highlight: add copyright and license header
Martin Geisler <mg@lazybytes.net>
parents: 7948
diff changeset
     4
#
7fc30044b514 highlight: add copyright and license header
Martin Geisler <mg@lazybytes.net>
parents: 7948
diff changeset
     5
# This software may be used and distributed according to the terms of the
10263
25e572394f5c Update license to GPLv2+
Matt Mackall <mpm@selenic.com>
parents: 9424
diff changeset
     6
# GNU General Public License version 2 or any later version.
6938
ce94b3236ea4 highlight: split code to improve startup times
Patrick Mezard <pmezard@gmail.com>
parents: 6666
diff changeset
     7
#
ce94b3236ea4 highlight: split code to improve startup times
Patrick Mezard <pmezard@gmail.com>
parents: 6666
diff changeset
     8
# The original module was split in an interface and an implementation
ce94b3236ea4 highlight: split code to improve startup times
Patrick Mezard <pmezard@gmail.com>
parents: 6666
diff changeset
     9
# file to defer pygments loading and speedup extension setup.
5532
40a06e39f010 extension for synax highlighting in the hgweb file revision view
Adam Hupp <adam@hupp.org>
parents:
diff changeset
    10
29485
6a98f9408a50 py3: make files use absolute_import and print_function
Pulkit Goyal <7895pulkit@gmail.com>
parents: 29216
diff changeset
    11
from __future__ import absolute_import
6a98f9408a50 py3: make files use absolute_import and print_function
Pulkit Goyal <7895pulkit@gmail.com>
parents: 29216
diff changeset
    12
5532
40a06e39f010 extension for synax highlighting in the hgweb file revision view
Adam Hupp <adam@hupp.org>
parents:
diff changeset
    13
from mercurial import demandimport
43076
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 42923
diff changeset
    14
43077
687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals
Augie Fackler <augie@google.com>
parents: 43076
diff changeset
    15
demandimport.IGNORES.update([b'pkgutil', b'pkg_resources', b'__main__'])
29485
6a98f9408a50 py3: make files use absolute_import and print_function
Pulkit Goyal <7895pulkit@gmail.com>
parents: 29216
diff changeset
    16
6a98f9408a50 py3: make files use absolute_import and print_function
Pulkit Goyal <7895pulkit@gmail.com>
parents: 29216
diff changeset
    17
from mercurial import (
6a98f9408a50 py3: make files use absolute_import and print_function
Pulkit Goyal <7895pulkit@gmail.com>
parents: 29216
diff changeset
    18
    encoding,
42923
a7abc6081bc5 highlight: fix encoding issues to enable Py3 compatibility
Connor Sheehan <sheehan@mozilla.com>
parents: 38383
diff changeset
    19
    pycompat,
37084
f0b6fbea00cf stringutil: bulk-replace call sites to point to new module
Yuya Nishihara <yuya@tcha.org>
parents: 35329
diff changeset
    20
)
f0b6fbea00cf stringutil: bulk-replace call sites to point to new module
Yuya Nishihara <yuya@tcha.org>
parents: 35329
diff changeset
    21
43076
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 42923
diff changeset
    22
from mercurial.utils import stringutil
5532
40a06e39f010 extension for synax highlighting in the hgweb file revision view
Adam Hupp <adam@hupp.org>
parents:
diff changeset
    23
32908
661025fd3e1c highlight: put pygments import inside demandimport.deactivated
Augie Fackler <raf@durin42.com>
parents: 29485
diff changeset
    24
with demandimport.deactivated():
661025fd3e1c highlight: put pygments import inside demandimport.deactivated
Augie Fackler <raf@durin42.com>
parents: 29485
diff changeset
    25
    import pygments
661025fd3e1c highlight: put pygments import inside demandimport.deactivated
Augie Fackler <raf@durin42.com>
parents: 29485
diff changeset
    26
    import pygments.formatters
661025fd3e1c highlight: put pygments import inside demandimport.deactivated
Augie Fackler <raf@durin42.com>
parents: 29485
diff changeset
    27
    import pygments.lexers
35329
169d66db5920 highlight: eagerly discover plugin lexers while demandimport is off
Augie Fackler <augie@google.com>
parents: 32908
diff changeset
    28
    import pygments.plugin
32908
661025fd3e1c highlight: put pygments import inside demandimport.deactivated
Augie Fackler <raf@durin42.com>
parents: 29485
diff changeset
    29
    import pygments.util
661025fd3e1c highlight: put pygments import inside demandimport.deactivated
Augie Fackler <raf@durin42.com>
parents: 29485
diff changeset
    30
35329
169d66db5920 highlight: eagerly discover plugin lexers while demandimport is off
Augie Fackler <augie@google.com>
parents: 32908
diff changeset
    31
    for unused in pygments.plugin.find_plugin_lexers():
169d66db5920 highlight: eagerly discover plugin lexers while demandimport is off
Augie Fackler <augie@google.com>
parents: 32908
diff changeset
    32
        pass
169d66db5920 highlight: eagerly discover plugin lexers while demandimport is off
Augie Fackler <augie@google.com>
parents: 32908
diff changeset
    33
29485
6a98f9408a50 py3: make files use absolute_import and print_function
Pulkit Goyal <7895pulkit@gmail.com>
parents: 29216
diff changeset
    34
highlight = pygments.highlight
6a98f9408a50 py3: make files use absolute_import and print_function
Pulkit Goyal <7895pulkit@gmail.com>
parents: 29216
diff changeset
    35
ClassNotFound = pygments.util.ClassNotFound
6a98f9408a50 py3: make files use absolute_import and print_function
Pulkit Goyal <7895pulkit@gmail.com>
parents: 29216
diff changeset
    36
guess_lexer = pygments.lexers.guess_lexer
6a98f9408a50 py3: make files use absolute_import and print_function
Pulkit Goyal <7895pulkit@gmail.com>
parents: 29216
diff changeset
    37
guess_lexer_for_filename = pygments.lexers.guess_lexer_for_filename
6a98f9408a50 py3: make files use absolute_import and print_function
Pulkit Goyal <7895pulkit@gmail.com>
parents: 29216
diff changeset
    38
TextLexer = pygments.lexers.TextLexer
6a98f9408a50 py3: make files use absolute_import and print_function
Pulkit Goyal <7895pulkit@gmail.com>
parents: 29216
diff changeset
    39
HtmlFormatter = pygments.formatters.HtmlFormatter
5532
40a06e39f010 extension for synax highlighting in the hgweb file revision view
Adam Hupp <adam@hupp.org>
parents:
diff changeset
    40
43076
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 42923
diff changeset
    41
SYNTAX_CSS = (
43077
687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals
Augie Fackler <augie@google.com>
parents: 43076
diff changeset
    42
    b'\n<link rel="stylesheet" href="{url}highlightcss" ' b'type="text/css" />'
43076
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 42923
diff changeset
    43
)
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 42923
diff changeset
    44
5532
40a06e39f010 extension for synax highlighting in the hgweb file revision view
Adam Hupp <adam@hupp.org>
parents:
diff changeset
    45
26680
7a3f6490ef97 highlight: add option to prevent content-only based fallback
Gregory Szorc <gregory.szorc@gmail.com>
parents: 25899
diff changeset
    46
def pygmentize(field, fctx, style, tmpl, guessfilenameonly=False):
6394
55bc0a035e1f highlight: some small cleanups
Dirkjan Ochtman <dirkjan@ochtman.nl>
parents: 6393
diff changeset
    47
5532
40a06e39f010 extension for synax highlighting in the hgweb file revision view
Adam Hupp <adam@hupp.org>
parents:
diff changeset
    48
    # append a <link ...> to the syntax highlighting css
43077
687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals
Augie Fackler <augie@google.com>
parents: 43076
diff changeset
    49
    tmpl.load(b'header')
687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals
Augie Fackler <augie@google.com>
parents: 43076
diff changeset
    50
    old_header = tmpl.cache[b'header']
5532
40a06e39f010 extension for synax highlighting in the hgweb file revision view
Adam Hupp <adam@hupp.org>
parents:
diff changeset
    51
    if SYNTAX_CSS not in old_header:
27637
b502138f5faa cleanup: remove superfluous space after space after equals (python)
timeless <timeless@mozdev.org>
parents: 26680
diff changeset
    52
        new_header = old_header + SYNTAX_CSS
43077
687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals
Augie Fackler <augie@google.com>
parents: 43076
diff changeset
    53
        tmpl.cache[b'header'] = new_header
5532
40a06e39f010 extension for synax highlighting in the hgweb file revision view
Adam Hupp <adam@hupp.org>
parents:
diff changeset
    54
6194
fe54e7501de1 highlight: bail out if file is binary
Brendan Cully <brendan@kublai.com>
parents: 6193
diff changeset
    55
    text = fctx.data()
37084
f0b6fbea00cf stringutil: bulk-replace call sites to point to new module
Yuya Nishihara <yuya@tcha.org>
parents: 35329
diff changeset
    56
    if stringutil.binary(text):
6194
fe54e7501de1 highlight: bail out if file is binary
Brendan Cully <brendan@kublai.com>
parents: 6193
diff changeset
    57
        return
fe54e7501de1 highlight: bail out if file is binary
Brendan Cully <brendan@kublai.com>
parents: 6193
diff changeset
    58
23613
7b8ff3fd11d3 highlight: ignore Unicode's extra linebreaks (issue4291)
Matt Mackall <mpm@selenic.com>
parents: 19169
diff changeset
    59
    # str.splitlines() != unicode.splitlines() because "reasons"
43077
687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals
Augie Fackler <augie@google.com>
parents: 43076
diff changeset
    60
    for c in b"\x0c\x1c\x1d\x1e":
23613
7b8ff3fd11d3 highlight: ignore Unicode's extra linebreaks (issue4291)
Matt Mackall <mpm@selenic.com>
parents: 19169
diff changeset
    61
        if c in text:
43077
687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals
Augie Fackler <augie@google.com>
parents: 43076
diff changeset
    62
            text = text.replace(c, b'')
23613
7b8ff3fd11d3 highlight: ignore Unicode's extra linebreaks (issue4291)
Matt Mackall <mpm@selenic.com>
parents: 19169
diff changeset
    63
9424
799373ff2554 highlight: fixes garbled text in non-UTF-8 environment
Yuya Nishihara <yuya@tcha.org>
parents: 8360
diff changeset
    64
    # Pygments is best used with Unicode strings:
799373ff2554 highlight: fixes garbled text in non-UTF-8 environment
Yuya Nishihara <yuya@tcha.org>
parents: 8360
diff changeset
    65
    # <http://pygments.org/docs/unicode/>
42923
a7abc6081bc5 highlight: fix encoding issues to enable Py3 compatibility
Connor Sheehan <sheehan@mozilla.com>
parents: 38383
diff changeset
    66
    text = text.decode(pycompat.sysstr(encoding.encoding), 'replace')
7120
db7557359636 highlight: convert text to local before passing to pygmentize (issue1341)
Christian Ebert <blacktrash@gmx.net>
parents: 6938
diff changeset
    67
6193
2344da8eb9b4 highlight: support annotate, and reduce layering violations.
Brendan Cully <brendan@kublai.com>
parents: 5991
diff changeset
    68
    # To get multi-line strings right, we can't format line-by-line
2344da8eb9b4 highlight: support annotate, and reduce layering violations.
Brendan Cully <brendan@kublai.com>
parents: 5991
diff changeset
    69
    try:
42923
a7abc6081bc5 highlight: fix encoding issues to enable Py3 compatibility
Connor Sheehan <sheehan@mozilla.com>
parents: 38383
diff changeset
    70
        path = pycompat.sysstr(fctx.path())
43076
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 42923
diff changeset
    71
        lexer = guess_lexer_for_filename(path, text[:1024], stripnl=False)
6494
c30849d4c8ba highlight: backward compatibility with pygments 0.5.1
Benoit Allard <benoit@aeteurope.nl>
parents: 6212
diff changeset
    72
    except (ClassNotFound, ValueError):
26680
7a3f6490ef97 highlight: add option to prevent content-only based fallback
Gregory Szorc <gregory.szorc@gmail.com>
parents: 25899
diff changeset
    73
        # guess_lexer will return a lexer if *any* lexer matches. There is
7a3f6490ef97 highlight: add option to prevent content-only based fallback
Gregory Szorc <gregory.szorc@gmail.com>
parents: 25899
diff changeset
    74
        # no way to specify a minimum match score. This can give a high rate of
7a3f6490ef97 highlight: add option to prevent content-only based fallback
Gregory Szorc <gregory.szorc@gmail.com>
parents: 25899
diff changeset
    75
        # false positives on files with an unknown filename pattern.
7a3f6490ef97 highlight: add option to prevent content-only based fallback
Gregory Szorc <gregory.szorc@gmail.com>
parents: 25899
diff changeset
    76
        if guessfilenameonly:
7a3f6490ef97 highlight: add option to prevent content-only based fallback
Gregory Szorc <gregory.szorc@gmail.com>
parents: 25899
diff changeset
    77
            return
7a3f6490ef97 highlight: add option to prevent content-only based fallback
Gregory Szorc <gregory.szorc@gmail.com>
parents: 25899
diff changeset
    78
6198
358cc9cf54db highlight: guess by text when path name is ambiguous
Brendan Cully <brendan@kublai.com>
parents: 6197
diff changeset
    79
        try:
19169
bcdfb6078b9f highlight: fix page layout with empty first and last lines
Alexander Plavin <me@aplavin.ru>
parents: 18054
diff changeset
    80
            lexer = guess_lexer(text[:1024], stripnl=False)
6494
c30849d4c8ba highlight: backward compatibility with pygments 0.5.1
Benoit Allard <benoit@aeteurope.nl>
parents: 6212
diff changeset
    81
        except (ClassNotFound, ValueError):
25899
c35ee1bbbbdc highlight: exit early on textual and unknown files (issue3005)
Anton Shestakov <av6@dwimlabs.net>
parents: 25867
diff changeset
    82
            # Don't highlight unknown files
c35ee1bbbbdc highlight: exit early on textual and unknown files (issue3005)
Anton Shestakov <av6@dwimlabs.net>
parents: 25867
diff changeset
    83
            return
c35ee1bbbbdc highlight: exit early on textual and unknown files (issue3005)
Anton Shestakov <av6@dwimlabs.net>
parents: 25867
diff changeset
    84
c35ee1bbbbdc highlight: exit early on textual and unknown files (issue3005)
Anton Shestakov <av6@dwimlabs.net>
parents: 25867
diff changeset
    85
    # Don't highlight text files
c35ee1bbbbdc highlight: exit early on textual and unknown files (issue3005)
Anton Shestakov <av6@dwimlabs.net>
parents: 25867
diff changeset
    86
    if isinstance(lexer, TextLexer):
c35ee1bbbbdc highlight: exit early on textual and unknown files (issue3005)
Anton Shestakov <av6@dwimlabs.net>
parents: 25867
diff changeset
    87
        return
6193
2344da8eb9b4 highlight: support annotate, and reduce layering violations.
Brendan Cully <brendan@kublai.com>
parents: 5991
diff changeset
    88
42923
a7abc6081bc5 highlight: fix encoding issues to enable Py3 compatibility
Connor Sheehan <sheehan@mozilla.com>
parents: 38383
diff changeset
    89
    formatter = HtmlFormatter(nowrap=True, style=pycompat.sysstr(style))
5532
40a06e39f010 extension for synax highlighting in the hgweb file revision view
Adam Hupp <adam@hupp.org>
parents:
diff changeset
    90
6193
2344da8eb9b4 highlight: support annotate, and reduce layering violations.
Brendan Cully <brendan@kublai.com>
parents: 5991
diff changeset
    91
    colorized = highlight(text, lexer, formatter)
43076
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 42923
diff changeset
    92
    coloriter = (
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 42923
diff changeset
    93
        s.encode(pycompat.sysstr(encoding.encoding), 'replace')
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 42923
diff changeset
    94
        for s in colorized.splitlines()
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 42923
diff changeset
    95
    )
6193
2344da8eb9b4 highlight: support annotate, and reduce layering violations.
Brendan Cully <brendan@kublai.com>
parents: 5991
diff changeset
    96
43077
687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals
Augie Fackler <augie@google.com>
parents: 43076
diff changeset
    97
    tmpl._filters[b'colorize'] = lambda x: next(coloriter)
6193
2344da8eb9b4 highlight: support annotate, and reduce layering violations.
Brendan Cully <brendan@kublai.com>
parents: 5991
diff changeset
    98
2344da8eb9b4 highlight: support annotate, and reduce layering violations.
Brendan Cully <brendan@kublai.com>
parents: 5991
diff changeset
    99
    oldl = tmpl.cache[field]
43077
687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals
Augie Fackler <augie@google.com>
parents: 43076
diff changeset
   100
    newl = oldl.replace(b'line|escape', b'line|colorize')
6193
2344da8eb9b4 highlight: support annotate, and reduce layering violations.
Brendan Cully <brendan@kublai.com>
parents: 5991
diff changeset
   101
    tmpl.cache[field] = newl