hgext/highlight/highlight.py
author Alexander Plavin <me@aplavin.ru>
Fri, 10 May 2013 00:26:39 +0400
branchstable
changeset 19169 bcdfb6078b9f
parent 18054 b35e3364f94a
child 23613 7b8ff3fd11d3
permissions -rw-r--r--
highlight: fix page layout with empty first and last lines Repeated newlines were stripped by pygmentize, now give the option not to do so.
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
40a06e39f010 extension for synax highlighting in the hgweb file revision view
Adam Hupp <adam@hupp.org>
parents:
diff changeset
    11
from mercurial import demandimport
10394
4612cded5176 fix coding style (reported by pylint)
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 10263
diff changeset
    12
demandimport.ignore.extend(['pkgutil', 'pkg_resources', '__main__'])
7948
de377b1a9a84 move encoding bits from util to encoding
Matt Mackall <mpm@selenic.com>
parents: 7120
diff changeset
    13
from mercurial import util, encoding
5532
40a06e39f010 extension for synax highlighting in the hgweb file revision view
Adam Hupp <adam@hupp.org>
parents:
diff changeset
    14
40a06e39f010 extension for synax highlighting in the hgweb file revision view
Adam Hupp <adam@hupp.org>
parents:
diff changeset
    15
from pygments import highlight
40a06e39f010 extension for synax highlighting in the hgweb file revision view
Adam Hupp <adam@hupp.org>
parents:
diff changeset
    16
from pygments.util import ClassNotFound
6199
0068809347d7 highlight: forgot import
Brendan Cully <brendan@kublai.com>
parents: 6198
diff changeset
    17
from pygments.lexers import guess_lexer, guess_lexer_for_filename, TextLexer
5532
40a06e39f010 extension for synax highlighting in the hgweb file revision view
Adam Hupp <adam@hupp.org>
parents:
diff changeset
    18
from pygments.formatters import HtmlFormatter
40a06e39f010 extension for synax highlighting in the hgweb file revision view
Adam Hupp <adam@hupp.org>
parents:
diff changeset
    19
6485
938319418d8c highlight: Generate pygments style sheet dynamically
Isaac Jurado <diptongo@gmail.com>
parents: 6394
diff changeset
    20
SYNTAX_CSS = ('\n<link rel="stylesheet" href="{url}highlightcss" '
5533
6cf7d7fe7d3d highlight: clean up coding style a little
Bryan O'Sullivan <bos@serpentine.com>
parents: 5532
diff changeset
    21
              'type="text/css" />')
5532
40a06e39f010 extension for synax highlighting in the hgweb file revision view
Adam Hupp <adam@hupp.org>
parents:
diff changeset
    22
6394
55bc0a035e1f highlight: some small cleanups
Dirkjan Ochtman <dirkjan@ochtman.nl>
parents: 6393
diff changeset
    23
def pygmentize(field, fctx, style, tmpl):
55bc0a035e1f highlight: some small cleanups
Dirkjan Ochtman <dirkjan@ochtman.nl>
parents: 6393
diff changeset
    24
5532
40a06e39f010 extension for synax highlighting in the hgweb file revision view
Adam Hupp <adam@hupp.org>
parents:
diff changeset
    25
    # append a <link ...> to the syntax highlighting css
10959
d1f4657f55e4 highlight: fix to work with caching templater
Matt Mackall <mpm@selenic.com>
parents: 10394
diff changeset
    26
    old_header = tmpl.load('header')
5532
40a06e39f010 extension for synax highlighting in the hgweb file revision view
Adam Hupp <adam@hupp.org>
parents:
diff changeset
    27
    if SYNTAX_CSS not in old_header:
40a06e39f010 extension for synax highlighting in the hgweb file revision view
Adam Hupp <adam@hupp.org>
parents:
diff changeset
    28
        new_header =  old_header + SYNTAX_CSS
5616
88ca3e0fb6e5 highlight: adapt to hgweb_mode refactoring
Christian Ebert <blacktrash@gmx.net>
parents: 5533
diff changeset
    29
        tmpl.cache['header'] = new_header
5532
40a06e39f010 extension for synax highlighting in the hgweb file revision view
Adam Hupp <adam@hupp.org>
parents:
diff changeset
    30
6194
fe54e7501de1 highlight: bail out if file is binary
Brendan Cully <brendan@kublai.com>
parents: 6193
diff changeset
    31
    text = fctx.data()
fe54e7501de1 highlight: bail out if file is binary
Brendan Cully <brendan@kublai.com>
parents: 6193
diff changeset
    32
    if util.binary(text):
fe54e7501de1 highlight: bail out if file is binary
Brendan Cully <brendan@kublai.com>
parents: 6193
diff changeset
    33
        return
fe54e7501de1 highlight: bail out if file is binary
Brendan Cully <brendan@kublai.com>
parents: 6193
diff changeset
    34
9424
799373ff2554 highlight: fixes garbled text in non-UTF-8 environment
Yuya Nishihara <yuya@tcha.org>
parents: 8360
diff changeset
    35
    # 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
    36
    # <http://pygments.org/docs/unicode/>
799373ff2554 highlight: fixes garbled text in non-UTF-8 environment
Yuya Nishihara <yuya@tcha.org>
parents: 8360
diff changeset
    37
    text = text.decode(encoding.encoding, 'replace')
7120
db7557359636 highlight: convert text to local before passing to pygmentize (issue1341)
Christian Ebert <blacktrash@gmx.net>
parents: 6938
diff changeset
    38
6193
2344da8eb9b4 highlight: support annotate, and reduce layering violations.
Brendan Cully <brendan@kublai.com>
parents: 5991
diff changeset
    39
    # 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
    40
    try:
19169
bcdfb6078b9f highlight: fix page layout with empty first and last lines
Alexander Plavin <me@aplavin.ru>
parents: 18054
diff changeset
    41
        lexer = guess_lexer_for_filename(fctx.path(), text[:1024],
bcdfb6078b9f highlight: fix page layout with empty first and last lines
Alexander Plavin <me@aplavin.ru>
parents: 18054
diff changeset
    42
                                         stripnl=False)
6494
c30849d4c8ba highlight: backward compatibility with pygments 0.5.1
Benoit Allard <benoit@aeteurope.nl>
parents: 6212
diff changeset
    43
    except (ClassNotFound, ValueError):
6198
358cc9cf54db highlight: guess by text when path name is ambiguous
Brendan Cully <brendan@kublai.com>
parents: 6197
diff changeset
    44
        try:
19169
bcdfb6078b9f highlight: fix page layout with empty first and last lines
Alexander Plavin <me@aplavin.ru>
parents: 18054
diff changeset
    45
            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
    46
        except (ClassNotFound, ValueError):
19169
bcdfb6078b9f highlight: fix page layout with empty first and last lines
Alexander Plavin <me@aplavin.ru>
parents: 18054
diff changeset
    47
            lexer = TextLexer(stripnl=False)
6193
2344da8eb9b4 highlight: support annotate, and reduce layering violations.
Brendan Cully <brendan@kublai.com>
parents: 5991
diff changeset
    48
9424
799373ff2554 highlight: fixes garbled text in non-UTF-8 environment
Yuya Nishihara <yuya@tcha.org>
parents: 8360
diff changeset
    49
    formatter = HtmlFormatter(style=style)
5532
40a06e39f010 extension for synax highlighting in the hgweb file revision view
Adam Hupp <adam@hupp.org>
parents:
diff changeset
    50
6193
2344da8eb9b4 highlight: support annotate, and reduce layering violations.
Brendan Cully <brendan@kublai.com>
parents: 5991
diff changeset
    51
    colorized = highlight(text, lexer, formatter)
2344da8eb9b4 highlight: support annotate, and reduce layering violations.
Brendan Cully <brendan@kublai.com>
parents: 5991
diff changeset
    52
    # strip wrapping div
2344da8eb9b4 highlight: support annotate, and reduce layering violations.
Brendan Cully <brendan@kublai.com>
parents: 5991
diff changeset
    53
    colorized = colorized[:colorized.find('\n</pre>')]
18054
b35e3364f94a check-code: there must also be whitespace between ')' and operator
Mads Kiilerich <madski@unity3d.com>
parents: 10959
diff changeset
    54
    colorized = colorized[colorized.find('<pre>') + 5:]
9424
799373ff2554 highlight: fixes garbled text in non-UTF-8 environment
Yuya Nishihara <yuya@tcha.org>
parents: 8360
diff changeset
    55
    coloriter = (s.encode(encoding.encoding, 'replace')
799373ff2554 highlight: fixes garbled text in non-UTF-8 environment
Yuya Nishihara <yuya@tcha.org>
parents: 8360
diff changeset
    56
                 for s in colorized.splitlines())
6193
2344da8eb9b4 highlight: support annotate, and reduce layering violations.
Brendan Cully <brendan@kublai.com>
parents: 5991
diff changeset
    57
8360
acc202b71619 templater: provide the standard template filters by default
Dirkjan Ochtman <dirkjan@ochtman.nl>
parents: 8251
diff changeset
    58
    tmpl.filters['colorize'] = lambda x: coloriter.next()
6193
2344da8eb9b4 highlight: support annotate, and reduce layering violations.
Brendan Cully <brendan@kublai.com>
parents: 5991
diff changeset
    59
2344da8eb9b4 highlight: support annotate, and reduce layering violations.
Brendan Cully <brendan@kublai.com>
parents: 5991
diff changeset
    60
    oldl = tmpl.cache[field]
2344da8eb9b4 highlight: support annotate, and reduce layering violations.
Brendan Cully <brendan@kublai.com>
parents: 5991
diff changeset
    61
    newl = oldl.replace('line|escape', 'line|colorize')
2344da8eb9b4 highlight: support annotate, and reduce layering violations.
Brendan Cully <brendan@kublai.com>
parents: 5991
diff changeset
    62
    tmpl.cache[field] = newl