diff options
Diffstat (limited to 'pygments/cmdline.py')
| -rw-r--r-- | pygments/cmdline.py | 221 |
1 files changed, 119 insertions, 102 deletions
diff --git a/pygments/cmdline.py b/pygments/cmdline.py index b5b6aab1..c77f0a32 100644 --- a/pygments/cmdline.py +++ b/pygments/cmdline.py @@ -5,20 +5,25 @@ Command line interface. - :copyright: Copyright 2006-2013 by the Pygments team, see AUTHORS. + :copyright: Copyright 2006-2014 by the Pygments team, see AUTHORS. :license: BSD, see LICENSE for details. """ + +from __future__ import print_function + import sys import getopt from textwrap import dedent from pygments import __version__, highlight -from pygments.util import ClassNotFound, OptionError, docstring_headline -from pygments.lexers import get_all_lexers, get_lexer_by_name, get_lexer_for_filename, \ - find_lexer_class, guess_lexer, TextLexer +from pygments.util import ClassNotFound, OptionError, docstring_headline, \ + guess_decode, guess_decode_from_terminal, terminal_encoding +from pygments.lexers import get_all_lexers, get_lexer_by_name, guess_lexer, \ + get_lexer_for_filename, find_lexer_class, TextLexer +from pygments.formatters.latex import LatexEmbeddedLexer, LatexFormatter from pygments.formatters import get_all_formatters, get_formatter_by_name, \ - get_formatter_for_filename, find_formatter_class, \ - TerminalFormatter # pylint:disable-msg=E0611 + get_formatter_for_filename, find_formatter_class, \ + TerminalFormatter # pylint:disable-msg=E0611 from pygments.filters import get_all_filters, find_filter_class from pygments.styles import get_all_styles, get_style_by_name @@ -97,7 +102,7 @@ def _parse_options(o_strs): for o_arg in o_args: o_arg = o_arg.strip() try: - o_key, o_val = o_arg.split('=') + o_key, o_val = o_arg.split('=', 1) o_key = o_key.strip() o_val = o_val.strip() except ValueError: @@ -124,25 +129,25 @@ def _print_help(what, name): try: if what == 'lexer': cls = find_lexer_class(name) - print "Help on the %s lexer:" % cls.name - print dedent(cls.__doc__) + print("Help on the %s lexer:" % cls.name) + print(dedent(cls.__doc__)) elif what == 'formatter': cls = find_formatter_class(name) - print "Help on the %s formatter:" % cls.name - print dedent(cls.__doc__) + print("Help on the %s formatter:" % cls.name) + print(dedent(cls.__doc__)) elif what == 'filter': cls = find_filter_class(name) - print "Help on the %s filter:" % name - print dedent(cls.__doc__) + print("Help on the %s filter:" % name) + print(dedent(cls.__doc__)) except AttributeError: - print >>sys.stderr, "%s not found!" % what + print("%s not found!" % what, file=sys.stderr) def _print_list(what): if what == 'lexer': - print - print "Lexers:" - print "~~~~~~~" + print() + print("Lexers:") + print("~~~~~~~") info = [] for fullname, names, exts, _ in get_all_lexers(): @@ -151,12 +156,12 @@ def _print_list(what): info.append(tup) info.sort() for i in info: - print ('* %s\n %s %s') % i + print(('* %s\n %s %s') % i) elif what == 'formatter': - print - print "Formatters:" - print "~~~~~~~~~~~" + print() + print("Formatters:") + print("~~~~~~~~~~~") info = [] for cls in get_all_formatters(): @@ -166,27 +171,27 @@ def _print_list(what): info.append(tup) info.sort() for i in info: - print ('* %s\n %s %s') % i + print(('* %s\n %s %s') % i) elif what == 'filter': - print - print "Filters:" - print "~~~~~~~~" + print() + print("Filters:") + print("~~~~~~~~") for name in get_all_filters(): cls = find_filter_class(name) - print "* " + name + ':' - print " %s" % docstring_headline(cls) + print("* " + name + ':') + print(" %s" % docstring_headline(cls)) elif what == 'style': - print - print "Styles:" - print "~~~~~~~" + print() + print("Styles:") + print("~~~~~~~") for name in get_all_styles(): cls = get_style_by_name(name) - print "* " + name + ':' - print " %s" % docstring_headline(cls) + print("* " + name + ':') + print(" %s" % docstring_headline(cls)) def main(args=sys.argv): @@ -207,8 +212,8 @@ def main(args=sys.argv): try: popts, args = getopt.getopt(args[1:], "l:f:F:o:O:P:LS:a:N:hVHgs") - except getopt.GetoptError, err: - print >>sys.stderr, usage + except getopt.GetoptError: + print(usage, file=sys.stderr) return 2 opts = {} O_opts = [] @@ -223,23 +228,19 @@ def main(args=sys.argv): F_opts.append(arg) opts[opt] = arg - if not opts and not args: - print usage - return 0 - if opts.pop('-h', None) is not None: - print usage + print(usage) return 0 if opts.pop('-V', None) is not None: - print 'Pygments version %s, (c) 2006-2013 by Georg Brandl.' % __version__ + print('Pygments version %s, (c) 2006-2014 by Georg Brandl.' % __version__) return 0 # handle ``pygmentize -L`` L_opt = opts.pop('-L', None) if L_opt is not None: if opts: - print >>sys.stderr, usage + print(usage, file=sys.stderr) return 2 # print version @@ -254,12 +255,12 @@ def main(args=sys.argv): H_opt = opts.pop('-H', None) if H_opt is not None: if opts or len(args) != 2: - print >>sys.stderr, usage + print(usage, file=sys.stderr) return 2 what, name = args if what not in ('lexer', 'formatter', 'filter'): - print >>sys.stderr, usage + print(usage, file=sys.stderr) return 2 _print_help(what, name) @@ -279,18 +280,22 @@ def main(args=sys.argv): parsed_opts[name] = value opts.pop('-P', None) + # encodings + inencoding = parsed_opts.get('inencoding', parsed_opts.get('encoding')) + outencoding = parsed_opts.get('outencoding', parsed_opts.get('encoding')) + # handle ``pygmentize -N`` infn = opts.pop('-N', None) if infn is not None: try: lexer = get_lexer_for_filename(infn, **parsed_opts) - except ClassNotFound, err: + except ClassNotFound as err: lexer = TextLexer() - except OptionError, err: - print >>sys.stderr, 'Error:', err + except OptionError as err: + print('Error:', err, file=sys.stderr) return 1 - print lexer.aliases[0] + print(lexer.aliases[0]) return 0 # handle ``pygmentize -S`` @@ -299,30 +304,30 @@ def main(args=sys.argv): if S_opt is not None: f_opt = opts.pop('-f', None) if not f_opt: - print >>sys.stderr, usage + print(usage, file=sys.stderr) return 2 if opts or args: - print >>sys.stderr, usage + print(usage, file=sys.stderr) return 2 try: parsed_opts['style'] = S_opt fmter = get_formatter_by_name(f_opt, **parsed_opts) - except ClassNotFound, err: - print >>sys.stderr, err + except ClassNotFound as err: + print(err, file=sys.stderr) return 1 arg = a_opt or '' try: - print fmter.get_style_defs(arg) - except Exception, err: - print >>sys.stderr, 'Error:', err + print(fmter.get_style_defs(arg)) + except Exception as err: + print('Error:', err, file=sys.stderr) return 1 return 0 # if no -S is given, -a is not allowed if a_opt is not None: - print >>sys.stderr, usage + print(usage, file=sys.stderr) return 2 # parse -F options @@ -335,39 +340,44 @@ def main(args=sys.argv): if fmter: try: fmter = get_formatter_by_name(fmter, **parsed_opts) - except (OptionError, ClassNotFound), err: - print >>sys.stderr, 'Error:', err + except (OptionError, ClassNotFound) as err: + print('Error:', err, file=sys.stderr) return 1 if outfn: if not fmter: try: fmter = get_formatter_for_filename(outfn, **parsed_opts) - except (OptionError, ClassNotFound), err: - print >>sys.stderr, 'Error:', err + except (OptionError, ClassNotFound) as err: + print('Error:', err, file=sys.stderr) return 1 try: outfile = open(outfn, 'wb') - except Exception, err: - print >>sys.stderr, 'Error: cannot open outfile:', err + except Exception as err: + print('Error: cannot open outfile:', err, file=sys.stderr) return 1 else: if not fmter: fmter = TerminalFormatter(**parsed_opts) - outfile = sys.stdout + if sys.version_info > (3,): + # Python 3: we have to use .buffer to get a binary stream + outfile = sys.stdout.buffer + else: + outfile = sys.stdout # select lexer lexer = opts.pop('-l', None) if lexer: try: lexer = get_lexer_by_name(lexer, **parsed_opts) - except (OptionError, ClassNotFound), err: - print >>sys.stderr, 'Error:', err + except (OptionError, ClassNotFound) as err: + print('Error:', err, file=sys.stderr) return 1 + # read input code if args: if len(args) > 1: - print >>sys.stderr, usage + print(usage, file=sys.stderr) return 2 #if '-s' in opts: @@ -377,58 +387,65 @@ def main(args=sys.argv): infn = args[0] try: - code = open(infn, 'rb').read() - except Exception, err: - print >>sys.stderr, 'Error: cannot read infile:', err + with open(infn, 'rb') as infp: + code = infp.read() + except Exception as err: + print('Error: cannot read infile:', err, file=sys.stderr) return 1 + if not inencoding: + code, inencoding = guess_decode(code) + # do we have to guess the lexer? if not lexer: try: lexer = get_lexer_for_filename(infn, code, **parsed_opts) - except ClassNotFound, err: + except ClassNotFound as err: if '-g' in opts: try: lexer = guess_lexer(code, **parsed_opts) except ClassNotFound: lexer = TextLexer(**parsed_opts) else: - print >>sys.stderr, 'Error:', err + print('Error:', err, file=sys.stderr) return 1 - except OptionError, err: - print >>sys.stderr, 'Error:', err + except OptionError as err: + print('Error:', err, file=sys.stderr) return 1 - elif '-s' not in opts: #treat stdin as full file (-s support is later) - if '-g' in opts: + elif '-s' not in opts: # treat stdin as full file (-s support is later) + # read code from terminal, always in binary mode since we want to + # decode ourselves and be tolerant with it + if sys.version_info > (3,): + # Python 3: we have to use .buffer to get a binary stream + code = sys.stdin.buffer.read() + else: code = sys.stdin.read() + if not inencoding: + code, inencoding = guess_decode_from_terminal(code, sys.stdin) + # else the lexer will do the decoding + if not lexer: try: lexer = guess_lexer(code, **parsed_opts) except ClassNotFound: lexer = TextLexer(**parsed_opts) - elif not lexer: - print >>sys.stderr, 'Error: no lexer name given and reading ' + \ - 'from stdin (try using -g or -l <lexer>)' - return 2 - else: - code = sys.stdin.read() - # No encoding given? Use latin1 if output file given, - # stdin/stdout encoding otherwise. - # (This is a compromise, I'm not too happy with it...) - if 'encoding' not in parsed_opts and 'outencoding' not in parsed_opts: + # When using the LaTeX formatter and the option `escapeinside` is + # specified, we need a special lexer which collects escaped text + # before running the chosen language lexer. + escapeinside = parsed_opts.get('escapeinside', '') + if len(escapeinside) == 2 and isinstance(fmter, LatexFormatter): + left = escapeinside[0] + right = escapeinside[1] + lexer = LatexEmbeddedLexer(left, right, lexer) + + # determine output encoding if not explicitly selected + if not outencoding: if outfn: - # encoding pass-through - fmter.encoding = 'latin1' + # output file? -> encoding pass-through + fmter.encoding = inencoding else: - if sys.version_info < (3,): - # use terminal encoding; Python 3's terminals already do that - lexer.encoding = getattr(sys.stdin, 'encoding', - None) or 'ascii' - fmter.encoding = getattr(sys.stdout, 'encoding', - None) or 'ascii' - elif not outfn and sys.version_info > (3,): - # output to terminal with encoding -> use .buffer - outfile = sys.stdout.buffer + # else use terminal encoding + fmter.encoding = terminal_encoding(sys.stdout) # ... and do it! try: @@ -437,12 +454,12 @@ def main(args=sys.argv): lexer.add_filter(fname, **fopts) if '-s' not in opts: - #process whole input as per normal... + # process whole input as per normal... highlight(code, lexer, fmter, outfile) else: - # line input processing of stdout (eg: for 'tail -f')... + # line by line processing of stdin (eg: for 'tail -f')... try: - while True: + while 1: line = sys.stdin.readline() if not line: break @@ -450,16 +467,16 @@ def main(args=sys.argv): except KeyboardInterrupt: return 0 - except Exception, err: + except Exception: import traceback info = traceback.format_exception(*sys.exc_info()) msg = info[-1].strip() if len(info) >= 3: # extract relevant file and position info msg += '\n (f%s)' % info[-2].split('\n')[0].strip()[1:] - print >>sys.stderr - print >>sys.stderr, '*** Error while highlighting:' - print >>sys.stderr, msg + print(file=sys.stderr) + print('*** Error while highlighting:', file=sys.stderr) + print(msg, file=sys.stderr) return 1 return 0 |