minirst: make format() simply return a formatted text

It's a source of bugs to change the type of the return value conditionally.

mercurial/minirst.py

@@ -1,835 +1,831 b''
 # minirst.py - minimal reStructuredText parser
 #
 # Copyright 2009, 2010 Matt Mackall <[e-mail]> and others
 #
 # This software may be used and distributed according to the terms of the
 # GNU General Public License version 2 or any later version.
 
 """simplified reStructuredText parser.
 
 This parser knows just enough about reStructuredText to parse the
 Mercurial docstrings.
 
 It cheats in a major way: nested blocks are not really nested. They
 are just indented blocks that look like they are nested. This relies
 on the user to keep the right indentation for the blocks.
 
 Remember to update https://mercurial-scm.org/wiki/HelpStyleGuide
 when adding support for new constructs.
 """
 
 from __future__ import absolute_import
 
 import re
 
 from .i18n import _
 from . import (
     encoding,
     pycompat,
     url,
 )
 from .utils import (
     stringutil,
 )
 
 def section(s):
     return "%s\n%s\n\n" % (s, "\"" * encoding.colwidth(s))
 
 def subsection(s):
     return "%s\n%s\n\n" % (s, '=' * encoding.colwidth(s))
 
 def subsubsection(s):
     return "%s\n%s\n\n" % (s, "-" * encoding.colwidth(s))
 
 def subsubsubsection(s):
     return "%s\n%s\n\n" % (s, "." * encoding.colwidth(s))
 
 def replace(text, substs):
     '''
     Apply a list of (find, replace) pairs to a text.
 
     >>> replace(b"foo bar", [(b'f', b'F'), (b'b', b'B')])
     'Foo Bar'
     >>> encoding.encoding = b'latin1'
     >>> replace(b'\\x81\\\\', [(b'\\\\', b'/')])
     '\\x81/'
     >>> encoding.encoding = b'shiftjis'
     >>> replace(b'\\x81\\\\', [(b'\\\\', b'/')])
     '\\x81\\\\'
     '''
 
     # some character encodings (cp932 for Japanese, at least) use
     # ASCII characters other than control/alphabet/digit as a part of
     # multi-bytes characters, so direct replacing with such characters
     # on strings in local encoding causes invalid byte sequences.
     utext = text.decode(pycompat.sysstr(encoding.encoding))
     for f, t in substs:
         utext = utext.replace(f.decode("ascii"), t.decode("ascii"))
     return utext.encode(pycompat.sysstr(encoding.encoding))
 
 _blockre = re.compile(br"\n(?:\s*\n)+")
 
 def findblocks(text):
     """Find continuous blocks of lines in text.
 
     Returns a list of dictionaries representing the blocks. Each block
     has an 'indent' field and a 'lines' field.
     """
     blocks = []
     for b in _blockre.split(text.lstrip('\n').rstrip()):
         lines = b.splitlines()
         if lines:
             indent = min((len(l) - len(l.lstrip())) for l in lines)
             lines = [l[indent:] for l in lines]
             blocks.append({'indent': indent, 'lines': lines})
     return blocks
 
 def findliteralblocks(blocks):
     """Finds literal blocks and adds a 'type' field to the blocks.
 
     Literal blocks are given the type 'literal', all other blocks are
     given type the 'paragraph'.
     """
     i = 0
     while i < len(blocks):
         # Searching for a block that looks like this:
         #
         # +------------------------------+
         # | paragraph                    |
         # | (ends with "::")             |
         # +------------------------------+
         #    +---------------------------+
         #    | indented literal block    |
         #    +---------------------------+
         blocks[i]['type'] = 'paragraph'
         if blocks[i]['lines'][-1].endswith('::') and i + 1 < len(blocks):
             indent = blocks[i]['indent']
             adjustment = blocks[i + 1]['indent'] - indent
 
             if blocks[i]['lines'] == ['::']:
                 # Expanded form: remove block
                 del blocks[i]
                 i -= 1
             elif blocks[i]['lines'][-1].endswith(' ::'):
                 # Partially minimized form: remove space and both
                 # colons.
                 blocks[i]['lines'][-1] = blocks[i]['lines'][-1][:-3]
             elif len(blocks[i]['lines']) == 1 and \
                  blocks[i]['lines'][0].lstrip(' ').startswith('.. ') and \
                  blocks[i]['lines'][0].find(' ', 3) == -1:
                 # directive on its own line, not a literal block
                 i += 1
                 continue
             else:
                 # Fully minimized form: remove just one colon.
                 blocks[i]['lines'][-1] = blocks[i]['lines'][-1][:-1]
 
             # List items are formatted with a hanging indent. We must
             # correct for this here while we still have the original
             # information on the indentation of the subsequent literal
             # blocks available.
             m = _bulletre.match(blocks[i]['lines'][0])
             if m:
                 indent += m.end()
                 adjustment -= m.end()
 
             # Mark the following indented blocks.
             while i + 1 < len(blocks) and blocks[i + 1]['indent'] > indent:
                 blocks[i + 1]['type'] = 'literal'
                 blocks[i + 1]['indent'] -= adjustment
                 i += 1
         i += 1
     return blocks
 
 _bulletre = re.compile(br'(\*|-|[0-9A-Za-z]+\.|\(?[0-9A-Za-z]+\)|\|) ')
 _optionre = re.compile(br'^(-([a-zA-Z0-9]), )?(--[a-z0-9-]+)'
                        br'((.*)  +)(.*)$')
 _fieldre = re.compile(br':(?![: ])([^:]*)(?<! ):[ ]+(.*)')
 _definitionre = re.compile(br'[^ ]')
 _tablere = re.compile(br'(=+\s+)*=+')
 
 def splitparagraphs(blocks):
     """Split paragraphs into lists."""
     # Tuples with (list type, item regexp, single line items?). Order
     # matters: definition lists has the least specific regexp and must
     # come last.
     listtypes = [('bullet', _bulletre, True),
                  ('option', _optionre, True),
                  ('field', _fieldre, True),
                  ('definition', _definitionre, False)]
 
     def match(lines, i, itemre, singleline):
         """Does itemre match an item at line i?
 
         A list item can be followed by an indented line or another list
         item (but only if singleline is True).
         """
         line1 = lines[i]
         line2 = i + 1 < len(lines) and lines[i + 1] or ''
         if not itemre.match(line1):
             return False
         if singleline:
             return line2 == '' or line2[0:1] == ' ' or itemre.match(line2)
         else:
             return line2.startswith(' ')
 
     i = 0
     while i < len(blocks):
         if blocks[i]['type'] == 'paragraph':
             lines = blocks[i]['lines']
             for type, itemre, singleline in listtypes:
                 if match(lines, 0, itemre, singleline):
                     items = []
                     for j, line in enumerate(lines):
                         if match(lines, j, itemre, singleline):
                             items.append({'type': type, 'lines': [],
                                           'indent': blocks[i]['indent']})
                         items[-1]['lines'].append(line)
                     blocks[i:i + 1] = items
                     break
         i += 1
     return blocks
 
 _fieldwidth = 14
 
 def updatefieldlists(blocks):
     """Find key for field lists."""
     i = 0
     while i < len(blocks):
         if blocks[i]['type'] != 'field':
             i += 1
             continue
 
         j = i
         while j < len(blocks) and blocks[j]['type'] == 'field':
             m = _fieldre.match(blocks[j]['lines'][0])
             key, rest = m.groups()
             blocks[j]['lines'][0] = rest
             blocks[j]['key'] = key
             j += 1
 
         i = j + 1
 
     return blocks
 
 def updateoptionlists(blocks):
     i = 0
     while i < len(blocks):
         if blocks[i]['type'] != 'option':
             i += 1
             continue
 
         optstrwidth = 0
         j = i
         while j < len(blocks) and blocks[j]['type'] == 'option':
             m = _optionre.match(blocks[j]['lines'][0])
 
             shortoption = m.group(2)
             group3 = m.group(3)
             longoption = group3[2:].strip()
             desc = m.group(6).strip()
             longoptionarg = m.group(5).strip()
             blocks[j]['lines'][0] = desc
 
             noshortop = ''
             if not shortoption:
                 noshortop = '   '
 
             opt = "%s%s" %   (shortoption and "-%s " % shortoption or '',
                             ("%s--%s %s") % (noshortop, longoption,
                                              longoptionarg))
             opt = opt.rstrip()
             blocks[j]['optstr'] = opt
             optstrwidth = max(optstrwidth, encoding.colwidth(opt))
             j += 1
 
         for block in blocks[i:j]:
             block['optstrwidth'] = optstrwidth
         i = j + 1
     return blocks
 
 def prunecontainers(blocks, keep):
     """Prune unwanted containers.
 
     The blocks must have a 'type' field, i.e., they should have been
     run through findliteralblocks first.
     """
     pruned = []
     i = 0
     while i + 1 < len(blocks):
         # Searching for a block that looks like this:
         #
         # +-------+---------------------------+
         # | ".. container ::" type            |
         # +---+                               |
         #     | blocks                        |
         #     +-------------------------------+
         if (blocks[i]['type'] == 'paragraph' and
             blocks[i]['lines'][0].startswith('.. container::')):
             indent = blocks[i]['indent']
             adjustment = blocks[i + 1]['indent'] - indent
             containertype = blocks[i]['lines'][0][15:]
             prune = True
             for c in keep:
                 if c in containertype.split('.'):
                     prune = False
             if prune:
                 pruned.append(containertype)
 
             # Always delete "..container:: type" block
             del blocks[i]
             j = i
             i -= 1
             while j < len(blocks) and blocks[j]['indent'] > indent:
                 if prune:
                     del blocks[j]
                 else:
                     blocks[j]['indent'] -= adjustment
                     j += 1
         i += 1
     return blocks, pruned
 
 _sectionre = re.compile(br"""^([-=`:.'"~^_*+#])\1+$""")
 
 def findtables(blocks):
     '''Find simple tables
 
        Only simple one-line table elements are supported
     '''
 
     for block in blocks:
         # Searching for a block that looks like this:
         #
         # === ==== ===
         #  A    B   C
         # === ==== ===  <- optional
         #  1    2   3
         #  x    y   z
         # === ==== ===
         if (block['type'] == 'paragraph' and
             len(block['lines']) > 2 and
             _tablere.match(block['lines'][0]) and
             block['lines'][0] == block['lines'][-1]):
             block['type'] = 'table'
             block['header'] = False
             div = block['lines'][0]
 
             # column markers are ASCII so we can calculate column
             # position in bytes
             columns = [x for x in pycompat.xrange(len(div))
                        if div[x:x + 1] == '=' and (x == 0 or
                                                    div[x - 1:x] == ' ')]
             rows = []
             for l in block['lines'][1:-1]:
                 if l == div:
                     block['header'] = True
                     continue
                 row = []
                 # we measure columns not in bytes or characters but in
                 # colwidth which makes things tricky
                 pos = columns[0] # leading whitespace is bytes
                 for n, start in enumerate(columns):
                     if n + 1 < len(columns):
                         width = columns[n + 1] - start
                         v = encoding.getcols(l, pos, width) # gather columns
                         pos += len(v) # calculate byte position of end
                         row.append(v.strip())
                     else:
                         row.append(l[pos:].strip())
                 rows.append(row)
 
             block['table'] = rows
 
     return blocks
 
 def findsections(blocks):
     """Finds sections.
 
     The blocks must have a 'type' field, i.e., they should have been
     run through findliteralblocks first.
     """
     for block in blocks:
         # Searching for a block that looks like this:
         #
         # +------------------------------+
         # | Section title                |
         # | -------------                |
         # +------------------------------+
         if (block['type'] == 'paragraph' and
             len(block['lines']) == 2 and
             encoding.colwidth(block['lines'][0]) == len(block['lines'][1]) and
             _sectionre.match(block['lines'][1])):
             block['underline'] = block['lines'][1][0:1]
             block['type'] = 'section'
             del block['lines'][1]
     return blocks
 
 def inlineliterals(blocks):
     substs = [('``', '"')]
     for b in blocks:
         if b['type'] in ('paragraph', 'section'):
             b['lines'] = [replace(l, substs) for l in b['lines']]
     return blocks
 
 def hgrole(blocks):
     substs = [(':hg:`', "'hg "), ('`', "'")]
     for b in blocks:
         if b['type'] in ('paragraph', 'section'):
             # Turn :hg:`command` into "hg command". This also works
             # when there is a line break in the command and relies on
             # the fact that we have no stray back-quotes in the input
             # (run the blocks through inlineliterals first).
             b['lines'] = [replace(l, substs) for l in b['lines']]
     return blocks
 
 def addmargins(blocks):
     """Adds empty blocks for vertical spacing.
 
     This groups bullets, options, and definitions together with no vertical
     space between them, and adds an empty block between all other blocks.
     """
     i = 1
     while i < len(blocks):
         if (blocks[i]['type'] == blocks[i - 1]['type'] and
             blocks[i]['type'] in ('bullet', 'option', 'field')):
             i += 1
         elif not blocks[i - 1]['lines']:
             # no lines in previous block, do not separate
             i += 1
         else:
             blocks.insert(i, {'lines': [''], 'indent': 0, 'type': 'margin'})
             i += 2
     return blocks
 
 def prunecomments(blocks):
     """Remove comments."""
     i = 0
     while i < len(blocks):
         b = blocks[i]
         if b['type'] == 'paragraph' and (b['lines'][0].startswith('.. ') or
                                          b['lines'] == ['..']):
             del blocks[i]
             if i < len(blocks) and blocks[i]['type'] == 'margin':
                 del blocks[i]
         else:
             i += 1
     return blocks
 
 
 def findadmonitions(blocks, admonitions=None):
     """
     Makes the type of the block an admonition block if
     the first line is an admonition directive
     """
     admonitions = admonitions or _admonitiontitles.keys()
 
     admonitionre = re.compile(br'\.\. (%s)::' % '|'.join(sorted(admonitions)),
                               flags=re.IGNORECASE)
 
     i = 0
     while i < len(blocks):
         m = admonitionre.match(blocks[i]['lines'][0])
         if m:
             blocks[i]['type'] = 'admonition'
             admonitiontitle = blocks[i]['lines'][0][3:m.end() - 2].lower()
 
             firstline = blocks[i]['lines'][0][m.end() + 1:]
             if firstline:
                 blocks[i]['lines'].insert(1, '   ' + firstline)
 
             blocks[i]['admonitiontitle'] = admonitiontitle
             del blocks[i]['lines'][0]
         i = i + 1
     return blocks
 
 _admonitiontitles = {
     'attention': _('Attention:'),
     'caution': _('Caution:'),
     'danger': _('!Danger!'),
     'error': _('Error:'),
     'hint': _('Hint:'),
     'important': _('Important:'),
     'note': _('Note:'),
     'tip': _('Tip:'),
     'warning': _('Warning!'),
 }
 
 def formatoption(block, width):
     desc = ' '.join(map(bytes.strip, block['lines']))
     colwidth = encoding.colwidth(block['optstr'])
     usablewidth = width - 1
     hanging = block['optstrwidth']
     initindent = '%s%s  ' % (block['optstr'], ' ' * ((hanging - colwidth)))
     hangindent = ' ' * (encoding.colwidth(initindent) + 1)
     return ' %s\n' % (stringutil.wrap(desc, usablewidth,
                                       initindent=initindent,
                                       hangindent=hangindent))
 
 def formatblock(block, width):
     """Format a block according to width."""
     if width <= 0:
         width = 78
     indent = ' ' * block['indent']
     if block['type'] == 'admonition':
         admonition = _admonitiontitles[block['admonitiontitle']]
         if not block['lines']:
             return indent + admonition + '\n'
         hang = len(block['lines'][-1]) - len(block['lines'][-1].lstrip())
 
         defindent = indent + hang * ' '
         text = ' '.join(map(bytes.strip, block['lines']))
         return '%s\n%s\n' % (indent + admonition,
                              stringutil.wrap(text, width=width,
                                              initindent=defindent,
                                              hangindent=defindent))
     if block['type'] == 'margin':
         return '\n'
     if block['type'] == 'literal':
         indent += '  '
         return indent + ('\n' + indent).join(block['lines']) + '\n'
     if block['type'] == 'section':
         underline = encoding.colwidth(block['lines'][0]) * block['underline']
         return "%s%s\n%s%s\n" % (indent, block['lines'][0],indent, underline)
     if block['type'] == 'table':
         table = block['table']
         # compute column widths
         widths = [max([encoding.colwidth(e) for e in c]) for c in zip(*table)]
         text = ''
         span = sum(widths) + len(widths) - 1
         indent = ' ' * block['indent']
         hang = ' ' * (len(indent) + span - widths[-1])
 
         for row in table:
             l = []
             for w, v in zip(widths, row):
                 pad = ' ' * (w - encoding.colwidth(v))
                 l.append(v + pad)
             l = ' '.join(l)
             l = stringutil.wrap(l, width=width,
                                 initindent=indent,
                                 hangindent=hang)
             if not text and block['header']:
                 text = l + '\n' + indent + '-' * (min(width, span)) + '\n'
             else:
                 text += l + "\n"
         return text
     if block['type'] == 'definition':
         term = indent + block['lines'][0]
         hang = len(block['lines'][-1]) - len(block['lines'][-1].lstrip())
         defindent = indent + hang * ' '
         text = ' '.join(map(bytes.strip, block['lines'][1:]))
         return '%s\n%s\n' % (term, stringutil.wrap(text, width=width,
                                                    initindent=defindent,
                                                    hangindent=defindent))
     subindent = indent
     if block['type'] == 'bullet':
         if block['lines'][0].startswith('| '):
             # Remove bullet for line blocks and add no extra
             # indentation.
             block['lines'][0] = block['lines'][0][2:]
         else:
             m = _bulletre.match(block['lines'][0])
             subindent = indent + m.end() * ' '
     elif block['type'] == 'field':
         key = block['key']
         subindent = indent + _fieldwidth * ' '
         if len(key) + 2 > _fieldwidth:
             # key too large, use full line width
             key = key.ljust(width)
         else:
             # key fits within field width
             key = key.ljust(_fieldwidth)
         block['lines'][0] = key + block['lines'][0]
     elif block['type'] == 'option':
         return formatoption(block, width)
 
     text = ' '.join(map(bytes.strip, block['lines']))
     return stringutil.wrap(text, width=width,
                            initindent=indent,
                            hangindent=subindent) + '\n'
 
 def formathtml(blocks):
     """Format RST blocks as HTML"""
 
     out = []
     headernest = ''
     listnest = []
 
     def escape(s):
         return url.escape(s, True)
 
     def openlist(start, level):
         if not listnest or listnest[-1][0] != start:
             listnest.append((start, level))
             out.append('<%s>\n' % start)
 
     blocks = [b for b in blocks if b['type'] != 'margin']
 
     for pos, b in enumerate(blocks):
         btype = b['type']
         level = b['indent']
         lines = b['lines']
 
         if btype == 'admonition':
             admonition = escape(_admonitiontitles[b['admonitiontitle']])
             text = escape(' '.join(map(bytes.strip, lines)))
             out.append('<p>\n<b>%s</b> %s\n</p>\n' % (admonition, text))
         elif btype == 'paragraph':
             out.append('<p>\n%s\n</p>\n' % escape('\n'.join(lines)))
         elif btype == 'margin':
             pass
         elif btype == 'literal':
             out.append('<pre>\n%s\n</pre>\n' % escape('\n'.join(lines)))
         elif btype == 'section':
             i = b['underline']
             if i not in headernest:
                 headernest += i
             level = headernest.index(i) + 1
             out.append('<h%d>%s</h%d>\n' % (level, escape(lines[0]), level))
         elif btype == 'table':
             table = b['table']
             out.append('<table>\n')
             for row in table:
                 out.append('<tr>')
                 for v in row:
                     out.append('<td>')
                     out.append(escape(v))
                     out.append('</td>')
                     out.append('\n')
                 out.pop()
                 out.append('</tr>\n')
             out.append('</table>\n')
         elif btype == 'definition':
             openlist('dl', level)
             term = escape(lines[0])
             text = escape(' '.join(map(bytes.strip, lines[1:])))
             out.append(' <dt>%s\n <dd>%s\n' % (term, text))
         elif btype == 'bullet':
             bullet, head = lines[0].split(' ', 1)
             if bullet in ('*', '-'):
                 openlist('ul', level)
             else:
                 openlist('ol', level)
             out.append(' <li> %s\n' % escape(' '.join([head] + lines[1:])))
         elif btype == 'field':
             openlist('dl', level)
             key = escape(b['key'])
             text = escape(' '.join(map(bytes.strip, lines)))
             out.append(' <dt>%s\n <dd>%s\n' % (key, text))
         elif btype == 'option':
             openlist('dl', level)
             opt = escape(b['optstr'])
             desc = escape(' '.join(map(bytes.strip, lines)))
             out.append(' <dt>%s\n <dd>%s\n' % (opt, desc))
 
         # close lists if indent level of next block is lower
         if listnest:
             start, level = listnest[-1]
             if pos == len(blocks) - 1:
                 out.append('</%s>\n' % start)
                 listnest.pop()
             else:
                 nb = blocks[pos + 1]
                 ni = nb['indent']
                 if (ni < level or
                     (ni == level and
                      nb['type'] not in 'definition bullet field option')):
                     out.append('</%s>\n' % start)
                     listnest.pop()
 
     return ''.join(out)
 
 def parse(text, indent=0, keep=None, admonitions=None):
     """Parse text into a list of blocks"""
     pruned = []
     blocks = findblocks(text)
     for b in blocks:
         b['indent'] += indent
     blocks = findliteralblocks(blocks)
     blocks = findtables(blocks)
     blocks, pruned = prunecontainers(blocks, keep or [])
     blocks = findsections(blocks)
     blocks = inlineliterals(blocks)
     blocks = hgrole(blocks)
     blocks = splitparagraphs(blocks)
     blocks = updatefieldlists(blocks)
     blocks = updateoptionlists(blocks)
     blocks = findadmonitions(blocks, admonitions=admonitions)
     blocks = addmargins(blocks)
     blocks = prunecomments(blocks)
     return blocks, pruned
 
 def formatblocks(blocks, width):
     text = ''.join(formatblock(b, width) for b in blocks)
     return text
 
 def formatplain(blocks, width):
     """Format parsed blocks as plain text"""
     return ''.join(formatblock(b, width) for b in blocks)
 
 def format(text, width=80, indent=0, keep=None, style='plain', section=None):
     """Parse and format the text according to width."""
     blocks, pruned = parse(text, indent, keep or [])
     if section:
         blocks = filtersections(blocks, section)
     if style == 'html':
-        text = formathtml(blocks)
+        return formathtml(blocks)
     else:
-        text = formatplain(blocks, width=width)
-    if keep is None:
-        return text
-    else:
-        return text, pruned
+        return formatplain(blocks, width=width)
 
 def filtersections(blocks, section):
     """Select parsed blocks under the specified section"""
     parents = []
     sections = getsections(blocks)
     blocks = []
     i = 0
     lastparents = []
     synthetic = []
     collapse = True
     while i < len(sections):
         name, nest, b = sections[i]
         del parents[nest:]
         parents.append(i)
         if name == section:
             if lastparents != parents:
                 llen = len(lastparents)
                 plen = len(parents)
                 if llen and llen != plen:
                     collapse = False
                 s = []
                 for j in pycompat.xrange(3, plen - 1):
                     parent = parents[j]
                     if (j >= llen or
                         lastparents[j] != parent):
                         s.append(len(blocks))
                         sec = sections[parent][2]
                         blocks.append(sec[0])
                         blocks.append(sec[-1])
                 if s:
                     synthetic.append(s)
 
             lastparents = parents[:]
             blocks.extend(b)
 
             ## Also show all subnested sections
             while i + 1 < len(sections) and sections[i + 1][1] > nest:
                 i += 1
                 blocks.extend(sections[i][2])
         i += 1
     if collapse:
         synthetic.reverse()
         for s in synthetic:
             path = [blocks[syn]['lines'][0] for syn in s]
             real = s[-1] + 2
             realline = blocks[real]['lines']
             realline[0] = ('"%s"' %
                            '.'.join(path + [realline[0]]).replace('"', ''))
             del blocks[s[0]:real]
 
     return blocks
 
 def getsections(blocks):
     '''return a list of (section name, nesting level, blocks) tuples'''
     nest = ""
     level = 0
     secs = []
 
     def getname(b):
         if b['type'] == 'field':
             x = b['key']
         else:
             x = b['lines'][0]
         x = encoding.lower(x).strip('"')
         if '(' in x:
             x = x.split('(')[0]
         return x
 
     for b in blocks:
         if b['type'] == 'section':
             i = b['underline']
             if i not in nest:
                 nest += i
             level = nest.index(i) + 1
             nest = nest[:level]
             secs.append((getname(b), level, [b]))
         elif b['type'] in ('definition', 'field'):
             i = ' '
             if i not in nest:
                 nest += i
             level = nest.index(i) + 1
             nest = nest[:level]
             for i in range(1, len(secs) + 1):
                 sec = secs[-i]
                 if sec[1] < level:
                     break
                 siblings = [a for a in sec[2] if a['type'] == 'definition']
                 if siblings:
                     siblingindent = siblings[-1]['indent']
                     indent = b['indent']
                     if siblingindent < indent:
                         level += 1
                         break
                     elif siblingindent == indent:
                         level = sec[1]
                         break
             secs.append((getname(b), level, [b]))
         else:
             if not secs:
                 # add an initial empty section
                 secs = [('', 0, [])]
             if b['type'] != 'margin':
                 pointer = 1
                 bindent = b['indent']
                 while pointer < len(secs):
                     section = secs[-pointer][2][0]
                     if section['type'] != 'margin':
                         sindent = section['indent']
                         if len(section['lines']) > 1:
                             sindent += len(section['lines'][1]) - \
                               len(section['lines'][1].lstrip(' '))
                         if bindent >= sindent:
                             break
                     pointer += 1
                 if pointer > 1:
                     blevel = secs[-pointer][1]
                     if section['type'] != b['type']:
                         blevel += 1
                     secs.append(('', blevel, []))
             secs[-1][2].append(b)
     return secs
 
 def decorateblocks(blocks, width):
     '''generate a list of (section name, line text) pairs for search'''
     lines = []
     for s in getsections(blocks):
         section = s[0]
         text = formatblocks(s[2], width)
         lines.append([(section, l) for l in text.splitlines(True)])
     return lines
 
 def maketable(data, indent=0, header=False):
     '''Generate an RST table for the given table data as a list of lines'''
 
     widths = [max(encoding.colwidth(e) for e in c) for c in zip(*data)]
     indent = ' ' * indent
     div = indent + ' '.join('=' * w for w in widths) + '\n'
 
     out = [div]
     for row in data:
         l = []
         for w, v in zip(widths, row):
             if '\n' in v:
                 # only remove line breaks and indentation, long lines are
                 # handled by the next tool
                 v = ' '.join(e.lstrip() for e in v.split('\n'))
             pad = ' ' * (w - encoding.colwidth(v))
             l.append(v + pad)
         out.append(indent + ' '.join(l) + "\n")
     if header and len(data) > 1:
         out.insert(2, div)
     out.append(div)
     return out

mercurial/templatefuncs.py

@@ -1,718 +1,718 b''
 # templatefuncs.py - common template functions
 #
 # Copyright 2005, 2006 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.
 
 from __future__ import absolute_import
 
 import re
 
 from .i18n import _
 from .node import (
     bin,
     wdirid,
 )
 from . import (
     color,
     encoding,
     error,
     minirst,
     obsutil,
     registrar,
     revset as revsetmod,
     revsetlang,
     scmutil,
     templatefilters,
     templatekw,
     templateutil,
     util,
 )
 from .utils import (
     dateutil,
     stringutil,
 )
 
 evalrawexp = templateutil.evalrawexp
 evalwrapped = templateutil.evalwrapped
 evalfuncarg = templateutil.evalfuncarg
 evalboolean = templateutil.evalboolean
 evaldate = templateutil.evaldate
 evalinteger = templateutil.evalinteger
 evalstring = templateutil.evalstring
 evalstringliteral = templateutil.evalstringliteral
 
 # dict of template built-in functions
 funcs = {}
 templatefunc = registrar.templatefunc(funcs)
 
 @templatefunc('date(date[, fmt])')
 def date(context, mapping, args):
     """Format a date. See :hg:`help dates` for formatting
     strings. The default is a Unix date format, including the timezone:
     "Mon Sep 04 15:13:13 2006 0700"."""
     if not (1 <= len(args) <= 2):
         # i18n: "date" is a keyword
         raise error.ParseError(_("date expects one or two arguments"))
 
     date = evaldate(context, mapping, args[0],
                     # i18n: "date" is a keyword
                     _("date expects a date information"))
     fmt = None
     if len(args) == 2:
         fmt = evalstring(context, mapping, args[1])
     if fmt is None:
         return dateutil.datestr(date)
     else:
         return dateutil.datestr(date, fmt)
 
 @templatefunc('dict([[key=]value...])', argspec='*args **kwargs')
 def dict_(context, mapping, args):
     """Construct a dict from key-value pairs. A key may be omitted if
     a value expression can provide an unambiguous name."""
     data = util.sortdict()
 
     for v in args['args']:
         k = templateutil.findsymbolicname(v)
         if not k:
             raise error.ParseError(_('dict key cannot be inferred'))
         if k in data or k in args['kwargs']:
             raise error.ParseError(_("duplicated dict key '%s' inferred") % k)
         data[k] = evalfuncarg(context, mapping, v)
 
     data.update((k, evalfuncarg(context, mapping, v))
                 for k, v in args['kwargs'].iteritems())
     return templateutil.hybriddict(data)
 
 @templatefunc('diff([includepattern [, excludepattern]])', requires={'ctx'})
 def diff(context, mapping, args):
     """Show a diff, optionally
     specifying files to include or exclude."""
     if len(args) > 2:
         # i18n: "diff" is a keyword
         raise error.ParseError(_("diff expects zero, one, or two arguments"))
 
     def getpatterns(i):
         if i < len(args):
             s = evalstring(context, mapping, args[i]).strip()
             if s:
                 return [s]
         return []
 
     ctx = context.resource(mapping, 'ctx')
     chunks = ctx.diff(match=ctx.match([], getpatterns(0), getpatterns(1)))
 
     return ''.join(chunks)
 
 @templatefunc('extdata(source)', argspec='source', requires={'ctx', 'cache'})
 def extdata(context, mapping, args):
     """Show a text read from the specified extdata source. (EXPERIMENTAL)"""
     if 'source' not in args:
         # i18n: "extdata" is a keyword
         raise error.ParseError(_('extdata expects one argument'))
 
     source = evalstring(context, mapping, args['source'])
     if not source:
         sym = templateutil.findsymbolicname(args['source'])
         if sym:
             raise error.ParseError(_('empty data source specified'),
                                    hint=_("did you mean extdata('%s')?") % sym)
         else:
             raise error.ParseError(_('empty data source specified'))
     cache = context.resource(mapping, 'cache').setdefault('extdata', {})
     ctx = context.resource(mapping, 'ctx')
     if source in cache:
         data = cache[source]
     else:
         data = cache[source] = scmutil.extdatasource(ctx.repo(), source)
     return data.get(ctx.rev(), '')
 
 @templatefunc('files(pattern)', requires={'ctx'})
 def files(context, mapping, args):
     """All files of the current changeset matching the pattern. See
     :hg:`help patterns`."""
     if not len(args) == 1:
         # i18n: "files" is a keyword
         raise error.ParseError(_("files expects one argument"))
 
     raw = evalstring(context, mapping, args[0])
     ctx = context.resource(mapping, 'ctx')
     m = ctx.match([raw])
     files = list(ctx.matches(m))
     return templateutil.compatlist(context, mapping, "file", files)
 
 @templatefunc('fill(text[, width[, initialident[, hangindent]]])')
 def fill(context, mapping, args):
     """Fill many
     paragraphs with optional indentation. See the "fill" filter."""
     if not (1 <= len(args) <= 4):
         # i18n: "fill" is a keyword
         raise error.ParseError(_("fill expects one to four arguments"))
 
     text = evalstring(context, mapping, args[0])
     width = 76
     initindent = ''
     hangindent = ''
     if 2 <= len(args) <= 4:
         width = evalinteger(context, mapping, args[1],
                             # i18n: "fill" is a keyword
                             _("fill expects an integer width"))
         try:
             initindent = evalstring(context, mapping, args[2])
             hangindent = evalstring(context, mapping, args[3])
         except IndexError:
             pass
 
     return templatefilters.fill(text, width, initindent, hangindent)
 
 @templatefunc('filter(iterable[, expr])')
 def filter_(context, mapping, args):
     """Remove empty elements from a list or a dict. If expr specified, it's
     applied to each element to test emptiness."""
     if not (1 <= len(args) <= 2):
         # i18n: "filter" is a keyword
         raise error.ParseError(_("filter expects one or two arguments"))
     iterable = evalwrapped(context, mapping, args[0])
     if len(args) == 1:
         def select(w):
             return w.tobool(context, mapping)
     else:
         def select(w):
             if not isinstance(w, templateutil.mappable):
                 raise error.ParseError(_("not filterable by expression"))
             lm = context.overlaymap(mapping, w.tomap(context))
             return evalboolean(context, lm, args[1])
     return iterable.filter(context, mapping, select)
 
 @templatefunc('formatnode(node)', requires={'ui'})
 def formatnode(context, mapping, args):
     """Obtain the preferred form of a changeset hash. (DEPRECATED)"""
     if len(args) != 1:
         # i18n: "formatnode" is a keyword
         raise error.ParseError(_("formatnode expects one argument"))
 
     ui = context.resource(mapping, 'ui')
     node = evalstring(context, mapping, args[0])
     if ui.debugflag:
         return node
     return templatefilters.short(node)
 
 @templatefunc('mailmap(author)', requires={'repo', 'cache'})
 def mailmap(context, mapping, args):
     """Return the author, updated according to the value
     set in the .mailmap file"""
     if len(args) != 1:
         raise error.ParseError(_("mailmap expects one argument"))
 
     author = evalstring(context, mapping, args[0])
 
     cache = context.resource(mapping, 'cache')
     repo = context.resource(mapping, 'repo')
 
     if 'mailmap' not in cache:
         data = repo.wvfs.tryread('.mailmap')
         cache['mailmap'] = stringutil.parsemailmap(data)
 
     return stringutil.mapname(cache['mailmap'], author)
 
 @templatefunc('pad(text, width[, fillchar=\' \'[, left=False]])',
               argspec='text width fillchar left')
 def pad(context, mapping, args):
     """Pad text with a
     fill character."""
     if 'text' not in args or 'width' not in args:
         # i18n: "pad" is a keyword
         raise error.ParseError(_("pad() expects two to four arguments"))
 
     width = evalinteger(context, mapping, args['width'],
                         # i18n: "pad" is a keyword
                         _("pad() expects an integer width"))
 
     text = evalstring(context, mapping, args['text'])
 
     left = False
     fillchar = ' '
     if 'fillchar' in args:
         fillchar = evalstring(context, mapping, args['fillchar'])
         if len(color.stripeffects(fillchar)) != 1:
             # i18n: "pad" is a keyword
             raise error.ParseError(_("pad() expects a single fill character"))
     if 'left' in args:
         left = evalboolean(context, mapping, args['left'])
 
     fillwidth = width - encoding.colwidth(color.stripeffects(text))
     if fillwidth <= 0:
         return text
     if left:
         return fillchar * fillwidth + text
     else:
         return text + fillchar * fillwidth
 
 @templatefunc('indent(text, indentchars[, firstline])')
 def indent(context, mapping, args):
     """Indents all non-empty lines
     with the characters given in the indentchars string. An optional
     third parameter will override the indent for the first line only
     if present."""
     if not (2 <= len(args) <= 3):
         # i18n: "indent" is a keyword
         raise error.ParseError(_("indent() expects two or three arguments"))
 
     text = evalstring(context, mapping, args[0])
     indent = evalstring(context, mapping, args[1])
 
     if len(args) == 3:
         firstline = evalstring(context, mapping, args[2])
     else:
         firstline = indent
 
     # the indent function doesn't indent the first line, so we do it here
     return templatefilters.indent(firstline + text, indent)
 
 @templatefunc('get(dict, key)')
 def get(context, mapping, args):
     """Get an attribute/key from an object. Some keywords
     are complex types. This function allows you to obtain the value of an
     attribute on these types."""
     if len(args) != 2:
         # i18n: "get" is a keyword
         raise error.ParseError(_("get() expects two arguments"))
 
     dictarg = evalwrapped(context, mapping, args[0])
     key = evalrawexp(context, mapping, args[1])
     try:
         return dictarg.getmember(context, mapping, key)
     except error.ParseError as err:
         # i18n: "get" is a keyword
         hint = _("get() expects a dict as first argument")
         raise error.ParseError(bytes(err), hint=hint)
 
 @templatefunc('if(expr, then[, else])')
 def if_(context, mapping, args):
     """Conditionally execute based on the result of
     an expression."""
     if not (2 <= len(args) <= 3):
         # i18n: "if" is a keyword
         raise error.ParseError(_("if expects two or three arguments"))
 
     test = evalboolean(context, mapping, args[0])
     if test:
         return evalrawexp(context, mapping, args[1])
     elif len(args) == 3:
         return evalrawexp(context, mapping, args[2])
 
 @templatefunc('ifcontains(needle, haystack, then[, else])')
 def ifcontains(context, mapping, args):
     """Conditionally execute based
     on whether the item "needle" is in "haystack"."""
     if not (3 <= len(args) <= 4):
         # i18n: "ifcontains" is a keyword
         raise error.ParseError(_("ifcontains expects three or four arguments"))
 
     haystack = evalwrapped(context, mapping, args[1])
     try:
         needle = evalrawexp(context, mapping, args[0])
         found = haystack.contains(context, mapping, needle)
     except error.ParseError:
         found = False
 
     if found:
         return evalrawexp(context, mapping, args[2])
     elif len(args) == 4:
         return evalrawexp(context, mapping, args[3])
 
 @templatefunc('ifeq(expr1, expr2, then[, else])')
 def ifeq(context, mapping, args):
     """Conditionally execute based on
     whether 2 items are equivalent."""
     if not (3 <= len(args) <= 4):
         # i18n: "ifeq" is a keyword
         raise error.ParseError(_("ifeq expects three or four arguments"))
 
     test = evalstring(context, mapping, args[0])
     match = evalstring(context, mapping, args[1])
     if test == match:
         return evalrawexp(context, mapping, args[2])
     elif len(args) == 4:
         return evalrawexp(context, mapping, args[3])
 
 @templatefunc('join(list, sep)')
 def join(context, mapping, args):
     """Join items in a list with a delimiter."""
     if not (1 <= len(args) <= 2):
         # i18n: "join" is a keyword
         raise error.ParseError(_("join expects one or two arguments"))
 
     joinset = evalwrapped(context, mapping, args[0])
     joiner = " "
     if len(args) > 1:
         joiner = evalstring(context, mapping, args[1])
     return joinset.join(context, mapping, joiner)
 
 @templatefunc('label(label, expr)', requires={'ui'})
 def label(context, mapping, args):
     """Apply a label to generated content. Content with
     a label applied can result in additional post-processing, such as
     automatic colorization."""
     if len(args) != 2:
         # i18n: "label" is a keyword
         raise error.ParseError(_("label expects two arguments"))
 
     ui = context.resource(mapping, 'ui')
     thing = evalstring(context, mapping, args[1])
     # preserve unknown symbol as literal so effects like 'red', 'bold',
     # etc. don't need to be quoted
     label = evalstringliteral(context, mapping, args[0])
 
     return ui.label(thing, label)
 
 @templatefunc('latesttag([pattern])')
 def latesttag(context, mapping, args):
     """The global tags matching the given pattern on the
     most recent globally tagged ancestor of this changeset.
     If no such tags exist, the "{tag}" template resolves to
     the string "null". See :hg:`help revisions.patterns` for the pattern
     syntax.
     """
     if len(args) > 1:
         # i18n: "latesttag" is a keyword
         raise error.ParseError(_("latesttag expects at most one argument"))
 
     pattern = None
     if len(args) == 1:
         pattern = evalstring(context, mapping, args[0])
     return templatekw.showlatesttags(context, mapping, pattern)
 
 @templatefunc('localdate(date[, tz])')
 def localdate(context, mapping, args):
     """Converts a date to the specified timezone.
     The default is local date."""
     if not (1 <= len(args) <= 2):
         # i18n: "localdate" is a keyword
         raise error.ParseError(_("localdate expects one or two arguments"))
 
     date = evaldate(context, mapping, args[0],
                     # i18n: "localdate" is a keyword
                     _("localdate expects a date information"))
     if len(args) >= 2:
         tzoffset = None
         tz = evalfuncarg(context, mapping, args[1])
         if isinstance(tz, bytes):
             tzoffset, remainder = dateutil.parsetimezone(tz)
             if remainder:
                 tzoffset = None
         if tzoffset is None:
             try:
                 tzoffset = int(tz)
             except (TypeError, ValueError):
                 # i18n: "localdate" is a keyword
                 raise error.ParseError(_("localdate expects a timezone"))
     else:
         tzoffset = dateutil.makedate()[1]
     return templateutil.date((date[0], tzoffset))
 
 @templatefunc('max(iterable)')
 def max_(context, mapping, args, **kwargs):
     """Return the max of an iterable"""
     if len(args) != 1:
         # i18n: "max" is a keyword
         raise error.ParseError(_("max expects one argument"))
 
     iterable = evalwrapped(context, mapping, args[0])
     try:
         return iterable.getmax(context, mapping)
     except error.ParseError as err:
         # i18n: "max" is a keyword
         hint = _("max first argument should be an iterable")
         raise error.ParseError(bytes(err), hint=hint)
 
 @templatefunc('min(iterable)')
 def min_(context, mapping, args, **kwargs):
     """Return the min of an iterable"""
     if len(args) != 1:
         # i18n: "min" is a keyword
         raise error.ParseError(_("min expects one argument"))
 
     iterable = evalwrapped(context, mapping, args[0])
     try:
         return iterable.getmin(context, mapping)
     except error.ParseError as err:
         # i18n: "min" is a keyword
         hint = _("min first argument should be an iterable")
         raise error.ParseError(bytes(err), hint=hint)
 
 @templatefunc('mod(a, b)')
 def mod(context, mapping, args):
     """Calculate a mod b such that a / b + a mod b == a"""
     if not len(args) == 2:
         # i18n: "mod" is a keyword
         raise error.ParseError(_("mod expects two arguments"))
 
     func = lambda a, b: a % b
     return templateutil.runarithmetic(context, mapping,
                                       (func, args[0], args[1]))
 
 @templatefunc('obsfateoperations(markers)')
 def obsfateoperations(context, mapping, args):
     """Compute obsfate related information based on markers (EXPERIMENTAL)"""
     if len(args) != 1:
         # i18n: "obsfateoperations" is a keyword
         raise error.ParseError(_("obsfateoperations expects one argument"))
 
     markers = evalfuncarg(context, mapping, args[0])
 
     try:
         data = obsutil.markersoperations(markers)
         return templateutil.hybridlist(data, name='operation')
     except (TypeError, KeyError):
         # i18n: "obsfateoperations" is a keyword
         errmsg = _("obsfateoperations first argument should be an iterable")
         raise error.ParseError(errmsg)
 
 @templatefunc('obsfatedate(markers)')
 def obsfatedate(context, mapping, args):
     """Compute obsfate related information based on markers (EXPERIMENTAL)"""
     if len(args) != 1:
         # i18n: "obsfatedate" is a keyword
         raise error.ParseError(_("obsfatedate expects one argument"))
 
     markers = evalfuncarg(context, mapping, args[0])
 
     try:
         # TODO: maybe this has to be a wrapped list of date wrappers?
         data = obsutil.markersdates(markers)
         return templateutil.hybridlist(data, name='date', fmt='%d %d')
     except (TypeError, KeyError):
         # i18n: "obsfatedate" is a keyword
         errmsg = _("obsfatedate first argument should be an iterable")
         raise error.ParseError(errmsg)
 
 @templatefunc('obsfateusers(markers)')
 def obsfateusers(context, mapping, args):
     """Compute obsfate related information based on markers (EXPERIMENTAL)"""
     if len(args) != 1:
         # i18n: "obsfateusers" is a keyword
         raise error.ParseError(_("obsfateusers expects one argument"))
 
     markers = evalfuncarg(context, mapping, args[0])
 
     try:
         data = obsutil.markersusers(markers)
         return templateutil.hybridlist(data, name='user')
     except (TypeError, KeyError, ValueError):
         # i18n: "obsfateusers" is a keyword
         msg = _("obsfateusers first argument should be an iterable of "
                 "obsmakers")
         raise error.ParseError(msg)
 
 @templatefunc('obsfateverb(successors, markers)')
 def obsfateverb(context, mapping, args):
     """Compute obsfate related information based on successors (EXPERIMENTAL)"""
     if len(args) != 2:
         # i18n: "obsfateverb" is a keyword
         raise error.ParseError(_("obsfateverb expects two arguments"))
 
     successors = evalfuncarg(context, mapping, args[0])
     markers = evalfuncarg(context, mapping, args[1])
 
     try:
         return obsutil.obsfateverb(successors, markers)
     except TypeError:
         # i18n: "obsfateverb" is a keyword
         errmsg = _("obsfateverb first argument should be countable")
         raise error.ParseError(errmsg)
 
 @templatefunc('relpath(path)', requires={'repo'})
 def relpath(context, mapping, args):
     """Convert a repository-absolute path into a filesystem path relative to
     the current working directory."""
     if len(args) != 1:
         # i18n: "relpath" is a keyword
         raise error.ParseError(_("relpath expects one argument"))
 
     repo = context.resource(mapping, 'repo')
     path = evalstring(context, mapping, args[0])
     return repo.pathto(path)
 
 @templatefunc('revset(query[, formatargs...])', requires={'repo', 'cache'})
 def revset(context, mapping, args):
     """Execute a revision set query. See
     :hg:`help revset`."""
     if not len(args) > 0:
         # i18n: "revset" is a keyword
         raise error.ParseError(_("revset expects one or more arguments"))
 
     raw = evalstring(context, mapping, args[0])
     repo = context.resource(mapping, 'repo')
 
     def query(expr):
         m = revsetmod.match(repo.ui, expr, lookup=revsetmod.lookupfn(repo))
         return m(repo)
 
     if len(args) > 1:
         formatargs = [evalfuncarg(context, mapping, a) for a in args[1:]]
         revs = query(revsetlang.formatspec(raw, *formatargs))
         revs = list(revs)
     else:
         cache = context.resource(mapping, 'cache')
         revsetcache = cache.setdefault("revsetcache", {})
         if raw in revsetcache:
             revs = revsetcache[raw]
         else:
             revs = query(raw)
             revs = list(revs)
             revsetcache[raw] = revs
     return templatekw.showrevslist(context, mapping, "revision", revs)
 
 @templatefunc('rstdoc(text, style)')
 def rstdoc(context, mapping, args):
     """Format reStructuredText."""
     if len(args) != 2:
         # i18n: "rstdoc" is a keyword
         raise error.ParseError(_("rstdoc expects two arguments"))
 
     text = evalstring(context, mapping, args[0])
     style = evalstring(context, mapping, args[1])
 
-    return minirst.format(text, style=style, keep=['verbose'])[0]
+    return minirst.format(text, style=style, keep=['verbose'])
 
 @templatefunc('separate(sep, args...)', argspec='sep *args')
 def separate(context, mapping, args):
     """Add a separator between non-empty arguments."""
     if 'sep' not in args:
         # i18n: "separate" is a keyword
         raise error.ParseError(_("separate expects at least one argument"))
 
     sep = evalstring(context, mapping, args['sep'])
     first = True
     for arg in args['args']:
         argstr = evalstring(context, mapping, arg)
         if not argstr:
             continue
         if first:
             first = False
         else:
             yield sep
         yield argstr
 
 @templatefunc('shortest(node, minlength=4)', requires={'repo', 'cache'})
 def shortest(context, mapping, args):
     """Obtain the shortest representation of
     a node."""
     if not (1 <= len(args) <= 2):
         # i18n: "shortest" is a keyword
         raise error.ParseError(_("shortest() expects one or two arguments"))
 
     hexnode = evalstring(context, mapping, args[0])
 
     minlength = 4
     if len(args) > 1:
         minlength = evalinteger(context, mapping, args[1],
                                 # i18n: "shortest" is a keyword
                                 _("shortest() expects an integer minlength"))
 
     repo = context.resource(mapping, 'repo')
     if len(hexnode) > 40:
         return hexnode
     elif len(hexnode) == 40:
         try:
             node = bin(hexnode)
         except TypeError:
             return hexnode
     else:
         try:
             node = scmutil.resolvehexnodeidprefix(repo, hexnode)
         except error.WdirUnsupported:
             node = wdirid
         except error.LookupError:
             return hexnode
         if not node:
             return hexnode
     cache = context.resource(mapping, 'cache')
     try:
         return scmutil.shortesthexnodeidprefix(repo, node, minlength, cache)
     except error.RepoLookupError:
         return hexnode
 
 @templatefunc('strip(text[, chars])')
 def strip(context, mapping, args):
     """Strip characters from a string. By default,
     strips all leading and trailing whitespace."""
     if not (1 <= len(args) <= 2):
         # i18n: "strip" is a keyword
         raise error.ParseError(_("strip expects one or two arguments"))
 
     text = evalstring(context, mapping, args[0])
     if len(args) == 2:
         chars = evalstring(context, mapping, args[1])
         return text.strip(chars)
     return text.strip()
 
 @templatefunc('sub(pattern, replacement, expression)')
 def sub(context, mapping, args):
     """Perform text substitution
     using regular expressions."""
     if len(args) != 3:
         # i18n: "sub" is a keyword
         raise error.ParseError(_("sub expects three arguments"))
 
     pat = evalstring(context, mapping, args[0])
     rpl = evalstring(context, mapping, args[1])
     src = evalstring(context, mapping, args[2])
     try:
         patre = re.compile(pat)
     except re.error:
         # i18n: "sub" is a keyword
         raise error.ParseError(_("sub got an invalid pattern: %s") % pat)
     try:
         yield patre.sub(rpl, src)
     except re.error:
         # i18n: "sub" is a keyword
         raise error.ParseError(_("sub got an invalid replacement: %s") % rpl)
 
 @templatefunc('startswith(pattern, text)')
 def startswith(context, mapping, args):
     """Returns the value from the "text" argument
     if it begins with the content from the "pattern" argument."""
     if len(args) != 2:
         # i18n: "startswith" is a keyword
         raise error.ParseError(_("startswith expects two arguments"))
 
     patn = evalstring(context, mapping, args[0])
     text = evalstring(context, mapping, args[1])
     if text.startswith(patn):
         return text
     return ''
 
 @templatefunc('word(number, text[, separator])')
 def word(context, mapping, args):
     """Return the nth word from a string."""
     if not (2 <= len(args) <= 3):
         # i18n: "word" is a keyword
         raise error.ParseError(_("word expects two or three arguments, got %d")
                                % len(args))
 
     num = evalinteger(context, mapping, args[0],
                       # i18n: "word" is a keyword
                       _("word expects an integer index"))
     text = evalstring(context, mapping, args[1])
     if len(args) == 3:
         splitter = evalstring(context, mapping, args[2])
     else:
         splitter = None
 
     tokens = text.split(splitter)
     if num >= len(tokens) or num < -len(tokens):
         return ''
     else:
         return tokens[num]
 
 def loadfunction(ui, extname, registrarobj):
     """Load template function from specified registrarobj
     """
     for name, func in registrarobj._table.iteritems():
         funcs[name] = func
 
 # tell hggettext to extract docstrings from these functions:
 i18nfunctions = funcs.values()

tests/test-minirst.py

@@ -1,268 +1,267 b''
 from __future__ import absolute_import, print_function
 from mercurial import (
     minirst,
 )
 from mercurial.utils import (
     stringutil,
 )
 
 def debugformat(text, form, **kwargs):
+    blocks, pruned = minirst.parse(text, **kwargs)
     if form == b'html':
         print("html format:")
         out = minirst.format(text, style=form, **kwargs)
     else:
         print("%d column format:" % form)
         out = minirst.format(text, width=form, **kwargs)
 
     print("-" * 70)
-    if type(out) == tuple:
-        print(out[0][:-1].decode('utf8'))
+    print(out[:-1].decode('utf8'))
+    if kwargs.get('keep'):
         print("-" * 70)
-        print(stringutil.pprint(out[1]).decode('utf8'))
-    else:
-        print(out[:-1].decode('utf8'))
+        print(stringutil.pprint(pruned).decode('utf8'))
     print("-" * 70)
     print()
 
 def debugformats(title, text, **kwargs):
     print("== %s ==" % title)
     debugformat(text, 60, **kwargs)
     debugformat(text, 30, **kwargs)
     debugformat(text, b'html', **kwargs)
 
 paragraphs = b"""
 This is some text in the first paragraph.
 
   A small indented paragraph.
   It is followed by some lines
   containing random whitespace.
  \n  \n   \nThe third and final paragraph.
 """
 
 debugformats('paragraphs', paragraphs)
 
 definitions = b"""
 A Term
   Definition. The indented
   lines make up the definition.
 Another Term
   Another definition. The final line in the
    definition determines the indentation, so
     this will be indented with four spaces.
 
   A Nested/Indented Term
     Definition.
 """
 
 debugformats('definitions', definitions)
 
 literals = br"""
 The fully minimized form is the most
 convenient form::
 
   Hello
     literal
       world
 
 In the partially minimized form a paragraph
 simply ends with space-double-colon. ::
 
   ////////////////////////////////////////
   long un-wrapped line in a literal block
   \\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\
 
 ::
 
   This literal block is started with '::',
     the so-called expanded form. The paragraph
       with '::' disappears in the final output.
 """
 
 debugformats('literals', literals)
 
 lists = b"""
 - This is the first list item.
 
   Second paragraph in the first list item.
 
 - List items need not be separated
   by a blank line.
 - And will be rendered without
   one in any case.
 
 We can have indented lists:
 
   - This is an indented list item
 
   - Another indented list item::
 
       - A literal block in the middle
             of an indented list.
 
       (The above is not a list item since we are in the literal block.)
 
 ::
 
   Literal block with no indentation (apart from
   the two spaces added to all literal blocks).
 
 1. This is an enumerated list (first item).
 2. Continuing with the second item.
 
 (1) foo
 (2) bar
 
 1) Another
 2) List
 
 Line blocks are also a form of list:
 
 | This is the first line.
   The line continues here.
 | This is the second line.
 
 Bullet lists are also detected:
 
 * This is the first bullet
 * This is the second bullet
   It has 2 lines
 * This is the third bullet
 """
 
 debugformats('lists', lists)
 
 options = b"""
 There is support for simple option lists,
 but only with long options:
 
 -X, --exclude  filter  an option with a short and long option with an argument
 -I, --include          an option with both a short option and a long option
 --all                  Output all.
 --both                 Output both (this description is
                        quite long).
 --long                 Output all day long.
 
 --par                 This option has two paragraphs in its description.
                       This is the first.
 
                       This is the second.  Blank lines may be omitted between
                       options (as above) or left in (as here).
 
 
 The next paragraph looks like an option list, but lacks the two-space
 marker after the option. It is treated as a normal paragraph:
 
 --foo bar baz
 """
 
 debugformats('options', options)
 
 fields = b"""
 :a: First item.
 :ab: Second item. Indentation and wrapping
      is handled automatically.
 
 Next list:
 
 :small: The larger key below triggers full indentation here.
 :much too large: This key is big enough to get its own line.
 """
 
 debugformats('fields', fields)
 
 containers = b"""
 Normal output.
 
 .. container:: debug
 
    Initial debug output.
 
 .. container:: verbose
 
    Verbose output.
 
    .. container:: debug
 
       Debug output.
 """
 
 debugformats('containers (normal)', containers)
 debugformats('containers (verbose)', containers, keep=[b'verbose'])
 debugformats('containers (debug)', containers, keep=[b'debug'])
 debugformats('containers (verbose debug)', containers,
             keep=[b'verbose', b'debug'])
 
 roles = b"""Please see :hg:`add`."""
 debugformats('roles', roles)
 
 
 sections = b"""
 Title
 =====
 
 Section
 -------
 
 Subsection
 ''''''''''
 
 Markup: ``foo`` and :hg:`help`
 ------------------------------
 """
 debugformats('sections', sections)
 
 
 admonitions = b"""
 .. note::
 
    This is a note
 
    - Bullet 1
    - Bullet 2
 
    .. warning:: This is a warning Second
       input line of warning
 
 .. danger::
    This is danger
 """
 
 debugformats('admonitions', admonitions)
 
 comments = b"""
 Some text.
 
 .. A comment
 
    .. An indented comment
 
    Some indented text.
 
 ..
 
 Empty comment above
 """
 
 debugformats('comments', comments)
 
 
 data = [[b'a', b'b', b'c'],
          [b'1', b'2', b'3'],
          [b'foo', b'bar', b'baz this list is very very very long man']]
 
 rst = minirst.maketable(data, 2, True)
 table = b''.join(rst)
 
 print(table.decode('utf8'))
 
 debugformats('table', table)
 
 data = [[b's', b'long', b'line\ngoes on here'],
         [b'', b'xy', b'tried to fix here\n        by indenting']]
 
 rst = minirst.maketable(data, 1, False)
 table = b''.join(rst)
 
 print(table.decode('utf8'))
 
 debugformats('table+nl', table)