Do a little bit more work on a table-driven basis. A little cleanup.
Seems to work at the moment.
diff --git a/Doc/tools/partparse.py b/Doc/tools/partparse.py
index 36ffb6c..d3b850f 100644
--- a/Doc/tools/partparse.py
+++ b/Doc/tools/partparse.py
@@ -836,9 +836,10 @@
self.data = self.data + data
# ignore these commands
-ignoredcommands = ('bcode', 'ecode', 'hline', 'small', '/')
+ignoredcommands = ('hline', 'small', '/', 'tableofcontents', 'Large')
# map commands like these to themselves as plaintext
-wordsselves = ('UNIX', 'ABC', 'C', 'ASCII', 'EOF', 'LaTeX', 'POSIX')
+wordsselves = ('UNIX', 'ABC', 'C', 'ASCII', 'EOF', 'LaTeX', 'POSIX', 'TeX',
+ 'SliTeX')
# \{ --> {, \} --> }, etc
themselves = ('{', '}', ',', '.', '@', ' ', '\n') + wordsselves
# these ones also themselves (see argargs macro in myformat.sty)
@@ -851,10 +852,6 @@
# recognise patter {\FONTCHANGE-CMD TEXT} to \MAPPED-FC-CMD{TEXT}
fontchanges = {'rm': 'r', 'it': 'i', 'em': 'emph', 'bf': 'b', 'tt': 't'}
-# transparent for these commands
-for_texi = ('emph', 'var', 'strong', 'code', 'kbd', 'key', 'dfn', 'samp',
- 'file', 'r', 'i', 't')
-
# try to remove macros and return flat text
def flattext(buf, pp):
@@ -874,11 +871,10 @@
pass
elif ch.chtype == chunk_type[CSNAME]:
s_buf_data = s(buf, ch.data)
- if s_buf_data in themselves or hist.inargs and s_buf_data in inargsselves:
+ if convertible_csname(s_buf_data):
+ ch.chtype, ch.data, nix = conversion(s_buf_data)
+ if hist.inargs and s_buf_data in inargsselves:
ch.chtype = chunk_type[PLAIN]
- elif s_buf_data == 'e':
- ch.chtype = chunk_type[PLAIN]
- ch.data = '\\'
elif len(s_buf_data) == 1 \
and s_buf_data in onlylatexspecial:
ch.chtype = chunk_type[PLAIN]
@@ -993,6 +989,7 @@
def startchange():
global hist, out
+ hist.chaptertype = "chapter"
hist.inenv = []
hist.nodenames = []
hist.cindex = []
@@ -1013,17 +1010,42 @@
# same for enumerate
enumeratesymbols = ['1', 'A', 'a']
+# Map of things that convert one-to-one. Each entry is a 3-tuple:
+#
+# new_chtype, new_data, nix_trailing_empty_group
+#
d = {}
+# add stuff that converts from one name to another:
for name in ('url', 'module', 'function', 'cfunction',
'keyword', 'method', 'exception', 'constant',
'email', 'class', 'member', 'cdata', 'ctype',
- 'member'):
- d[name] = 'code'
-d['program'] = 'strong'
-d['sectcode'] = 'code'
+ 'member', 'sectcode', 'verb'):
+ d[name] = chunk_type[CSNAME], 'code', 0
+for name in ('emph', 'var', 'strong', 'code', 'kbd', 'key',
+ 'dfn', 'samp', 'file', 'r', 'i', 't'):
+ d[name] = chunk_type[CSNAME], name, 0
+d['program'] = chunk_type[CSNAME], 'strong', 0
+d['\\'] = chunk_type[CSNAME], '*', 0
+# add stuff that converts to text:
+for name in themselves:
+ d[name] = chunk_type[PLAIN], name, 0
+for name in wordsselves:
+ d[name] = chunk_type[PLAIN], name, 1
+for name in ',[]()':
+ d[name] = chunk_type[PLAIN], name, 0
+# a lot of these are LaTeX2e additions
+for name, value in [('quotedblbase', ',,'), ('quotesinglbase', ','),
+ ('textquotedbl', '"'), ('LaTeXe', 'LaTeX2e'),
+ ('e', '\\'), ('textquotedblleft', "``"),
+ ('textquotedblright', "''"), ('textquoteleft', "`"),
+ ('textquoteright', "'"), ('textbackslash', '\\'),
+ ('textbar', '|'), ('textless', '<'),
+ ('textgreater', '>'), ('textasciicircum', '^'),
+ ('Cpp', 'C++'), ('copyright', '')]:
+ d[name] = chunk_type[PLAIN], value, 1
convertible_csname = d.has_key
conversion = d.get
-del d, name
+del d, name, value
##
## \begin{ {func,data,exc}desc }{name}...
@@ -1261,6 +1283,7 @@
while 1:
# sanity check: length should always equal len(pp)
if len(pp) != length:
+ print i, pp[i]
raise 'FATAL', 'inconsistent length. thought ' + `length` + ', but should really be ' + `len(pp)`
if i >= length:
break
@@ -1404,7 +1427,7 @@
raise error, 'Sorry, expected plain text argument'
hist.itemargmacro = s(buf, pp[i].data)
if convertible_csname(hist.itemargmacro):
- hist.itemargmacro = conversion(hist.itemargmacro)
+ hist.itemargmacro = conversion(hist.itemargmacro)[1]
del pp[i:newi]
length = length - (newi-i)
@@ -1466,7 +1489,7 @@
length = length + len(chunks) - 1
i = i + len(chunks) - 1
- elif envname in ('sloppypar', 'flushleft'):
+ elif envname in ('sloppypar', 'flushleft', 'document'):
pass
else:
@@ -1523,7 +1546,7 @@
chunk(PLAIN, ch.where, "deffn")])]
i, length = i+2, length+2
- elif envname in ('seealso', 'sloppypar', 'flushleft'):
+ elif envname in ('seealso', 'sloppypar', 'flushleft', 'document'):
pass
else:
@@ -1541,9 +1564,11 @@
pp[i:i+1]=cp + [
chunk(PLAIN, ch.where, ']')]
length = length+len(cp)
+
elif s_buf_data in ignoredcommands:
del pp[i-1]
i, length = i-1, length-1
+
elif s_buf_data == '@' and \
i != length and \
pp[i].chtype == chunk_type[PLAIN] and \
@@ -1551,10 +1576,22 @@
# \@. --> \. --> @.
ch.data = '.'
del pp[i]
- length = length-1
+ length = length - 1
+
+ elif convertible_csname(s_buf_data):
+ ch.chtype, ch.data, nix = conversion(s_buf_data)
+ try:
+ if nix and pp[i].chtype == chunk_type[GROUP] \
+ and len(pp[i].data) == 0:
+ del pp[i]
+ length = length - 1
+ except IndexError:
+ pass
+
elif s_buf_data == '\\':
# \\ --> \* --> @*
ch.data = '*'
+
elif len(s_buf_data) == 1 and \
s_buf_data in onlylatexspecial:
ch.chtype = chunk_type[PLAIN]
@@ -1567,6 +1604,11 @@
del pp[i]
length = length-1
+ elif s_buf_data == "appendix":
+ hist.chaptertype = "appendix"
+ del pp[i-1]
+ i, length = i-1, length-1
+
elif hist.inargs and s_buf_data in inargsselves:
# This is the special processing of the
# arguments of the \begin{funcdesc}... or
@@ -1597,6 +1639,13 @@
i = i - 1
length = length - 3
+ elif s_buf_data == 'renewcommand':
+ print "ignoring redefinition of \\" \
+ + s(buf, pp[i].data[0].data)
+ del pp[i-1:i+2]
+ i = i - 1
+ length = length - 3
+
elif s_buf_data == 'mbox':
stuff = pp[i].data
pp[i-1:i+1] = stuff
@@ -1672,6 +1721,7 @@
or pp[i].data != []:
pp.insert(i, chunk(GROUP, ch.where, []))
i, length = i+1, length+1
+
elif s_buf_data in themselves:
# \UNIX --> &UNIX;
ch.chtype = chunk_type[PLAIN]
@@ -1680,8 +1730,14 @@
and pp[i].data == []:
del pp[i]
length = length-1
- elif s_buf_data in for_texi:
- pass
+
+## elif s_buf_data == 'copyright':
+## if (pp[i].chtype == chunk_type[GROUP]
+## and not pp[i].data):
+## del pp[i]
+## length = length - 1
+## del pp[i-1]
+## i, length = i-1, length-1
elif s_buf_data == 'manpage':
ch.data = 'emph'
@@ -1689,10 +1745,6 @@
pp[i+1].data = "(%s)" % sect
pp[i+1].chtype = chunk_type[PLAIN]
- elif s_buf_data == 'e':
- # "\e" --> "\"
- ch.data = '\\'
- ch.chtype = chunk_type[PLAIN]
elif s_buf_data in ('lineiii', 'lineii'):
# This is the most tricky one
# \lineiii{a1}{a2}[{a3}] -->
@@ -1726,11 +1778,14 @@
if length != len(pp):
raise 'IN LINEIII IS THE ERR', `i`
- elif s_buf_data in ('chapter', 'section', 'subsection', 'subsubsection'):
+ elif s_buf_data in ('chapter', 'section',
+ 'subsection', 'subsubsection'):
#\xxxsection{A} ---->
# @node A, , ,
# @xxxsection A
## also: remove commas and quotes
+ if s_buf_data == "chapter":
+ ch.data = hist.chaptertype
ch.chtype = chunk_type[CSLINE]
length, newi = getnextarg(length, buf, pp, i)
afternodenamecmd = next_command_p(length, buf,
@@ -1915,7 +1970,7 @@
pp.insert(i, chunk(GROUP, ch.where, ingroupch))
length, i = length+1, i+1
- elif s_buf_data in ('stindex', 'kwindex'):
+ elif s_buf_data == 'stindex':
# XXX must actually go to newindex st
what = (s_buf_data[:2] == "st") and "statement" or "keyword"
wh = ch.where
@@ -2091,12 +2146,13 @@
ch.chtype = PLAIN
ch.data = " "
+ elif s_buf_data in ('usepackage', 'input'):
+ del pp[i-1:i+1]
+ i, length = i-1, length-2
+
elif s_buf_data in ('noindent', 'indexsubitem', 'footnote'):
pass
- elif convertible_csname(s_buf_data):
- ch.data = conversion(s_buf_data)
-
elif s_buf_data == 'label':
name = s(buf, pp[i].data[0].data)
del pp[i-1:i+1]
@@ -2110,11 +2166,6 @@
del pp[i]
length = length - 1
- elif s_buf_data == 'Large':
- del pp[i-1]
- i = i - 1
- length = length - 1
-
elif s_buf_data == 'ref':
name = s(buf, pp[i].data[0].data)
if label_nodes.has_key(name):
@@ -2154,12 +2205,12 @@
if ch.chtype == chunk_type[CSNAME]:
s_buf_data = s(buf, ch.data)
- if s_buf_data == 'e':
- wm('\\')
- continue
- if s_buf_data == '$':
- wm('$')
- continue
+## if s_buf_data == 'e':
+## wm('\\')
+## continue
+## if s_buf_data == '$':
+## wm('$')
+## continue
wm('@' + s_buf_data)
if s_buf_data == 'node' and \
pp[i].chtype == chunk_type[PLAIN] and \
@@ -2232,13 +2283,11 @@
wm('\n')
elif ch.chtype == chunk_type[COMMENT]:
-## print 'COMMENT: previous chunk =', pp[i-2]
-## if pp[i-2].chtype == chunk_type[PLAIN]:
-## print 'PLAINTEXT =', `s(buf, pp[i-2].data)`
if s(buf, ch.data) and \
regex.match('^[ \t]*$', s(buf, ch.data)) < 0:
if i >= 2 \
- and pp[i-2].chtype not in (chunk_type[ENDLINE], chunk_type[DENDLINE]) \
+ and pp[i-2].chtype not in (chunk_type[ENDLINE],
+ chunk_type[DENDLINE]) \
and not (pp[i-2].chtype == chunk_type[PLAIN]
and regex.match('\\(.\\|\n\\)*[ \t]*\n$', s(buf, pp[i-2].data)) >= 0):
wm('\n')