#! /usr/local/bin/python
# Convert GNU texinfo files into HTML, one file per node.
# Based on Texinfo 2.14.
# Usage: texi2html [-d] [-d] inputfile outputdirectory
# The input file must be a complete texinfo file, e.g. emacs.texi.
# This creates many files (one per info node) in the output directory,
# overwriting existing files of the same name. All files created have
# ".html" as their extension.
# XXX To do:
# - handle @comment*** correctly
# - handle @xref {some words} correctly
# - handle @ftable correctly (items aren't indexed?)
# - handle @itemx properly
# - handle @exdent properly
# - add links directly to the proper line from indices
# - check against the definitive list of @-cmds; we still miss (among others):
# - @set, @clear, @ifset, @ifclear
# - @defindex (hard)
# - @c(omment) in the middle of a line (rarely used)
# - @this* (not really needed, only used in headers anyway)
# - @today{} (ever used outside title page?)
import os
import regex
import regsub
import string
MAGIC = '\\input texinfo'
cmprog = regex.compile('^@\([a-z]+\)\([ \t]\|$\)') # Command (line-oriented)
blprog = regex.compile('^[ \t]*$') # Blank line
kwprog = regex.compile('@[a-z]+') # Keyword (embedded, usually with {} args)
spprog = regex.compile('[\n@{}&<>]') # Special characters in running text
miprog = regex.compile( \
'^\* \([^:]*\):\(:\|[ \t]*\([^\t,\n.]+\)\([^ \t\n]*\)\)[ \t\n]*')
# menu item (Yuck!)
class TexinfoParser:
# Initialize an instance
def __init__(self):
self.unknown = {} # statistics about unknown @-commands
self.filenames = {} # Check for identical filenames
self.debugging = 0 # larger values produce more output
self.nodefp = None # open file we're writing to
self.nodelineno = 0 # Linenumber relative to node
self.links = None # Links from current node
self.savetext = None # If not None, save text head instead
self.dirname = 'tmp' # directory where files are created
self.includedir = '.' # directory to search @include files
self.nodename = '' # name of current node
self.topname = '' # name of top node (first node seen)
self.title = '' # title of this whole Texinfo tree
self.resetindex() # Reset all indices
self.contents = [] # Reset table of contents
self.numbering = [] # Reset section numbering counters
self.nofill = 0 # Normal operation: fill paragraphs
self.goodset=['html'] # Names that should be parsed in ifset
self.stackinfo={} # Keep track of state in the stack
# XXX The following should be reset per node?!
self.footnotes = [] # Reset list of footnotes
self.itemarg = None # Reset command used by @item
self.itemnumber = None # Reset number for @item in @enumerate
self.itemindex = None # Reset item index name
# Set (output) directory name
def setdirname(self, dirname):
self.dirname = dirname
# Set include directory name
def setincludedir(self, includedir):
self.includedir = includedir
# Parse the contents of an entire file
def parse(self, fp):
line = fp.readline()
lineno = 1
while line and (line[0] == '%' or blprog.match(line) >= 0):
line = fp.readline()
lineno = lineno + 1
if line[:len(MAGIC)] <> MAGIC:
raise SyntaxError, 'file does not begin with '+`MAGIC`
self.parserest(fp, lineno)
# Parse the contents of a file, not expecting a MAGIC header
def parserest(self, fp, initial_lineno):
lineno = initial_lineno
self.done = 0
self.skip = 0
self.stack = []
accu = []
while not self.done:
line = fp.readline()
self.nodelineno = self.nodelineno + 1
if not line:
if accu:
if not self.skip: self.process(accu)
accu = []
if initial_lineno > 0:
print '*** EOF before @bye'
break
lineno = lineno + 1
if cmprog.match(line) >= 0:
a, b = cmprog.regs[1]
cmd = line[a:b]
if cmd in ('noindent', 'refill'):
accu.append(line)
else:
if accu:
if not self.skip:
self.process(accu)
accu = []
self.command(line)
elif blprog.match(line) >= 0 and \
'format' not in self.stack and \
'example' not in self.stack:
if accu:
if not self.skip:
self.process(accu)
self.write('\n')
accu = []
else:
# Append the line including trailing \n!
accu.append(line)
#
if self.skip:
print '*** Still skipping at the end'
if self.stack:
print '*** Stack not empty at the end'
print '***', self.stack
# Start saving text in a buffer instead of writing it to a file
def startsaving(self):
if self.savetext <> None:
print '*** Recursively saving text, expect trouble'
self.savetext = ''
# Return the text saved so far and start writing to file again
def collectsavings(self):
savetext = self.savetext
self.savetext = None
return savetext or ''
# Write text to file, or save it in a buffer, or ignore it
def write(self, *args):
text = string.joinfields(args, '')
if self.savetext <> None:
self.savetext = self.savetext + text
elif self.nodefp:
self.nodefp.write(text)
# Complete the current node -- write footnotes and close file
def endnode(self):
if self.savetext <> None:
print '*** Still saving text at end of node'
dummy = self.collectsavings()
if self.footnotes:
self.writefootnotes()
if self.nodefp:
if self.nodelineno > 20:
self.write ('
\n')
[name, next, prev, up] = self.nodelinks[:4]
self.link('Next', next)
self.link('Prev', prev)
self.link('Up', up)
if self.nodename <> self.topname:
self.link('Top', self.topname)
self.write ('
\n')
self.write(' \n')
self.nodefp.close()
self.nodefp = None
self.nodename = ''
# Process a list of lines, expanding embedded @-commands
# This mostly distinguishes between menus and normal text
def process(self, accu):
if self.debugging > 1:
print self.skip, self.stack,
if accu: print accu[0][:30],
if accu[0][30:] or accu[1:]: print '...',
print
if self.stack and self.stack[-1] == 'menu':
# XXX should be done differently
for line in accu:
if miprog.match(line) < 0:
line = string.strip(line) + '\n'
self.expand(line)
continue
(bgn, end), (a, b), (c, d), (e, f), (g, h) = \
miprog.regs[:5]
label = line[a:b]
nodename = line[c:d]
if nodename[0] == ':': nodename = label
else: nodename = line[e:f]
punct = line[g:h]
self.write('