diff options
Diffstat (limited to 'buildtools/makeman')
| -rwxr-xr-x | buildtools/makeman | 20 |
1 files changed, 11 insertions, 9 deletions
diff --git a/buildtools/makeman b/buildtools/makeman index 2e122779..94ee2172 100755 --- a/buildtools/makeman +++ b/buildtools/makeman @@ -1,4 +1,4 @@ -#!/bin/env python +#!/usr/bin/python # # makeman -- compile netpbm's stereotyped HTML to troff markup # @@ -51,9 +51,7 @@ def makeman(name, file, indoc): # Protect escapes before we try generating font changes. indoc = indoc.replace("\\", r"\e") # Header-bashing - indoc = indoc.replace('<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.1//EN" "DTD/xhtml11.dtd">', "") - indoc = indoc.replace('<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">', "") - indoc = indoc.replace('<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 3.2//EN">\n',"") + indoc = re.sub('(?i)<!DOCTYPE html[^>]*>', "", indoc) indoc = indoc.replace('<meta http-equiv="Content-Type" content="text/html; charset=iso-8859-1">', "") indoc = indoc.replace('<meta http-equiv="Content-Type" content="text/html; charset=us-ascii"/>', "") indoc = indoc.replace('<?xml version="1.1" encoding="iso-8859-1" ?>\n',"") @@ -121,7 +119,10 @@ def makeman(name, file, indoc): indoc = re.sub("(?i)</SUP>", r"\\d", indoc) # Paragraph handling indoc = re.sub("(?i)\n*<P>\n*", r"\n.PP\n", indoc) + indoc = re.sub("(?i)<br */>", r"\n.PP\n", indoc) indoc = re.sub("(?i)</P>", "", indoc) + indoc = re.sub("(?i)<!--[^>]*-->", "", indoc) + indoc = re.sub("(?i)<meta[^>]*>", "", indoc) lines = indoc.split("\n") listdepth = 0 for i in range(len(lines)): @@ -138,13 +139,14 @@ def makeman(name, file, indoc): indoc = re.sub('(?i)<A[ \n]+HREF="mailto:[^>]+">([^<]+)</A>', r'\\fI\1\\fP', indoc) # Format manual crossreferences def xrefmatch(match): - xrefto = match.group(1) - xrefsection = sectmap.get(xrefto, 1) + xrefto = match.group(2) + xrefurl = match.group(1) + xrefsection = sectmap.get(xrefurl, 1) if xrefsection == 0: return "\n.I " + xrefto else: - return "\n.BR %s (%d)" % (xrefto, xrefsection) - indoc = re.sub(r'(?i)\n* *(?:\\fB)?<A[ \n]+HREF="[^>]+.html">([^<]+)</A>(?:\\fP)?', + return "\n.BR %s (%d)\n" % (xrefto, xrefsection) + indoc = re.sub(r'(?i)\n* *(?:\\fB)?<A[ \n]+HREF="?([^>]+.html)"?>([^<]+)</A>(?:\\fP)?', xrefmatch, indoc) # Format URLs def urlmatch(match): @@ -159,7 +161,7 @@ def makeman(name, file, indoc): indoc = indoc.replace("®", r"\*R") indoc = indoc.replace("©", r"\(co") # Turn anchors into .UN tags - indoc = re.sub('(?i)<A NAME *= *"#?([a-zA-Z][a-zA-Z0-9.-]+)">(?: )*</A>\s*', ".UN \\1\n", indoc) + indoc = re.sub('(?i)<A NAME *= *"#?([a-zA-Z_][a-zA-Z_0-9.-]+)">(?: )*</A>\s*', ".UN \\1\n", indoc) # Strip off the index trailer trailer = re.compile('<HR */*>.*', re.DOTALL | re.IGNORECASE) indoc = re.sub(trailer, "", indoc) |