#! /usr/bin/env python """Perform massive transformations on a document tree created from the LaTeXof the Python documentation, and dump the ESIS data for the transformed tree.""" import errnoimport esistoolsimport reimport sysimport xml.domimport xml.dom.minidom ELEMENT = xml.dom.Node.ELEMENT_NODEENTITY_REFERENCE = xml.dom.Node.ENTITY_REFERENCE_NODETEXT = xml.dom.Node.TEXT_NODE class ConversionError(Exception): pass ewrite = sys.stderr.writetry: # We can only do this trick on Unix (if tput is on $PATH)! if sys.platform != "posix" or not sys.stderr.isatty(): raise ImportError import commandsexcept ImportError: bwrite = ewriteelse: def bwrite(s, BOLDON=commands.getoutput("tput bold"), BOLDOFF=commands.getoutput("tput sgr0")): ewrite("%s%s%s" % (BOLDON, s, BOLDOFF)) PARA_ELEMENT = "para" DEBUG_PARA_FIXER = 0 if DEBUG_PARA_FIXER: def para_msg(s): ewrite("*** %s\n" % s)else: def para_msg(s): pass def get_first_element(doc, gi): for n in doc.childNodes: if n.nodeName == gi: return n def extract_first_element(doc, gi): node = get_first_element(doc, gi) if node is not None: doc.removeChild(node) return node def get_documentElement(node): result = None for child in node.childNodes: if child.nodeType == ELEMENT: result = child return result def set_tagName(elem, gi): elem.nodeName = elem.tagName = gi def find_all_elements(doc, gi): nodes = [] if doc.nodeName == gi: nodes.append(doc) for child in doc.childNodes: if child.nodeType == ELEMENT: if child.tagName == gi: nodes.append(child) for node in child.getElementsByTagName(gi): nodes.append(node) return nodes def find_all_child_elements(doc, gi): nodes = [] for child in doc.childNodes: if child.nodeName == gi: nodes.append(child) return nodes def find_all_elements_from_set(doc, gi_set): return __find_all_elements_from_set(doc, gi_set, []) def __find_all_elements_from_set(doc, gi_set, nodes): if doc.nodeName in gi_set: nodes.append(doc) for child in doc.childNodes: if child.nodeType == ELEMENT: __find_all_elements_from_set(child, gi_set, nodes) return nodes def simplify(doc, fragment): # Try to rationalize the document a bit, since these things are simply # not valid SGML/XML documents as they stand, and need a little work. documentclass = "document" inputs = [] node = extract_first_element(fragment, "documentclass") if node is not None: documentclass = node.getAttribute("classname") node = extract_first_element(fragment, "title") if node is not None: inputs.append(node) # update the name of the root element node = get_first_element(fragment, "document") if node is not None: set_tagName(node, documentclass) # Move everything that comes before this node into this node; # this will be the document element. nodelist = fragment.childNodes point = node.firstChild while not nodelist[0].isSameNode(node): node.insertBefore(nodelist[0], point) while 1: node = extract_first_element(fragment, "input") if node is None: break inputs.append(node) if inputs: docelem = get_documentElement(fragment) inputs.reverse() for node in inputs: text = doc.createTextNode("\n") docelem.insertBefore(text, docelem.firstChild) docelem.insertBefore(node, text) docelem.insertBefore(doc.createTextNode("\n"), docelem.firstChild) while fragment.firstChild and fragment.firstChild.nodeType == TEXT: fragment.removeChild(fragment.firstChild) def cleanup_root_text(doc): discards = [] skip = 0 for n in doc.childNodes: prevskip = skip skip = 0 if n.nodeType == TEXT and not prevskip: discards.append(n) elif n.nodeName == "COMMENT": skip = 1 for node in discards: doc.removeChild(node) DESCRIPTOR_ELEMENTS = ( "cfuncdesc", "cvardesc", "ctypedesc", "classdesc", "memberdesc", "memberdescni", "methoddesc", "methoddescni", "excdesc", "funcdesc", "funcdescni", "opcodedesc", "datadesc", "datadescni", ) def fixup_descriptors(doc, fragment): sections = find_all_elements(fragment, "section") for section in sections: find_and_fix_descriptors(doc, section) def find_and_fix_descriptors(doc, container): children = container.childNodes for child in children: if child.nodeType == ELEMENT: tagName = child.tagName if tagName in DESCRIPTOR_ELEMENTS: rewrite_descriptor(doc, child) elif tagName == "subsection": find_and_fix_descriptors(doc, child) def rewrite_descriptor(doc, descriptor): # # Do these things: # 1. Add an "index='no'" attribute to the element if the tagName # ends in 'ni', removing the 'ni' from the name. # 2. Create a <signature> from the name attribute # 2a.Create an <args> if it appears to be available. # 3. Create additional <signature>s from <*line{,ni}> elements, # if found. # 4. If a <versionadded> is found, move it to an attribute on the # descriptor. # 5. Move remaining child nodes to a <description> element. # 6. Put it back together. # # 1. descname = descriptor.tagName index = descriptor.getAttribute("name") != "no" desctype = descname[:-4] # remove 'desc' linename = desctype + "line" if not index: linename = linename + "ni" # 2. signature = doc.createElement("signature") name = doc.createElement("name") signature.appendChild(doc.createTextNode("\n ")) signature.appendChild(name) name.appendChild(doc.createTextNode(descriptor.getAttribute("name"))) descriptor.removeAttribute("name") # 2a. if descriptor.hasAttribute("var"): if descname != "opcodedesc": raise RuntimeError, \ "got 'var' attribute on descriptor other than opcodedesc" variable = descriptor.getAttribute("var") if variable: args = doc.createElement("args") args.appendChild(doc.createTextNode(variable)) signature.appendChild(doc.createTextNode("\n ")) signature.appendChild(args) descriptor.removeAttribute("var") newchildren = [signature] children = descriptor.childNodes pos = skip_leading_nodes(children) if pos < len(children): child = children[pos] if child.nodeName == "args": # move <args> to <signature>, or remove if empty: child.parentNode.removeChild(child) if len(child.childNodes): signature.appendChild(doc.createTextNode("\n ")) signature.appendChild(child) signature.appendChild(doc.createTextNode("\n ")) # 3, 4. pos = skip_leading_nodes(children, pos) while pos < len(children) \ and children[pos].nodeName in (linename, "versionadded"): if children[pos].tagName == linename: # this is really a supplemental signature, create <signature> oldchild = children[pos].cloneNode(1) try: sig = methodline_to_signature(doc, children[pos]) except KeyError: print oldchild.toxml() raise newchildren.append(sig) else: # <versionadded added=...> descriptor.setAttribute( "added", children[pos].getAttribute("version")) pos = skip_leading_nodes(children, pos + 1) # 5. description = doc.createElement("description") description.appendChild(doc.createTextNode("\n")) newchildren.append(description) move_children(descriptor, description, pos) last = description.childNodes[-1] if last.nodeType == TEXT: last.data = last.data.rstrip() + "\n " # 6. # should have nothing but whitespace and signature lines in <descriptor>; # discard them while descriptor.childNodes: descriptor.removeChild(descriptor.childNodes[0]) for node in newchildren: descriptor.appendChild(doc.createTextNode("\n ")) descriptor.appendChild(node) descriptor.appendChild(doc.createTextNode("\n")) def methodline_to_signature(doc, methodline): signature = doc.createElement("signature") signature.appendChild(doc.createTextNode("\n ")) name = doc.createElement("name") name.appendChild(doc.createTextNode(methodline.getAttribute("name"))) methodline.removeAttribute("name") signature.appendChild(name) if len(methodline.childNodes): args = doc.createElement("args") signature.appendChild(doc.createTextNode("\n ")) signature.appendChild(args) move_children(methodline, args) signature.appendChild(doc.createTextNode("\n ")) return signature def move_children(origin, dest, start=0): children = origin.childNodes while start < len(children): node = children[start] origin.removeChild(node) dest.appendChild(node) def handle_appendix(doc, fragment): # must be called after simplfy() if document is multi-rooted to begin with docelem = get_documentElement(fragment) toplevel = docelem.tagName == "manual" and "chapter" or "section" appendices = 0 nodes = [] for node in docelem.childNodes: if appendices: nodes.append(node) elif node.nodeType == ELEMENT: appnodes = node.getElementsByTagName("appendix") if appnodes: appendices = 1 parent = appnodes[0].parentNode parent.removeChild(appnodes[0]) parent.normalize() if nodes: map(docelem.removeChild, nodes) docelem.appendChild(doc.createTextNode("\n\n\n")) back = doc.createElement("back-matter") docelem.appendChild(back) back.appendChild(doc.createTextNode("\n")) while nodes and nodes[0].nodeType == TEXT \ and not nodes[0].data.strip(): del nodes[0] map(back.appendChild, nodes) docelem.appendChild(doc.createTextNode("\n")) def handle_labels(doc, fragment): for label in find_all_elements(fragment, "label"): id = label.getAttribute("id") if not id: continue parent = label.parentNode parentTagName = parent.tagName if parentTagName == "title": parent.parentNode.setAttribute("id", id) else: parent.setAttribute("id", id) # now, remove <label id="..."/> from parent: parent.removeChild(label) if parentTagName == "title": parent.normalize() children = parent.childNodes if children[-1].nodeType == TEXT: children[-1].data = children[-1].data.rstrip() def fixup_trailing_whitespace(doc, fragment, wsmap): queue = [fragment] fixups = [] while queue: node = queue[0] del queue[0] if wsmap.has_key(node.nodeName): fixups.append(node) for child in node.childNodes: if child.nodeType == ELEMENT: queue.append(child) # reverse the list to process from the inside out fixups.reverse() for node in fixups: node.parentNode.normalize() lastchild = node.lastChild before, after = wsmap[node.tagName] if lastchild.nodeType == TEXT: data = lastchild.data.rstrip() + before lastchild.data = data norm = 0 if wsmap[node.tagName]: nextnode = node.nextSibling if nextnode and nextnode.nodeType == TEXT: nextnode.data = after + nextnode.data.lstrip() else: wsnode = doc.createTextNode(after) node.parentNode.insertBefore(wsnode, nextnode) # hack to get the title in place: if node.tagName == "title" \ and node.parentNode.firstChild.nodeType == ELEMENT: node.parentNode.insertBefore(doc.createTextNode("\n "), node.parentNode.firstChild) node.parentNode.normalize() def normalize(doc): for node in doc.childNodes: if node.nodeType == ELEMENT: node.normalize() def cleanup_trailing_parens(doc, element_names): d = {} for gi in element_names: d[gi] = gi rewrite_element = d.has_key queue = [node for node in doc.childNodes if node.nodeType == ELEMENT] while queue: node = queue[0] del queue[0] if rewrite_element(node.tagName): lastchild = node.lastChild if lastchild and lastchild.nodeType == TEXT: data = lastchild.data if data.endswith("()"): lastchild.data = data[:-2] else: for child in node.childNodes: if child.nodeType == ELEMENT: queue.append(child) def contents_match(left, right): left_children = left.childNodes right_children = right.childNodes if len(left_children) != len(right_children): return 0 for l, r in map(None, left_children, right_children): nodeType = l.nodeType if nodeType != r.nodeType: return 0 if nodeType == ELEMENT: if l.tagName != r.tagName: return 0 # should check attributes, but that's not a problem here if not contents_match(l, r): return 0 elif nodeType == TEXT: if l.data != r.data: return 0 else: # not quite right, but good enough return 0 return 1 def create_module_info(doc, section): # Heavy. node = extract_first_element(section, "modulesynopsis") if node is None: return set_tagName(node, "synopsis") lastchild = node.childNodes[-1] if lastchild.nodeType == TEXT \ and lastchild.data[-1:] == ".": lastchild.data = lastchild.data[:-1] modauthor = extract_first_element(section, "moduleauthor") if modauthor: set_tagName(modauthor, "author") modauthor.appendChild(doc.createTextNode( modauthor.getAttribute("name"))) modauthor.removeAttribute("name") platform = extract_first_element(section, "platform") if section.tagName == "section": modinfo_pos = 2 modinfo = doc.createElement("moduleinfo") moddecl = extract_first_element(section, "declaremodule") name = None if moddecl: modinfo.appendChild(doc.createTextNode("\n ")) name = moddecl.attributes["name"].value namenode = doc.createElement("name") namenode.appendChild(doc.createTextNode(name)) modinfo.appendChild(namenode) type = moddecl.attributes.get("type") if type: type = type.value modinfo.appendChild(doc.createTextNode("\n ")) typenode = doc.createElement("type") typenode.appendChild(doc.createTextNode(type)) modinfo.appendChild(typenode) versionadded = extract_first_element(section, "versionadded") if versionadded: modinfo.setAttribute("added", versionadded.getAttribute("version")) title = get_first_element(section, "title") if title: children = title.childNodes if len(children) >= 2 \ and children[0].nodeName == "module" \ and children[0].childNodes[0].data == name: # this is it; morph the <title> into <short-synopsis> first_data = children[1] if first_data.data[:4] == " ---": first_data.data = first_data.data[4:].lstrip() set_tagName(title, "short-synopsis") if children[-1].nodeType == TEXT \ and children[-1].data[-1:] == ".": children[-1].data = children[-1].data[:-1] section.removeChild(title) section.removeChild(section.childNodes[0]) title.removeChild(children[0]) modinfo_pos = 0 else: ewrite("module name in title doesn't match" " <declaremodule/>; no <short-synopsis/>\n") else: ewrite("Unexpected condition: <section/> without <title/>\n") modinfo.appendChild(doc.createTextNode("\n ")) modinfo.appendChild(node) if title and not contents_match(title, node): # The short synopsis is actually different, # and needs to be stored: modinfo.appendChild(doc.createTextNode("\n ")) modinfo.appendChild(title) if modauthor:Lines 1–500 of 1073. Load more
#! /usr/bin/env python """Perform massive transformations on a document tree created from the LaTeX of the Python documentation, and dump the ESIS data for the transformed tree. """ import errno import esistools import re import sys import xml.dom import xml.dom.minidom ELEMENT = xml.dom.Node.ELEMENT_NODE ENTITY_REFERENCE = xml.dom.Node.ENTITY_REFERENCE_NODE TEXT = xml.dom.Node.TEXT_NODE class ConversionError(Exception): pass ewrite = sys.stderr.write try: # We can only do this trick on Unix (if tput is on $PATH)! if sys.platform != "posix" or not sys.stderr.isatty(): raise ImportError import commands except ImportError: bwrite = ewrite else: def bwrite(s, BOLDON=commands.getoutput("tput bold"), BOLDOFF=commands.getoutput("tput sgr0")): ewrite("%s%s%s" % (BOLDON, s, BOLDOFF)) PARA_ELEMENT = "para" DEBUG_PARA_FIXER = 0 if DEBUG_PARA_FIXER: def para_msg(s): ewrite("*** %s\n" % s) else: def para_msg(s): pass def get_first_element(doc, gi): for n in doc.childNodes: if n.nodeName == gi: return n def extract_first_element(doc, gi): node = get_first_element(doc, gi) if node is not None: doc.removeChild(node) return node def get_documentElement(node): result = None for child in node.childNodes: if child.nodeType == ELEMENT: result = child return result def set_tagName(elem, gi): elem.nodeName = elem.tagName = gi def find_all_elements(doc, gi): nodes = [] if doc.nodeName == gi: nodes.append(doc) for child in doc.childNodes: if child.nodeType == ELEMENT: if child.tagName == gi: nodes.append(child) for node in child.getElementsByTagName(gi): nodes.append(node) return nodes def find_all_child_elements(doc, gi): nodes = [] for child in doc.childNodes: if child.nodeName == gi: nodes.append(child) return nodes def find_all_elements_from_set(doc, gi_set): return __find_all_elements_from_set(doc, gi_set, []) def __find_all_elements_from_set(doc, gi_set, nodes): if doc.nodeName in gi_set: nodes.append(doc) for child in doc.childNodes: if child.nodeType == ELEMENT: __find_all_elements_from_set(child, gi_set, nodes) return nodes def simplify(doc, fragment): # Try to rationalize the document a bit, since these things are simply # not valid SGML/XML documents as they stand, and need a little work. documentclass = "document" inputs = [] node = extract_first_element(fragment, "documentclass") if node is not None: documentclass = node.getAttribute("classname") node = extract_first_element(fragment, "title") if node is not None: inputs.append(node) # update the name of the root element node = get_first_element(fragment, "document") if node is not None: set_tagName(node, documentclass) # Move everything that comes before this node into this node; # this will be the document element. nodelist = fragment.childNodes point = node.firstChild while not nodelist[0].isSameNode(node): node.insertBefore(nodelist[0], point) while 1: node = extract_first_element(fragment, "input") if node is None: break inputs.append(node) if inputs: docelem = get_documentElement(fragment) inputs.reverse() for node in inputs: text = doc.createTextNode("\n") docelem.insertBefore(text, docelem.firstChild) docelem.insertBefore(node, text) docelem.insertBefore(doc.createTextNode("\n"), docelem.firstChild) while fragment.firstChild and fragment.firstChild.nodeType == TEXT: fragment.removeChild(fragment.firstChild) def cleanup_root_text(doc): discards = [] skip = 0 for n in doc.childNodes: prevskip = skip skip = 0 if n.nodeType == TEXT and not prevskip: discards.append(n) elif n.nodeName == "COMMENT": skip = 1 for node in discards: doc.removeChild(node) DESCRIPTOR_ELEMENTS = ( "cfuncdesc", "cvardesc", "ctypedesc", "classdesc", "memberdesc", "memberdescni", "methoddesc", "methoddescni", "excdesc", "funcdesc", "funcdescni", "opcodedesc", "datadesc", "datadescni", ) def fixup_descriptors(doc, fragment): sections = find_all_elements(fragment, "section") for section in sections: find_and_fix_descriptors(doc, section) def find_and_fix_descriptors(doc, container): children = container.childNodes for child in children: if child.nodeType == ELEMENT: tagName = child.tagName if tagName in DESCRIPTOR_ELEMENTS: rewrite_descriptor(doc, child) elif tagName == "subsection": find_and_fix_descriptors(doc, child) def rewrite_descriptor(doc, descriptor): # # Do these things: # 1. Add an "index='no'" attribute to the element if the tagName # ends in 'ni', removing the 'ni' from the name. # 2. Create a <signature> from the name attribute # 2a.Create an <args> if it appears to be available. # 3. Create additional <signature>s from <*line{,ni}> elements, # if found. # 4. If a <versionadded> is found, move it to an attribute on the # descriptor. # 5. Move remaining child nodes to a <description> element. # 6. Put it back together. # # 1. descname = descriptor.tagName index = descriptor.getAttribute("name") != "no" desctype = descname[:-4] # remove 'desc' linename = desctype + "line" if not index: linename = linename + "ni" # 2. signature = doc.createElement("signature") name = doc.createElement("name") signature.appendChild(doc.createTextNode("\n ")) signature.appendChild(name) name.appendChild(doc.createTextNode(descriptor.getAttribute("name"))) descriptor.removeAttribute("name") # 2a. if descriptor.hasAttribute("var"): if descname != "opcodedesc": raise RuntimeError, \ "got 'var' attribute on descriptor other than opcodedesc" variable = descriptor.getAttribute("var") if variable: args = doc.createElement("args") args.appendChild(doc.createTextNode(variable)) signature.appendChild(doc.createTextNode("\n ")) signature.appendChild(args) descriptor.removeAttribute("var") newchildren = [signature] children = descriptor.childNodes pos = skip_leading_nodes(children) if pos < len(children): child = children[pos] if child.nodeName == "args": # move <args> to <signature>, or remove if empty: child.parentNode.removeChild(child) if len(child.childNodes): signature.appendChild(doc.createTextNode("\n ")) signature.appendChild(child) signature.appendChild(doc.createTextNode("\n ")) # 3, 4. pos = skip_leading_nodes(children, pos) while pos < len(children) \ and children[pos].nodeName in (linename, "versionadded"): if children[pos].tagName == linename: # this is really a supplemental signature, create <signature> oldchild = children[pos].cloneNode(1) try: sig = methodline_to_signature(doc, children[pos]) except KeyError: print oldchild.toxml() raise newchildren.append(sig) else: # <versionadded added=...> descriptor.setAttribute( "added", children[pos].getAttribute("version")) pos = skip_leading_nodes(children, pos + 1) # 5. description = doc.createElement("description") description.appendChild(doc.createTextNode("\n")) newchildren.append(description) move_children(descriptor, description, pos) last = description.childNodes[-1] if last.nodeType == TEXT: last.data = last.data.rstrip() + "\n " # 6. # should have nothing but whitespace and signature lines in <descriptor>; # discard them while descriptor.childNodes: descriptor.removeChild(descriptor.childNodes[0]) for node in newchildren: descriptor.appendChild(doc.createTextNode("\n ")) descriptor.appendChild(node) descriptor.appendChild(doc.createTextNode("\n")) def methodline_to_signature(doc, methodline): signature = doc.createElement("signature") signature.appendChild(doc.createTextNode("\n ")) name = doc.createElement("name") name.appendChild(doc.createTextNode(methodline.getAttribute("name"))) methodline.removeAttribute("name") signature.appendChild(name) if len(methodline.childNodes): args = doc.createElement("args") signature.appendChild(doc.createTextNode("\n ")) signature.appendChild(args) move_children(methodline, args) signature.appendChild(doc.createTextNode("\n ")) return signature def move_children(origin, dest, start=0): children = origin.childNodes while start < len(children): node = children[start] origin.removeChild(node) dest.appendChild(node) def handle_appendix(doc, fragment): # must be called after simplfy() if document is multi-rooted to begin with docelem = get_documentElement(fragment) toplevel = docelem.tagName == "manual" and "chapter" or "section" appendices = 0 nodes = [] for node in docelem.childNodes: if appendices: nodes.append(node) elif node.nodeType == ELEMENT: appnodes = node.getElementsByTagName("appendix") if appnodes: appendices = 1 parent = appnodes[0].parentNode parent.removeChild(appnodes[0]) parent.normalize() if nodes: map(docelem.removeChild, nodes) docelem.appendChild(doc.createTextNode("\n\n\n")) back = doc.createElement("back-matter") docelem.appendChild(back) back.appendChild(doc.createTextNode("\n")) while nodes and nodes[0].nodeType == TEXT \ and not nodes[0].data.strip(): del nodes[0] map(back.appendChild, nodes) docelem.appendChild(doc.createTextNode("\n")) def handle_labels(doc, fragment): for label in find_all_elements(fragment, "label"): id = label.getAttribute("id") if not id: continue parent = label.parentNode parentTagName = parent.tagName if parentTagName == "title": parent.parentNode.setAttribute("id", id) else: parent.setAttribute("id", id) # now, remove <label id="..."/> from parent: parent.removeChild(label) if parentTagName == "title": parent.normalize() children = parent.childNodes if children[-1].nodeType == TEXT: children[-1].data = children[-1].data.rstrip() def fixup_trailing_whitespace(doc, fragment, wsmap): queue = [fragment] fixups = [] while queue: node = queue[0] del queue[0] if wsmap.has_key(node.nodeName): fixups.append(node) for child in node.childNodes: if child.nodeType == ELEMENT: queue.append(child) # reverse the list to process from the inside out fixups.reverse() for node in fixups: node.parentNode.normalize() lastchild = node.lastChild before, after = wsmap[node.tagName] if lastchild.nodeType == TEXT: data = lastchild.data.rstrip() + before lastchild.data = data norm = 0 if wsmap[node.tagName]: nextnode = node.nextSibling if nextnode and nextnode.nodeType == TEXT: nextnode.data = after + nextnode.data.lstrip() else: wsnode = doc.createTextNode(after) node.parentNode.insertBefore(wsnode, nextnode) # hack to get the title in place: if node.tagName == "title" \ and node.parentNode.firstChild.nodeType == ELEMENT: node.parentNode.insertBefore(doc.createTextNode("\n "), node.parentNode.firstChild) node.parentNode.normalize() def normalize(doc): for node in doc.childNodes: if node.nodeType == ELEMENT: node.normalize() def cleanup_trailing_parens(doc, element_names): d = {} for gi in element_names: d[gi] = gi rewrite_element = d.has_key queue = [node for node in doc.childNodes if node.nodeType == ELEMENT] while queue: node = queue[0] del queue[0] if rewrite_element(node.tagName): lastchild = node.lastChild if lastchild and lastchild.nodeType == TEXT: data = lastchild.data if data.endswith("()"): lastchild.data = data[:-2] else: for child in node.childNodes: if child.nodeType == ELEMENT: queue.append(child) def contents_match(left, right): left_children = left.childNodes right_children = right.childNodes if len(left_children) != len(right_children): return 0 for l, r in map(None, left_children, right_children): nodeType = l.nodeType if nodeType != r.nodeType: return 0 if nodeType == ELEMENT: if l.tagName != r.tagName: return 0 # should check attributes, but that's not a problem here if not contents_match(l, r): return 0 elif nodeType == TEXT: if l.data != r.data: return 0 else: # not quite right, but good enough return 0 return 1 def create_module_info(doc, section): # Heavy. node = extract_first_element(section, "modulesynopsis") if node is None: return set_tagName(node, "synopsis") lastchild = node.childNodes[-1] if lastchild.nodeType == TEXT \ and lastchild.data[-1:] == ".": lastchild.data = lastchild.data[:-1] modauthor = extract_first_element(section, "moduleauthor") if modauthor: set_tagName(modauthor, "author") modauthor.appendChild(doc.createTextNode( modauthor.getAttribute("name"))) modauthor.removeAttribute("name") platform = extract_first_element(section, "platform") if section.tagName == "section": modinfo_pos = 2 modinfo = doc.createElement("moduleinfo") moddecl = extract_first_element(section, "declaremodule") name = None if moddecl: modinfo.appendChild(doc.createTextNode("\n ")) name = moddecl.attributes["name"].value namenode = doc.createElement("name") namenode.appendChild(doc.createTextNode(name)) modinfo.appendChild(namenode) type = moddecl.attributes.get("type") if type: type = type.value modinfo.appendChild(doc.createTextNode("\n ")) typenode = doc.createElement("type") typenode.appendChild(doc.createTextNode(type)) modinfo.appendChild(typenode) versionadded = extract_first_element(section, "versionadded") if versionadded: modinfo.setAttribute("added", versionadded.getAttribute("version")) title = get_first_element(section, "title") if title: children = title.childNodes if len(children) >= 2 \ and children[0].nodeName == "module" \ and children[0].childNodes[0].data == name: # this is it; morph the <title> into <short-synopsis> first_data = children[1] if first_data.data[:4] == " ---": first_data.data = first_data.data[4:].lstrip() set_tagName(title, "short-synopsis") if children[-1].nodeType == TEXT \ and children[-1].data[-1:] == ".": children[-1].data = children[-1].data[:-1] section.removeChild(title) section.removeChild(section.childNodes[0]) title.removeChild(children[0]) modinfo_pos = 0 else: ewrite("module name in title doesn't match" " <declaremodule/>; no <short-synopsis/>\n") else: ewrite("Unexpected condition: <section/> without <title/>\n") modinfo.appendChild(doc.createTextNode("\n ")) modinfo.appendChild(node) if title and not contents_match(title, node): # The short synopsis is actually different, # and needs to be stored: modinfo.appendChild(doc.createTextNode("\n ")) modinfo.appendChild(title) if modauthor: