diff options
author | Fred Drake <fdrake@acm.org> | 1999-08-02 14:46:15 (GMT) |
---|---|---|
committer | Fred Drake <fdrake@acm.org> | 1999-08-02 14:46:15 (GMT) |
commit | 080c1b5af693afd471728e7fa66f744a61e75ed4 (patch) | |
tree | 584894ee7a50bb4d0bfcc7f4af91a9f75ab46245 /Doc | |
parent | 4fbdf9779c968cf0827c3191dd4a00361d2354fc (diff) | |
download | cpython-080c1b5af693afd471728e7fa66f744a61e75ed4.zip cpython-080c1b5af693afd471728e7fa66f744a61e75ed4.tar.gz cpython-080c1b5af693afd471728e7fa66f744a61e75ed4.tar.bz2 |
bwrite(), ewrite(): Helpers for diagnostic output; essentially the
same as sys.stderr.write(), but bwrite() will make it
bold if stderr is a tty (for visibility).
Use these everywhere instead of accessing
sys.stderr.write() directly.
find_all_child_elements(): Find all elements by type name that are
immediate children of another.
cleanup_synopses(): Receive both the document and fragment objects as
parameters; if only the document is passed, synopsis
information is never found since it's in the fragment!
build_para(): Add a chunk of code to ensure that a newline always
follows the </para> end-tag. Not really important, but
makes presentation consistent for authoring/
maintenance.
fixup_bifuncindexes_chunk(): Use find_all_child_elements() instead of
find_all_elements() so we don't work with too many
levels at once; that could cause DOM exceptions with
the <elem>.removeChild() method.
Diffstat (limited to 'Doc')
-rwxr-xr-x | Doc/tools/sgmlconv/docfixer.py | 91 |
1 files changed, 61 insertions, 30 deletions
diff --git a/Doc/tools/sgmlconv/docfixer.py b/Doc/tools/sgmlconv/docfixer.py index 5328a50..983601b 100755 --- a/Doc/tools/sgmlconv/docfixer.py +++ b/Doc/tools/sgmlconv/docfixer.py @@ -22,13 +22,28 @@ 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 curses + 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): - sys.stderr.write("*** %s\n" % s) + ewrite("*** %s\n" % s) else: def para_msg(s): pass @@ -81,14 +96,23 @@ def find_all_elements(doc, gi): nodes.append(node) return nodes -def find_all_elements_from_set(doc, gi_set, nodes=None): - if nodes is None: - nodes = [] +def find_all_child_elements(doc, gi): + nodes = [] + for child in doc.childNodes: + if child.nodeType == ELEMENT: + if child.tagName == 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.nodeType == ELEMENT and doc.tagName in gi_set: nodes.append(doc) for child in doc.childNodes: if child.nodeType == ELEMENT: - find_all_elements_from_set(child, gi_set, nodes) + __find_all_elements_from_set(child, gi_set, nodes) return nodes @@ -165,7 +189,7 @@ def find_and_fix_descriptors(doc, container): def rewrite_descriptor(doc, descriptor): # # Do these things: - # 1. Add an "index=noindex" attribute to the element if the tagName + # 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 and <args>. # 3. Create additional <signature>s from <*line{,ni}> elements, @@ -179,7 +203,7 @@ def rewrite_descriptor(doc, descriptor): index = 1 if descname[-2:] == "ni": descname = descname[:-2] - descriptor.setAttribute("index", "noindex") + descriptor.setAttribute("index", "no") descriptor._node.name = descname index = 0 desctype = descname[:-4] # remove 'desc' @@ -459,12 +483,10 @@ def create_module_info(doc, section): title.removeChild(children[0]) modinfo_pos = 0 else: - sys.stderr.write( - "module name in title doesn't match" - " <declaremodule>; no <short-synopsis>\n") + ewrite("module name in title doesn't match" + " <declaremodule/>; no <short-synopsis/>\n") else: - sys.stderr.write( - "Unexpected condition: <section> without <title>\n") + ewrite("Unexpected condition: <section/> without <title/>\n") modinfo.appendChild(doc.createTextNode("\n ")) modinfo.appendChild(node) if title and not contents_match(title, node): @@ -484,7 +506,7 @@ def create_module_info(doc, section): # # The rest of this removes extra newlines from where we cut out # a lot of elements. A lot of code for minimal value, but keeps - # keeps the generated SGML from being too funny looking. + # keeps the generated *ML from being too funny looking. # section.normalize() children = section.childNodes @@ -499,8 +521,8 @@ def create_module_info(doc, section): nextnode.data = "\n\n\n" + string.lstrip(data) -def cleanup_synopses(doc): - for node in find_all_elements(doc, "section"): +def cleanup_synopses(doc, fragment): + for node in find_all_elements(fragment, "section"): create_module_info(doc, node) @@ -624,7 +646,7 @@ PARA_LEVEL_PRECEEDERS = ( "index", "indexii", "indexiii", "indexiv", "setindexsubitem", "stindex", "obindex", "COMMENT", "label", "input", "title", "versionadded", "versionchanged", "declaremodule", "modulesynopsis", - "moduleauthor", + "moduleauthor", "indexterm", ) @@ -710,9 +732,19 @@ def build_para(doc, parent, start, i): prev = node if have_last: parent.appendChild(para) + parent.appendChild(doc.createTextNode("\n\n")) return len(parent.childNodes) else: - parent.insertBefore(para, parent.childNodes[start]) + nextnode = parent.childNodes[start] + if nextnode.nodeType == TEXT: + if nextnode.data and nextnode.data[0] != "\n": + nextnode.data = "\n" + nextnode.data + else: + newnode = doc.createTextNode("\n") + parent.insertBefore(newnode, nextnode) + nextnode = newnode + start = start + 1 + parent.insertBefore(para, nextnode) return start + 1 @@ -838,7 +870,7 @@ def fixup_refmodindexes(fragment): def fixup_refmodindexes_chunk(container): # node is probably a <para>; let's see how often it isn't: if container.tagName != PARA_ELEMENT: - sys.stderr.write("--- fixup_refmodindexes_chunk(%s)\n" % container) + bwrite("--- fixup_refmodindexes_chunk(%s)\n" % container) module_entries = find_all_elements(container, "module") if not module_entries: return @@ -847,10 +879,9 @@ def fixup_refmodindexes_chunk(container): for entry in index_entries: children = entry.childNodes if len(children) != 0: - sys.stderr.write( - "--- unexpected number of children for %s node:\n" - % entry.tagName) - sys.stderr.write(entry.toxml() + "\n") + bwrite("--- unexpected number of children for %s node:\n" + % entry.tagName) + ewrite(entry.toxml() + "\n") continue found = 0 module_name = entry.getAttribute("name") @@ -860,7 +891,7 @@ def fixup_refmodindexes_chunk(container): this_name = node.childNodes[0].data if this_name == module_name: found = 1 - node.setAttribute("index", "index") + node.setAttribute("index", "yes") if found: removes.append(entry) for node in removes: @@ -870,6 +901,7 @@ def fixup_refmodindexes_chunk(container): def fixup_bifuncindexes(fragment): nodes = find_all_elements(fragment, 'bifuncindex') d = {} + # make sure that each parent is only processed once: for node in nodes: parent = node.parentNode d[parent._node.node_id] = parent @@ -879,8 +911,8 @@ def fixup_bifuncindexes(fragment): def fixup_bifuncindexes_chunk(container): removes = [] - entries = find_all_elements(container, "bifuncindex") - function_entries = find_all_elements(container, "function") + entries = find_all_child_elements(container, "bifuncindex") + function_entries = find_all_child_elements(container, "function") for entry in entries: function_name = entry.getAttribute("name") found = 0 @@ -890,12 +922,11 @@ def fixup_bifuncindexes_chunk(container): continue t2 = t2[:-2] if t2 == function_name: - - func_entry.setAttribute("index", "index") + func_entry.setAttribute("index", "yes") func_entry.setAttribute("module", "__builtin__") if not found: - removes.append(entry) found = 1 + removes.append(entry) for entry in removes: container.removeChild(entry) @@ -948,8 +979,8 @@ def convert(ifp, ofp): "subparagraph": "\n\n", }) cleanup_root_text(doc) - cleanup_trailing_parens(doc, ["function", "method", "cfunction"]) - cleanup_synopses(doc) + cleanup_trailing_parens(fragment, ["function", "method", "cfunction"]) + cleanup_synopses(doc, fragment) fixup_descriptors(doc, fragment) fixup_verbatims(fragment) normalize(fragment) |