Remove duplicate files from doclib

[senf.git] / doclib / fix-links.py

diff --git a/doclib/fix-links.py b/doclib/fix-links.py
index 440a33b..fb7450e 100644 (file)
--- a/doclib/fix-links.py
+++ b/doclib/fix-links.py
@@ -1,6 +1,32 @@
 #!/usr/bin/python
-
-import sys,os.path,fnmatch, HTMLParser, getopt, re
+#
+# This tool will hack the doxygen generated documentation to fix link
+# errors produced by doxygen.
+#
+# This works because most anchors doxygen generates are unique 32 char
+# hash values. To speed up the operation, the tool will not check all
+# the files itself but will let 'linklint' do the grunt
+# work. fix-links.py reads the 'errorX.txt' and 'errorAX.txt' files
+# generated by linklint. These files list links to missing files
+# (errorX.html) and links to missing anchors
+# (errorAX.html). fix-links.py works in the following way:
+#
+# - Build a complete index of all unique anchors found in any html
+#   file. The index will only include *unique* anchors. Anchors found
+#   multiple times are removed from the index
+#
+# - The index is extended to include all unique names of html files
+#
+# - Scn the linklint result and check the bad links against the
+#   index. If the file or anchor is found in the index, an accoringly
+#   corrected link is generated otherwise the link is removed.
+#
+# One additional twak is, that fix-links.py will successively remove
+# initial 'g' charachters from anchors until the link is found in the
+# index. Doxygen seems to create links with the wrong number of 'g'
+# charachters in front sometimes.
+
+import sys,os.path,fnmatch, HTMLParser, getopt, re, codecs
 
 class HTMLFilter(HTMLParser.HTMLParser):
 
@@ -187,11 +213,11 @@ class LinkFixer:
 
     def fix(self, path, target):
         self._files += 1
-        data = file(path).read()
+        data = codecs.open(path, "r", "utf-8").read()
         filt = LinkFixer.LinkFilter(self._index,
                                     target,
                                     "../" * (len(os.path.split(path)[0].split("/"))),
-                                    file(path,"w"))
+                                    codecs.open(path, "w", "utf-8") )
         filt.feed(data)
         filt.close()
         found, fixed, removed = filt.stats()
@@ -206,8 +232,7 @@ class LinkFixer:
     
 (opts, args) = getopt.getopt(sys.argv[1:], "vs:")
 if len(args) != 2:
-    sys.stderr.write("""Usage:
-       fix-links.py [-s skip-dir]... <errrorX.txt> <errorAX.txt>
+    sys.stderr.write("""Usage: fix-links.py [-s skip-dir]... <errrorX.txt> <errorAX.txt>
 
 Process the 'errorX.txt' and 'errorAX.txt' files as generated by
 'linklint': Check all invalid links and try to find the correct
@@ -256,7 +281,7 @@ total, found, fixed, removed = fixer.stats()
 if verbose:
     sys.stderr.write("\nRemoved links:\n")
     for (anchor, label), files in removed.items():
-        sys.stdout.write("%-36.36s %-48.48s %s\n"
+        sys.stderr.write("%-36.36s %-48.48s %s\n"
                          % ( anchor,
                              "(%s)" % label[:46],
                              " ".join(files.keys())) )