headerid extension patch

[thread13]

Basically it is an update for the issue 54: an empty string is still considered to be a valid unique id.

As the code has gone from toc.py to, I guess, headerid.py, here is a new patch:

--- headerid.py 2012-05-15 15:36:49.000000000 +1000
+++ headerid.py.new 2012-05-15 16:04:25.000000000 +1000
@@ -97,7 +97,7 @@

 def unique(id, ids):
     """ Ensure id is unique in set of ids. Append '_1', '_2'... if not """
-    while id in ids:
+    while id in ids or not id :
         m = IDCOUNT_RE.match(id)
         if m:
             id = '%s_%d'% (m.group(1), int(m.group(2))+1)

basically anything that will check for a non-empty id value will do, e.g. setting

id = id or '_0'

before the while loop, etc ; I assume that the id in ids check is more likely to be true, thus the suggested version .

Again, the point was that for non-ascii headers slugify() returns a completely empty string ( as one can clearly see from its code ), so it looks as a good idea to turn it to something non-empty ; which function -- "slugify()" or "unique()" -- should perform that check, is of course completely a matter of the author's choice -- but as the "slugify()" function could be currently, AFAIR, provided by the user -- it might be a good idea to perform a non-zero check afterwards, i.e. in "unique()" .

Bottom line: here is a patch for the current version for the issue 54 )