From: John L. Clark (jlc6@po.cwru.edu)
Date: Sun Apr 14 2002 - 23:18:35 EDT
Guys,
After exporting the attachment for this bug to HTML, I get the following
result:
http://djinn.dnsalias.net/abiword/bug1956
which at least is clean and approaches the desired goal. This is with
CVS HEAD, and tested in Konqueror, Netscape, Mozilla (some older build,
blah), and of course, w3m. ;-)
There is only one way that it does not reach the desired goal. That is,
it does not make note of when a list stops and then restarts later in
the document, as is allowed in AbiWord docs, but not (precisely) allowed
in HTML. That is, you can start the numbering where the list leaves
off, but you can't actually keep it all in one list.
To solve this problem, I had thought to record where each list in a
document was in its count, and then feed this information to the HTML.
To do this, however, I need some sort of hash table for
(list ID)->(count), but I can't seem to find a hash table in .../af/util
anywhere. What happened to it? Further, this is not a particularly
clean solution. One alternate is to use CSS counters[1], but these are
supported by exactly 0 browsers that I know of, so I don't know how
valuable they are. They do have the advantage that they are
(theoretically?) extremely flexible, and would be able to emulate all of
our list display types. And, of course, you can use one counter over
multiple lists, which indicates that the counting should be conistent
across those lists. Does anyone else have any other ideas?
Additionally, I would like to know right now if anyone has any other
1.0 critical bugs with the HTML exporter, or related to this bug in
particular. Does everyone else get good list generation in HTML? Other
elements, styles, etc?
Take care,
John
[1] http://www.w3.org/TR/REC-CSS2/generate.html#counters
This archive was generated by hypermail 2.1.4 : Sun Apr 14 2002 - 23:19:08 EDT