A big blooper that i stuck in where the bintable to search in for
the end of a para of exception text was incorrect has also been
fixed, this was a very nasty one. Usually identified by text with
no formatting directly above an empty paragraph that had all
the correct formatting for the line above it, but no actual text :-),
that was another toughy. Heavy stuff entirely, my head is very sore
today.
A few other sprms have been fixed so there should be other improvements.
There is one outstanding problem, and that is that if a paragraph begins
in the fastsaved region of a word file that it isnt actually emitting
a para begin token. If the fastsaved portion is in the middle of a para
that begins in the fullsaved portion we are ok, and many of them do. But
if it starts in the beginning (at least sometimes) then it is getting
registered incorrectly as not being the beginning of a para. Im looking
into this one, but I thought id like to share my rather incoherent pride
in figuring out two of the larger problems with fastsaves.
The missing para begin issue can be seen in the flesh if you do
/wvHtml feature-examples/supported-font-features.doc > ! out
scroll down in out to see the para with the text "White text",
above it you can see that Yellow text is inside in a nice little
paragraph of its own, which White text is outside any para, while
the para that it should belong to is below it. Ill grind away at
that one for a day or two. Once that is complete I can trundle on to do
lists and tables.
There hasn't been any extensive testing of these changes, which affect
fastsaves only. wvSimpleCLX will tell you if a file is fast or full.
the development ver (0.5.23) is available at
http://www.csn.ul.ie/~caolan/publink/mswordview/development/
and at cvs checkout wv from abisource
Note that wv can handle word 6 and word 7 files with a fair chance of
success as well as word 8. (wvSummary will tell you what ver the file is,
wvSummary is a summarystream displayer for all ole2 files, and can be
integrated into file, http://www.csn.ul.ie/~caolan/publink/file/,
though mind you while the file guy said he liked it, there has been
a notable lack of an updated file, anyway...) wv only handles para and
char properties, no tables lists etc. One solid step at a time.
C.
Real Life: Caolan McNamara * Doing: MSc in HCI
Work: Caolan.McNamara@ul.ie * Phone: +353-86-8790257
URL: http://www.csn.ul.ie/~caolan * Sig: an oblique strategy
Re-evaluation (a warm feeling)