Re: commit (head): toggle case fixes

>

Hi Tomas,
         I don;t understand all your changes. But now I can't type a
simple footnote without hitting a whole lot of asserts. In
addition bug 7369 after
being fully fixed in CVS HEAD, is now back.

Can you see what is going wrong? The undo in pt_PT_insertSpan.cpp:435 is
producing change records for fmtmark that point to the wrong block strux.

I suspect you're somewhow producing fmtmark records with the incorrect
strux positions.

However that is a total guess at this point.

Please have a look.

Cheers

Martin

>
> various fixes to the toggle case code, including bug 6430.
>
> In order to get this sorted I had to make changes to
> pt_PieceTable::_insertFmtMarkFragWithNotify -- basically the original
> code tried to be too clever, doing things that are better done on
> higher level than the PT. The two significant changes are:
>
> (1) the fmt mark simply holds the fmt that has been passed down to
> it; it does not try to supperimpose it on props and attrs derrived
> from the context; that is the job of the caller. (I created
> FV_View::getAttrPropForPoint() to simplyfy handling of fmt continuity
> at point).
>
> (2) Immediately after the start of a strux, fmt mark is inserted,
> even if it seemingly has no effect -- without this the
> _chooseIndexAP() does not work.
>
> I also came across a situation where UT_Worker was firing recursively
> inside _sFrequentRepeat(), so I added code to stop that from
> happening.
>
> files: number in ptbl/xp and fv_View*.h/cpp, ap_EditMethods.cpp
>
> Tomas
>
>
>
>
Received on Sun Sep 19 14:47:36 2004