Re: commit -- turned on FASTSQUIGGLE logic

Stephen Hack (shack@uiuc.edu)
Thu, 28 Oct 1999 15:24:57 -0500


I'm pretty sure the spaces are dropped in this case, since the problem lie in
the attempt to spell check a word (which does not exist)

On Thu, Oct 28, 1999 at 03:48:10PM -0400, Richard Jefts wrote:
> Adding right after it gets the pBlockText
>
> if (pBlockText == NULL) {
> return;
> }
>
> should take care of the problem. That's what I did for the crashing when
> changing font style on an empty doc (same prob, different function). In
> this case, it might hide a deeper problem. i'm not that familiar with the
> block internals. Shouldn't the block contain the space?
>
> richie
>
> On Thu, 28 Oct 1999, Stephen Hack wrote:
>
> >
> > Paul,
> >
> > I was get the following Asserts when I would
> > type space followed by an <Enter>
> > or left a trailing space at the end of a paragraph
> >
> > -shack
> >
> > **** (3) Assert ****
> > **** (3) !UT_isWordDelimiter( pBlockText[wordBeginning + wordLength] ) at
> > fl_BlockLayout.cpp:1551 ****
> > **** (3) Continue ? (y/n) [y] : y
> > DEBUG: Edit:InsertFmtMark [blockOffset 0]
> >
> >
> > **** (4) Assert ****
> > **** (4) eor <= pgb.getLength() at fl_BlockLayout.cpp:1547 ****
> > **** (4) Continue ? (y/n) [y] :
> >
> >
> > On Thu, Oct 21, 1999 at 09:58:29PM -0700, Paul Rohr wrote:
> > > Minimize spell activity during editing. Fixes #131, 133.
> > >
> > > M src/text/fmt/xp/fl_BlockLayout.cpp
> > > M src/text/fmt/xp/fl_BlockLayout.h
> > > M src/text/fmt/xp/fl_DocLayout.cpp
> > > M src/text/fmt/xp/fv_View.cpp
> > > M src/text/fmt/xp/fv_View.h
> > >
> > > OK, I finally went back in and finished up the code I started last spring
> > > for this. It seems to be working pretty smoothly, so I've checked it in.
> > > It could certainly use more testing, so hammer away at it!
> > >
> > > (Jeff's usually the one who gets to introduce destabilizing stuff just after
> > > we cut a release, but he's travelling this week, so I grabbed my chance.)
> > >
> > > Note that this change doesn't affect the existing dialog-driven or
> > > timer-driven spell logic at all. It only functions at edit time, replacing
> > > the old brain-damaged strategy (recheck the entire block for every
> > > keystroke) with a much more streamlined approach.
> > >
> > > Enjoy!
> > >
> > > Paul
> > >
> > > PS: If things get really gruesome, just turn off the #define FASTSQUIGGLE
> > > down in the bowels of fl_BlockLayout.cpp, and you should be back to the poky
> > > old reliable logic. :-)
> > >
> >



This archive was generated by hypermail 1.03b2.