On Fri, Aug 22, 2003 at 08:57:24AM +0200, Andre Poenitz spake thusly: > On Fri, Aug 22, 2003 at 12:12:23AM +0300, Martin Vermeer wrote: > > On Thu, Aug 21, 2003 at 05:01:11PM +0100, John Levon spake thusly: > > > > > > On Thu, Aug 21, 2003 at 06:34:37PM +0300, Martin Vermeer wrote: > > > > > > > You may want to try the attached, which is only the changes to the > > > > insets (i.e. they don't use needFullRow/display). I would be interested > > > > if it works without these artefacts etc. for you. > > > > > > It works just as well as without the patch, yes. But then that's > > > expected: insets still aren't constrained to be inside both margins, and > > > inline tex still takes up a whole row, as before. > > > > Are you seeing the following, also with stock CVS of this moment: > > > > the text of an inset extends to the right under the scrollbar (i.e. > > some 20 pixels to the right of the right margin of the main text)? > > I always thought it extends exactly up to the scrollbar... > > > And the text of an inset inside an inset extends even twice that > > amount, so that a whole short word can go missing? > > > > That's the most serious rendering bug I am seeing. If that's not my > > fault, I'm a happy camper :-) > > > > My complete patch (including the text*.C parts) at least fixes the > > inline tex behaviour. > > > > > My vote is to just apply the full patch now, with all the if 0'd code > > > removed, and see what stuff we need to sort out. > > > > > > I would be a little more conservative ;-) I see the rendering bugs. > > There have been bugs in this area for weeks now. > > > ...but I will do the cleanup and return to the matter... sometime > > mid next week, I expect. > > Your call. But at least the 'redoParagraph' chunk will be broken then. > > Andre'
OK. I made the attached, sanitized patch after you committed your last
changes.
The inset margin bug is still there, but I have ideas on that. The
current code is pretty convoluted, esp the cursor positioning.
As this meshes well with what you're doing, feel free to test and
check it in as I will be 'patchily on-line' the coming week... if
you're happy with it, that is.
2003-08-23 Martin Vermeer <[EMAIL PROTECTED]>
* lyxrow_funcs.C:
* lyxtext.h:
* text.C:
* text2.C:
* insets/inset.h:
* insets/insetbibtex.[Ch]:
* insets/insetbranch.[Ch]:
* insets/insetcollapsable.h:
* insets/insetert.[Ch]:
* insets/insetfloatlist.[Ch]:
* insets/insetfootlike.[Ch]:
* insets/insetinclude.[Ch]:
* insets/insetindex.[Ch]:
* insets/insetindex.h:
* insets/insetminipage.h:
* insets/insetnote.[Ch]:
* insets/insettext.C:
* insets/insettheorem.[Ch]:
* insets/insettoc.[Ch]: eliminates the needFullRow/display() stuff
altogether, putting the logic in metrics/draw in the insets.
Draws short single row insets with more or less realistic metrics.
Added 225 lines, removed 222. Mostly moved from core to insets.
- Martin
--
Martin Vermeer [EMAIL PROTECTED]
Helsinki University of Technology
Dept. of Surveying, Inst. of Geodesy
P.O. Box 1200, FIN-02015 HUT, Finland
:wq
BTW what does scroll() do?
width.diff.gz
Description: application/gunzip
pgp00000.pgp
Description: PGP signature
