GitLab

Still some alpha problems but this solved a lot

svn path=/trunk/koffice/; revision=488285

svn path=/trunk/koffice/; revision=488278

Unfortunately it's not used very much

Make UINT16_BLEND and UINT8_BLEND work correctly in 
every case. Signed arithmetic is needed.


svn path=/trunk/koffice/; revision=488271

- add support for ruby
- add a sample for ruby

svn path=/trunk/koffice/; revision=488251

  virtual QPoint applyViewTransformations( const QPoint& ) const;
  virtual QPoint reverseViewTransformations( const QPoint& ) const;
to KoView. The goal is to replace the matrix() virtual, which isn't flexible
enough to handle e.g. kword's preview mode with 4 pages per row.

I hope this doesn't break other koffice apps (embedding in kspread still
seems to work, at least when not using rotations. I doubt it fully worked
with rotations before this commit?). Anyhow, please double check if any
regressions are introduced in child doc handling.

Also made the KoPoint(QPoint) constructor explicit to avoid surprises.
Commented out some unused methods in KoChild.

This commit is BIC, make sure to recompile all koffice apps after upgrading lib.
CCMAIL: koffice-devel@kde.org

svn path=/trunk/koffice/; revision=488240

svn path=/trunk/koffice/; revision=488151

svn path=/trunk/koffice/; revision=488150

svn path=/trunk/koffice/; revision=488122

on ppc.

svn path=/trunk/koffice/; revision=488121

svn path=/trunk/koffice/; revision=488035

svn path=/trunk/koffice/; revision=488011

svn path=/trunk/koffice/; revision=488010

This way we are sure to notice it before release


svn path=/trunk/koffice/; revision=487985

Still not ready for primetime yet


svn path=/trunk/koffice/; revision=487971

svn path=/trunk/koffice/; revision=487962

in KoDocumentChild soon), and renamed the files to start with K while being at it.

svn path=/trunk/koffice/; revision=487956

svn path=/trunk/koffice/; revision=487901

svn path=/trunk/koffice/; revision=487857

svn path=/trunk/koffice/; revision=487743

Change some return values to unsigned.


svn path=/trunk/koffice/; revision=487682

svn path=/trunk/koffice/; revision=487629

svn path=/trunk/koffice/; revision=487613

svn path=/trunk/koffice/; revision=487606

svn path=/trunk/koffice/; revision=487569

svn path=/trunk/koffice/; revision=487565

colorspaces and hasAlpha from paintdevices

It is safe to assume that every colorspace has 
alpha. Alpha should be accessed through setAlpha 
and getAlpha if needed. The best is of course to 
let the colorspace handle alpha itself.


svn path=/trunk/koffice/; revision=487538

svn path=/trunk/koffice/; revision=487535

svn path=/trunk/koffice/; revision=487528

svn path=/trunk/koffice/; revision=487521

arent used and wont be in the future


svn path=/trunk/koffice/; revision=487516

based on lcms.
Implement function in abstract instead and use
standard lcms deltaE


svn path=/trunk/koffice/; revision=487513

svn path=/trunk/koffice/; revision=487502

svn path=/trunk/koffice/; revision=487487

- add a wrapper for KisHistogram
- add support for displaying progression
- add undo support
- add reshapehisto.py an example script for histogram usage


svn path=/trunk/koffice/; revision=487481

a QColor isn't a colorspace, and to know what the rgb
triplet means, we need a profile.

svn path=/trunk/koffice/; revision=487467

So the general agreement seems to be that processMessages is bad, so disable it. It doesn't like to make a big impact anyway, so I'll let this be for a while.

svn path=/trunk/koffice/; revision=487442

* Add a way to safely call processEvents in the KisImageRasteredCache
* Multithread KisAccumulatingHistogramProducer. This really gives a massive apparant speedup, since painting a big image doesn't block on this anymore.

svn path=/trunk/koffice/; revision=487418

svn path=/trunk/koffice/; revision=487333

svn path=/trunk/koffice/; revision=486995

svn path=/trunk/koffice/; revision=486894