* Improved performance by creating text page/object rects only on first generatePixmap call
* Code cleanup
svn path=/trunk/playground/graphics/okular/; revision=635240
Add a protected method to get the document, and hide the real document pointer as private, so the generators can't redefine it.
svn path=/trunk/playground/graphics/okular/; revision=598025
* Moved all implementations to generator.cpp
* Added 'const' where it make sense
* Adapted all generators (except gs)
svn path=/trunk/playground/graphics/okular/; revision=597525
Adapted the poppler(pdf) and the chm generators to return the title of the current document.
svn path=/trunk/playground/graphics/okular/; revision=568059
Now every generator has to implement this one and put (if necessary) all the code for cleaning up all the stuff related to the currently open document.
For now the return value it is not read, but generators as strongly suggested to return tru o false, whether all the operations in there went fine.
svn path=/trunk/playground/graphics/okular/; revision=562210
- more const'ness in signals
- no need to redeclare the signals in the generators, as they are already in Generator
svn path=/branches/work/kde4/playground/graphics/okular/; revision=549941
- What works:
+ viewing documents
+ thumbnail generation
+ calculating page sizes based on the size of their KHTMLView
+ table of contents
- What is pending:
+ using internal index, will code it once I comeback, wasted too much time on textpage generation
- What probably will not work:
+ printing the entire document, no CHM viewer provides it now
it would need some kind of being able to print multiple HTML files
at once or sth similar (maybe anyone else has an idea?)
- About generating the textpage:
I went a long way, with enormous, and I do mean enormous help from SadEagle explaining to me how
to use the KHTML API to generate the text page in a usable way (actually getCaretPos is depreciated
and not document, and if you try to read the code, you'll notice it delegate the important stuff
further and further into the KHTML core. Understanding KHTML's core is a bit too much to require
at the moment.
Well finally I had something close to working, some words are matched properly - the rectangles are ok,
most are matched with a one/one and a half offshot to the right, finally some are not matched because
getCaretPost returns -1 as the left X border, I have no idea why, when I am asking for a
This is why I am ccing to kfm-devel, if anyone of you guys, has a while to spare and would be kind
enough, I would like to ask you to look at generator_t.cpp and the
void TGenerator::recursiveExploreNodes(DOM::Node node,KPDFTextPage *tp)
function, and tell me what seems wrong there? Also if you knwo a better way to get per character
coordinates for every character in the text inside a HTML Page, please let me know.
CCMAIL:kfm-devel@kde.org
svn path=/trunk/playground/graphics/oKular/kpdf/; revision=450785
Pino Toscano
Tobias Koenig
Albert Astals Cid
Piotr Szymanski
Thiago Macieira