Hi Xavi, I don't understand this patch, it's against an older version of the files (missing my last minor update), you added conditional code for layered mode (why?) and you still try to force your setcancel hack. It also employs trick to track focus, and I'm not sure I like it.
IMO the issue to be resolved doesn't justify any sort of irregular tricks or any chance for performance drop down. Pls consult with Przemek about it. Brgds, Viktor On Sun, Apr 4, 2010 at 5:08 AM, Xavi <jara...@gmail.com> wrote: > Hi, > >> app crashes / hangs. EX_LAYERED will keep a >> buffer of the screen, which gives better >> performance for screen refreshes, and it also >> makes the app behave/look smoother. Regardless >> I think it's not the job of high-level code >> to ensure screen refresh. >> >> BTW, even with this I can see artifacts with >> GTWVT, f.e. visible pixel-wide vertical lines >> between full-block characters. Much worse is >> GTWIN, where complete pixels are missing/left >> on screen and which are Windows problems. > > Thank you very much for the hints Viktor and have a nice day. > > It's not a local problem of drivers because I can reproduce it on different > hardware/drives under Windows XP, also in VirtualBox. It's easy to reproduce > in the application that I'm moving to Harbour and in Clipper, under NTVDM, > does not happen. > Only enter in MemoEdit, move to the end, exit (this memorize positions and > text) and re-enter, move the cursor up and I've the mark of the previous > cursor. [Remember that this is few tested, I've fix recently MemoEdit's page > movements, before it broke the screen.] It also does with some GETs to close > the window press Esc but it depends on what is presented. > It's not easy to reproduce in a self-contained example because it uses a > sophisticated system of windows (Save/RestScreen) written in pure Clipper. I > can not simplify it, i.e. presents the positions and scroll bars > operational. > > I like to use EX_LAYERED but I think is an issue in Windows XP on Caret > refresher and the buffer. > Only found this but is very different .- > > http://www.drdobbs.com/windows/184416498;jsessionid=CEWLVUUVAYSALQE1GHOSKH4ATMY32JVN?pgno=1 > http://plainwonders.blogspot.com/2007/09/whats-wrong-with-wsexlayered-and.html > > I've made many tests to fix it but the only effective solution is invalidate > the previous cursor position. > Attached the changes to the current code, I would commit. > Please note that most positions (normal use) are already invalidated and is > affected by SetCursor(0). > The code is activated with EX_LAYERED by define initial. > > Best regards and Happy Easter to All! > -- > Xavi > > > _______________________________________________ > Harbour mailing list (attachment size limit: 40KB) > Harbour@harbour-project.org > http://lists.harbour-project.org/mailman/listinfo/harbour > > _______________________________________________ Harbour mailing list (attachment size limit: 40KB) Harbour@harbour-project.org http://lists.harbour-project.org/mailman/listinfo/harbour
