https://bugs.winehq.org/show_bug.cgi?id=40701
Bug ID: 40701 Summary: COCO simulator: Aborting of painting upon concurrent WM_PAINT messages Product: Wine Version: 1.9.9 Hardware: x86 OS: Linux Status: UNCONFIRMED Severity: normal Priority: P3 Component: -unknown Assignee: wine-bugs@winehq.org Reporter: kooijman@clarkson.edu Distribution: ---
Under wine v1.9.9 while using COCO v3.1.0.1(from http://cocosimulator.org/) I encountered a problem with the painting of window elements during a simulation. Normally this would cause parts of the flowsheets to light up in a different color while each part is being solved but now this is no longer the case. The problem is not related to any particular flowsheet being solved, you can take any example from the samples they have online at http://cocosimulator.org/index_sample.html
Note this bug occurs after the most recent bug update from COCO, titled "prevent problems due to re-entrancy in painting methods caused by STA threads in 3rd party modules", see http://cocosimulator.org/index_changes.html. I contacted author jasper van baten jasper@amsterchem.com. He commented that "COM is flawed in the matter of thread safety in case there is somewhere any STA compartment which means that WM_PAINT messages can come in concurrently. If this is detected, paint is aborted and a private message is posted (which is received non-concurrently in the main windowing thread) and the window update proceeds from there. I am guessing wine has a bug that makes this message does not arrive." Jasper recommended me to go back to using the stable version 1.8 for there the problem does not occur according to him. I fear that going back would just postpone the issue for me and hence I wanted to report it as a bug.
Harry