On 2013-12-04 13:36, Michael Meeks wrote:
Sure - the question really is how best to deal with the mess we have currently. It's not entirely clear that the above
scenario is even achievable with the functionality we've exposed to eg. scripting / Java users today =) someone needs
to sit down and carefully work through every threading use-case, and use of the SolarMutex across all the code. Then
they need to unwind the UNO appartment model (which was originally intended to be some sort of solution here), and
finally all the platform toolkit's thread safety / affinity & functionality in this area - which is really not a
trivial task. When we have all of that understood - at least in one brain, we can (perhaps) work out what to do for
the best =) ATB, Michael.
That would be ideal, for sure.
But perhaps there is a way to get there that does not require
- genius-level IQ
- years of LO experience
- tons of time
?
Something like
(*) Change the UNO/scripting processing model to do something like
while (Message m = readMessage() != null) {
invokeOnMainThread(void f() {
processMessage(m);
});
}
(*) add a configure debug option in SolarMutex that asserts if something
touches it from outside the main thread
(*) gradually fix the places that assert until we get to a point where we can
make the option always on in debug mode.
Disclaimer: http://www.peralex.com/disclaimer.html
_______________________________________________
LibreOffice mailing list
LibreOffice@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/libreoffice
