Well, I guess it could be something weird about single-core machines… It’s a pity because the code that’s wrapped by the spinlock in Timer is very small, and really shouldn’t need anything as complicated as a critical section, but yes, I guess I should revert it…
Will you please post here when this is reverted? Also, does anyone here have a machine for testing this on? My machines (Core2Duos) don’t display the bugs that seem related to this issue.
Thanks,
Sean Costello
Any progress on this? I just got pinged by a customer, so I am hoping that I can use the tip soon to create a build of my plugin that doesn’t have these graphic issues.
Thanks,
Sean Costello
I did it a couple of days ago. Sorry, but asking me to post when I commit something is a bit futile, as I will inevitably forget or not have time. You can follow the sourceforge RSS if you need to watch out for a commit.
I’m all about striving against futility here at Valhalla DSP! Seriously, plugin development feels like Sisyphus pushing that boulder up that hill. I’m sure you can relate - you’re the one having to deal with the low-level platform stuff.
Sean
P.S. More existential ranting about plugins: Cave Paintings, Wabi-Sabi, and Plugins – The Halls of Valhalla
I’m neck-deep in doo-doo with this project that I’ve been working on for 2+ years now. It seems every time I get something implemented, I discover subtle problems due to the interaction with other parts of the concurrent system (I’m up to 14 threads now). Sometimes it’s really difficult to keep my mental game up because it just feels overwhelming.
Update: I built my plugins with the latest tip (using VS2010) and sent them to a customer that was having the drawing issues on the Pentium IV. He reports back that things work perfectly now. Woo-hoo!
Sean Costello
I wish I understood why the spinlocks would cause a problem though! I hate it when I see stuff that doesn’t make sense like that…