Since the fix to restore the ability to open V1 mlc files, some Cartwall featurs seem very slow (?).
A looped player with an item duration of 1.9s can take half a second or more to ‘rewind’ and start playing again.
TRIGGERs and Actions on start/stop (all of which start or fadeout other Cartwall players) can result in gaps of up to 2s. (!)
Has anyone else noticed this odd behaviour in b640?
[later]
It’s definitely a CPU issue. I have the Cartwall in a separate window which isn’t open when mAirList starts. Opening the Cartwall window rockets the CPU to at least 60% and (if the 5×5 Cartwall is full) up to 100%!
This would explain the behaviour I originally reported. I also had another strange issue: I have an mlc with a news open sting, a six-minute backing bed, and a news closer sting. The news closer has a TRIGGER which should fadeout the bed. It does do that, but the bed player never stops: it sits in FADE state, so even opening another mlc will not clear it. Since my Cartwall players have only PFL, Next/Previous, and Fadeout buttons, this means restarting mAirList to clear the problem.
Sorry, unable to reproduce it here. And I don’t think this is related to the resolved v1 .mlc file issue, but rather a conincidence - or does it only happen with cart sets loaded from such an old v1 .mlc file?
I’ll go and save an affected mlc in V2 format, then close and re-open mAirList and try again (with Task Manager open so I can watch the CPU … unless you’d prefer me to use Process Explorer and report on the specific processes?
I loaded the ‘special’ build and got a Stop message box with the text:
“The procedure entry point BASS_ChannelSlideAttributes could not be located in the dynamic link library bass.dll.”
I had a quick look within bass.dll and the message is correct: the entry point is named BASS_ChannelSlideAttribute (no ‘s’ on the end).
PS: I also checked some older builds (back to b632) and there is the same ‘leap’ in CPU when I open the Cartwall window. Maybe the Cartwall window in v3 has always ‘eaten’ CPU, and I have just not noticed before? Certainly Saturday was the only time I’d tried doing much with it other than checking it opens. :-[
mairlist.exe is V3.0.10.641 (or .10.637, or .9.632).
Perhaps I should add that in mAirList V2.2.3.552 on the same PC, the Cartwall works flawlessly. At most, CPU goes up to maybe 50% while a Cart Set is loaded, but after that it settles back down to at most about 20%–25%.
I think you did not download the correct “special” version. Because “The procedure entry point BASS_ChannelSlideAttributes could not be located in the dynamic link library bass.dll.” is the error message you get when you try to use an old mAirList.exe (v2.2) with a new bass.dll (2.4.x, from mAirList 3.x).
I will send you an email with the link again.
I suspect that there is a delay in the internal message processing somewhere (which is totally different from the one in v2.2). The debug output from the special version will help to find out why.
It turned out that these effects were caused by a somewhat dated PC (1.5 GHz) with a cheap onboard video card which made the player drawing so slow that there was no CPU load left for processing the internal trigger etc. messages.
I have optimized the code a bit, and now it works smoothly again even on Cad’s PC
Well, one day I just might have some money to buy a newer PC. (Unlikely right now, given that I’m out of work. Unlike Continental Europe, the UK is not out of recession.)
But it was, as Torben says, a worthwhile exercise. We just slashed the minimum PC spec. to run mAirList in half! ;D
PS: In the interests of accuracy, it’s a 1.0 GHz processor, not 1.5 GHz. With a “massive” 1GB of RAM (woo-hoo!).
(Unlikely right now, given that I’m out of work. Unlike Continental Europe, the UK is not out of recession.)
With a “massive” 1GB of RAM (woo-hoo!).