CompuServe Thread

Autodesk Multimedia Forum · *MM Developers

#Feedback sought for…

7 messages in this thread

#98357From: Jonas Ruikis [ADESK]Apr 22, 1994 4:17 PM

I was asked to post the following thread and ask for feedback. "The point was brought up that some IPAS developers may be counting on the PharLap memory problem, that global variables are not re-initialized unless 3DS is restarted. Could you post a query on CompuServe as to whether fixing the memory releasing problem would break any existing IPAS routines?" Comments?

#98370From: Danny MercurioApr 22, 1994 6:09 PM

While I think that all memory used within an IPAS routine should be released after termination, I believe that a number of us are probably taking advantage of the fact that this is not currently the case. I personally use several structure arrays in an AXP which I count on remaining available as long as the user is in 3DS. Maybe there could be a kind of compromise like a free() function for globals that would give the programmer a choice whether or not to release memory.

#98404From: Jonas Ruikis [ADESK]Apr 23, 1994 9:46 AM

Thanks Danny. Anyone else?

#98625From: SyndesisApr 25, 1994 10:40 AM

If you used an APP_DATA, then you could keep this global info in a file, and preserve it even between runs of 3DS. Or for that matter, you could store it in a file of your own design.

#98754From: Yost GroupApr 26, 1994 10:30 AM

FYI, there's a "limitation" in the shipping version of 3DSr3 that inhibits storing app_data chunks larger than 64K. We've come up with a program that modifies 3ds.exe to take out that limitation, called appdata.exe (of course). This'll ship with out IPAS routines (which almost ALL use app_data very heavily), and I'll upload appdata.exe here to dl17 for any other developers that want to use it. I'll put it up here today. – G

#98469From: Don HansonApr 24, 1994 3:45 AM

Currently I use temp files in between frames for my AXPs, but I was definately planning on doing all that in memory next time. I had thought that the Exp_Dont_Unload command would work even if the memory retaining "bug" was fixed. Will this be the case?

#98569From: Douglas HoltApr 25, 1994 1:10 AM

I would also like to see info on Don's q. I use EXP_DONT_UNLOAD with just about everything I use… PS: My new address is 74401,3004, there should be a message in Software Dev Kits about this from my new account. I don't have access to this section yet from my new account.