Play! - Development Log

Play! is a PlayStation2 emulator for Windows, macOS, UNIX, Android, iOS & web browser platforms. For more information about this project please visit the "About" section of this site.

Fetching compatibility status...

Fix previously working programs
Posted on 2006-08-30 00:00:00

Quake and Slave of the VU seem to be sending too much data through the GIF when uploading some textures and that's what was causing the problems. So, I made some changes that'll make any excess data to be ignored.

Fixed a small bug in the implementation of the FILEIO module that caused the VM thread to be destroyed when reading a file beyond EOF (an event that generated an exception that was unhandled). With this fixed, Space Invaders works properly again.

I readded support for the V4-5 VIF compression mode that's used in the 3rd scene in Slave of the VU. The scene itself still isn't working properly, so I'll be working on making it work like it was before.

Fix previously working programs
Posted on 2006-08-28 00:00:00

Completed the changes in PADMAN so it can adapt itself to which version is requested.

Unfortunately, there's some more things I need to fix before I can really call this development complete:

- Some programs, like Quake and Slave of the VU, seem to be causing some problems when uploading image data to the GS.
- Slave of the VU is a bit screwed up, but this is probably because I commented out the implementation of the VIF V4-5 decompressor.
- Atelier Iris will crash during the intro sequence. This is due to the use of an unsupported IPU command which I'll have to implement.
- Add this "force flipping at v-sync" flag for the OpenGL GS handler.

That's pretty much some of the major bugs I can think about for the moment. So, I'll be working on fixing these next.

Fix previously working programs
Posted on 2006-08-26 00:00:00

Found the right place to call the function that flips the buffers : each time something is written to DISPFBx the buffers are flipped. This method works with the 3 games and it even seems to work with the demos that use double buffering. We'll need to add an option to force the flipping of the buffers every v-sync though, because some demos don't display anything at all with this new method.

Also fixed the crashing problem in Atelier Iris. Nothing of the threading stuff was touched... A small memory bound enforcement in the GS memory address handlers did the trick.

Fixed a bug with the threading stuff that was causing some demos not to start properly.

The last thing I want to do is to make the PADMAN IOP module able to select which pad data structure to use depending on the version of the module requested. I've started working on it, and I hope to have it done tomorrow.

Fix previously working programs
Posted on 2006-08-24 00:00:00

Completed the DSxxV fixing task.

I found out about a little bug with Atelier Iris, probably related to the new thread scheduler code, that makes the game crash during the movie playing phase. That'll mean I'll have to do some more tweaks to make it work properly.

But first, I really want to complete the flipping moment detection thing, so I'll be working on that next.

Fix previously working programs
Posted on 2006-08-21 00:00:00

Started working on the DSxxV fixing task. This is actually an excuse to rewrite the implementation of these instructions, so it might take some days to complete. It'll give me some time to think about the buffer flipping problem.

Fix previously working programs
Posted on 2006-08-19 00:00:00

Started checking about how to detect the "moment of back buffer flipping", but I don't really know where to start. Some demos don't use double buffering at all, so there must be a way to keep the default flipping moment which is at each vblank.

I think I'll try to use some kind of state machine to decide of which moment to use, but I'm not quite sure if it'll be efficient. I'll start implementing this idea next.

Fix previously working programs
Posted on 2006-08-17 00:00:00

I seemed to be able to tweak the threading stuff enough to make all 3 games work. So, I don't think there's any need to touch that anymore.

Next thing I wanna do is to have some kind of automatic screen buffer flipping moment detection, since the scheme used in the 3 games are different. It'll also have to work with demos.

Fix previously working programs
Posted on 2006-08-15 00:00:00

I was able to make the new thread scheduling system to work with Ys I & II. Now, all that remains is to make it work correctly with Castlevania.

Fix previously working programs
Posted on 2006-08-13 00:00:00

Scrapped the old status bar and replaced it with the standard one that comes in Windows. (task)

I still haven't thought of an idea for the thread stuff. I'll keep thinking about it and hope to come up with a solution soon.

Fix previously working programs
Posted on 2006-08-11 00:00:00

I checked why the current thread system wouldn't work with Ys I & II and I remembered that this game was kinda picky about them : if a sub-thread doesn't complete its work before the main thread is woken up, it'll simply get deleted and the game will stay in a state of uselessness... So, maybe a thread shouldn't ever get the control back if there's a higher priority thread still running...

Seems like I'll have to think this over again.

<< Older Log Entries

Newer Log Entries >>