Emulator

May be fixed with r1119. Try with that one ;p

Will do asap Sonic, let me compile that.

In r1125 it still hangs in both DX and OGL.

Also in OGL the game hangs 99% of the time just when closing the inventory, when it
tries to restore the game after using the FB.

what's the error msg? unhandled exception?
Can you copy/paste the msgbox error here please.
Could be issue 287 related.

I you have VS2008 enter in debug when you have that crash. Post here the asm (or code
line) where the breakpoint assertion occur AND the registers value (in "registers"
docked window).

I said hang, not crash, thus no error message. Open the savegame and equip the
Pictograph, if u can..., and take a shot, it'll hang if it didn't while trying to equip.

it Still Not Fixed in R1166
Take a One Shot..

Same for rev svn 1180. It is still hang after you take a shot.

hmm... maybe zelda's render-to-texture doesn't work properly....

It seems like returning from a rtt state is what's not working properly in OGL, or it
could also be interfacing with the memory card upon creation of the "screenshot", or
both.

Maybe once a more generic issue can be disclosed for this, you guys can rename this
to not sound like a game issue.

Federelli, if you can find more instances of this render-to-memory-card in other GC
games it would be great (hint: try thinking of games which use this feature... I have
the African savannah for the Wii...)

Speaking of which, has anyone tested Luigi's Mansion for it? I think it saves a
screenshot of the game upon saving.

By the way, do you see a message "Writing to memory card" before it hangs?

No message, goes to black screen before that, i'm thinking of other games that may
use this meanwhile.

There, SMMB uses this in camera mode in VS / Special Melee, but it appears to work
correctly.

Hm. If the game is sitting in a loop, waiting for the texels from the screenshot to
appear in RAM, this would currently hang in Dolphin, yes. We do not copy the pixels
back from render target into the corresponding location in RAM yet...

If there is an interrupt that waits for that, we should trap it and then copy the
pixels to the RAM

This issue was kinda fixed with the recent attempt to emulate this properly, whoever
the framebuffer that is captured seems to be bigger than the actual one, and u end up
getting the top right corner of the screen in the screenshot u take.

federelli: What Rev are you refering to?
I just tried this with Z:TWW and the pictobox on r2255, this still hangs the whole
thing (with either JIT or JITIL).

I thought we solved that by copying the EFB to RAM

This particular issue can be closed, yes the crash is resolved by copying to RAM,
unless your trying to make the hack not hang the game.

Good.

Issue 1136 has been merged into this issue.