Custom Query (13985 matches)

Debian 10 x86_64 ScummVM 2.0.0 Maniac Mansion C64/English

The graphics are corrupted when the alien appears in the telescope, see screenshot.

To reproduce: load the attached save game, and use really powerful telescope.

This may be due to a corruption in the disk image, but apparently disk images from different download sources are identical in this regard: in MANIAC2.D64, at offset 0x29fff (i.e. within the pic map of Room 35), there is a 0x4B value, which should rather be 0x82.

(Btw. this would also cause a buffer overrun by 24 bytes in ScummVM, if it weren't for the extra space allocated for GdiV1::_V1.picMap. To make it safer, there should be x<size checks in GdiV1::decodeV1Gfx() before all three occurrences of writing to dst[x], which would log a warning about incorrect encoding and return from the function if x is too large.)

I am testing this on a raspberry pi 4 with the bleeding edge scummvm which I compiled myself. I'm not sure if the same issue happens on other platforms, please check.

At the beginning of the game, when you are next to the dumpster. If you attempt to use your hand or leg on the dumpster there is a 50% chance of a crash. On the next screen (in front of the bar), when you use your eyes on the window of the bar, the game crashes on all attempts. Doing actions pretty much always end up in a crash.

Running scummvm with gdb gave me this result :

0x0007d0c0 in Scumm::IMuseDigital::switchToNextRegion (track=0x2bbd6d0, this=0x2bfa640) at engines/scumm/imuse_digi/dimuse.cpp:582

582 bool isJumpToStart = (soundDesc->jump[jumpId].dest == soundDesc->marker[2].pos && !scumm_stricmp(soundDesc->marker[2].ptr, "start"));

Program terminated with signal SIGSEGV, Segmentation fault.

I used git blame to check which commit actually made changes around that line and found 3 possible commits : 434d1f37f9398b6a7a1b08cd5464a1d5ecccdc41 8be297ea4fcb6dc56b28e7484913f149e6ff796d 5ea8e62e5315c940556b27d6fbf91a4e278e4e6c

The problem was solved after reverting these 3 commits. Do note however that I don't think all of these 3 commits caused the regression.

I got this a bunch of times (three or four) lately, when playing CoMI (first screen puzzles) on my Android 10 device (smartphone Mi Note 9 Pro). I am unsure if the issue is specific to Android 10 device or to the CoMI game.

Putting the device to "sleep" and then, after a while, resuming the app, leads to a crash to the Home screen. Oddly, it does not happen every time, and it seems that some (significant) time must pass with the device being in sleep/suspend mode, before the issue occurs. Maybe after a while in sleep mode, the device releases some resource (or completely shuts down the app?).

As a possibly related bug, I also got a corrupted save game (again in the first screen) but failed to record the issue with it (when trying to load it, it displayed a message but I forgot to store it somehow). I haven't been able to reproduce this one though. This is with disabled Cloud Saving (so it's not a side-effect of that module).

I will be debugging this one, since I am not sure what the cause is as of yet.

This is with the 2.3.0git version (recent daily build).