g00r00,

I just upgraded to the latest alpha as you can see in my tear line. I'm
trying to play some MPL games and I'm having some severe issues. I'm playing my latest double-up! and reversi. When I exit the games, the app hangs for a second, then it drops carrier on me. It doesn't happen with each mpl, but
it's consistant with these two. There's no entry in the node.log nor in the mis.log files.

The only time I see things like this happen is if arrays inadvertently get overrun or if the app tries to access element zero. I've tried a number of other my apps and only the ones that don't use multidimentional arrays are the ones that don't bork.

It's possible that I'm not doing enough error checking and am violating
memory space so I'll check my code again. But they were working just fine before the upgrade, and now they are consistently failing. But they only
fail after the game is played once or if I try to quit the game after
playing. If I play once, then try to play a second time in the same session, it will fail.

--- Mystic BBS v1.12 A37 2017/12/28 (Linux/64)
* Origin: Cyberia BBS | cyberia.darktech.org | San Jose, CA (21:1/120)

On 12/28/17, Gryphon said the following...

It's possible that I'm not doing enough error checking and am violating memory space so I'll check my code again. But they were working just
fine before the upgrade, and now they are consistently failing.

The code undergoing rearrangement and the prealphas being built with a lot
of debugging enabled will make them more sensitive to bugs like buffer overflows, so that's certainly a possibility. Unless something actively stops you, you can overflow an array and nothing bad will happen as long as the affected memory belongs to you and other parts of your code never need to
read or execute what used to be there. No signal or exception thrown, no
foul. If something changes the way the memory is mapped out significantly between builds though, it may end up in an arrangement that isn't as okay
with the situation as it was previously. Or the debugger itself could
introduce watchdogs or other enforcement, etc...

--- Mystic BBS v1.12 A36 2017/12/03 (Linux/64)
* Origin: Subcarrier BBS (21:2/140)

The only time I see things like this happen is if arrays inadvertently
get overrun or if the app tries to access element zero. I've tried a number of other my apps and only the ones that don't use
multidimentional arrays are the ones that don't bork.

This was a great observation, thank you for pointing me in the right direction. MPL can be a gigantic pain to debug.

I made some changes for 2D and 3D arrays and I tested it with Double Up which seems to run now regardless of how I compile Mystic. I worry this change
could break something else though, but it doesn't appear to have broken my
BBS which has some custom MPL.

--- Mystic BBS v1.12 A37 2017/12/29 (Windows/32)
* Origin: Sector 7 [Mystic BBS WHQ] (21:1/108)

On 12/29/17, g00r00 said the following...

The only time I see things like this happen is if arrays inadvertentl get overrun or if the app tries to access element zero. I've tried a number of other my apps and only the ones that don't use multidimentional arrays are the ones that don't bork.

This was a great observation, thank you for pointing me in the right direction. MPL can be a gigantic pain to debug.

I made some changes for 2D and 3D arrays and I tested it with Double Up which seems to run now regardless of how I compile Mystic. I worry this change could break something else though, but it doesn't appear to have broken my BBS which has some custom MPL.

FWIW, I noticed that the MPY apps I wrote are also having this issue. Both
the IBOL and SBBL apps are now just dropping carrier on me when I run them. They were running just fine before the upgrade. The only difference is that IBOL gives an error in the logs:

2017.12.30 18:01:28 Python execute: ibol.mpy
2017.12.30 18:01:28 An error has occured (216)
2017.12.30 18:01:28 Shutting down

I don't know if it's related but its definately a persistant issue.

--- Mystic BBS v1.12 A37 2017/12/28 (Linux/64)
* Origin: Cyberia BBS | cyberia.darktech.org | San Jose, CA (21:1/120)