-
Notifications
You must be signed in to change notification settings - Fork 10
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Segfaults #64
Comments
Can you show me some a minimal path to reproduce the problem ? (ie. what you type to compile, then what you type when you start the engine, in terms of UCI commands, until the segfault). |
Hello Lucas, When I compile demolito on a 32 bit Linux laptop there are no crashes. [antonio@manjaro coredump]$ coredumpctl gdb 15062 Stack trace of thread 19269: Stack trace of thread 19272: Stack trace of thread 19271: Stack trace of thread 19270: Stack trace of thread 19268: Stack trace of thread 15062: GNU gdb (GDB) 7.12
Kind regards,
On Thu, Nov 17, 2016 at 12:11 AM, lucasart notifications@github.com wrote:
|
Thanks. Unfortunately "I run the engine with Cutechess GUI" is not enough to reproduce. Can you do the same experiment with cutechess-cli instead. Please run as follows:
where $1 is the path to a debug compile of demolito, and $2 any other engine that won't crash. What I need is the last UCI communication (displayed on screen) and the file Thanks! |
Hello Lucas, Demolito didn't crash in 100 games with your command. 157944 >Demolito-rev-247(0): position fen And this is the message in errors.txt And this is the backtrace. (gdb) bt Kind regards,
On Sat, Nov 19, 2016 at 12:37 AM, lucasart notifications@github.com wrote:
|
Thanks. So the minimal reproduction should be:
Can you confirm that, just by starting demolito in command line, and pasting the 5 commands one by one, you can reproduce the crash ? I cannot reproduce on my machine. |
I can confirm that by pasting the 5 commands one by one demolito crashes On Sun, Nov 20, 2016 at 2:19 AM, lucasart notifications@github.com wrote:
|
Thanks. I can reproduce now. The trick (for me) was to compile with
So it's completely unrelated to the hash size, or to chess960. The bug should be easy to find now. I'll give it a crack when I have a bit of time. |
The bug is caused by an illegal move in the PV, which triggers an assert, as PV moves have to be played in
The last move is illegal. Seems to be linked to abruptly stopping the qsearch at |
fixed |
Hi Lucas,
I've compiled many commits of demolito in Linux 64 bits and I get nearly all games a segfault.
I've tried with GCC and Clang.
I used the make.sh script and also tried to compile on my own, always same result.
On 32 bits hardware no segfaults.
Does this engine only run on 32 bits?
The text was updated successfully, but these errors were encountered: