I am working with an app that uses rust lib. I've modified that lib that it causes crash.
I know what I did and it is quite easy for me to fix that problem without seeing any output. But if I was working with somebody's else program that crashed without any output I would have really hard time to find the bug.
How to force the output?
I cannot start debugging the app. Debugging app starts only after attaching it to a running process. Process crashes during the startup phase so I cannot attach to it.
This lib is written in rust. I was using it by including it in cargo.toml in dependencies section. I've modified it by making changes to its source code.
I know how to start debugger. Unfortunately debugging apps that use the cursive library is not that straightforward and you need to attach to them really.
What kind of output are you referring to? The panic message and backtrace printed to the standard error stream? And how is this hypothetical program suppressing the output?
i wanna say running ulimit -c unlimited prior to running your program if you run on linux, to enable core dump save to file for after crash debugging... but not sure if it fits your scenario... (or i understood the problem correctly)
Ah, that's because it's using ncurses, which hijacks the terminal as part of its functionality, and can't fix it up before Rust attempts to print the panic info. @bart2, can you redirect the stderr stream to a file, as suggested in a relevant GitHub issue?