Bug 6353 – No stacktraces on x86_64

This program void main() { assert(0); } prints this on x86_64: core.exception.AssertError@w(3): Assertion failure ---------------- ---------------- _All_ stack traces look like that on x86_64. For this one, if I were to compile with -m32, I'd get this: core.exception.AssertError@w(3): Assertion failure ---------------- ./w(onAssertError+0x2e) [0x806111e] ./w(_d_assertm+0x16) [0x805f096] ./w() [0x805ca76] ./w(_Dmain+0xd) [0x805ca61] ./w(_D2rt6dmain24mainUiPPaZi7runMainMFZv+0x1a) [0x805f53e] ./w(_D2rt6dmain24mainUiPPaZi7tryExecMFMDFZvZv+0x20) [0x805f1d8] ./w(_D2rt6dmain24mainUiPPaZi6runAllMFZv+0x32) [0x805f582] ./w(_D2rt6dmain24mainUiPPaZi7tryExecMFMDFZvZv+0x20) [0x805f1d8] ./w(main+0x94) [0x805f184] /usr/lib32/libc.so.6(__libc_start_main+0xf3) [0xf75aa233] ---------------- So, this is purely a 64-bit issue. But the lack of stacktraces on x86_64 is a definite impedement to debugging, and it really needs to get fixed.

Something that's never worked isn't a regression. I agree it's a major feature gap for x86-64 though.

I'm pretty darn sure that this was working before - hence why I changed it to regression. I'd have to go try older versions of dmd to be 100% sure though. But I have no problem with it being labeled critical instead of regression even if it is a regression. It's making my life _much_ harder when it comes to debugging.

This works now and has for a couple of releases, I believe.