2436 – Unexpected OPTLINK termination EIP = 00425303 with /co

optlink fails to link an object file compiled with dmd 2.020 with -g or -gc switch.
CL: link test,,,user32+kernel32/co/noi;

Created attachment 277 [details] 
offending obj

Probably a duplicate of #1439 or/and #424. I suggest voting for #424 as it already has a couple of votes :)

I tried linking this obj file, and it linked without complaint, and runs printing "Test".
I suspect the problem may be with multithreading on a multicore system. My system is single core. If you're running multicore, is there a way to use only one core and try it that way?

It fails on a single core. Are you sure you specified /co for the linker? It fails only with the CodeView stuff turned on. Please try to link it like this: dmd test.obj -g

This is not the same as bug 1439 or bug 424, which have both been fixed.
I can reproduce it by:
> link test.obj /co
(of course it doesn't link, since the runtime library is the wrong version;
but it still crashes).

I can duplicate the problem, and I know where in optlink it is failing, but I don't know why.

This test case was created in the days when I was overly enthusiastic about templates, ctfe and stuff. Now I don't have any code that triggers this error. Lowering the severity to normal as nobody else seems to have suffered from this particular bug.

Fixed Optlink 8.00.9
http://ftp.digitalmars.com/link.8.00.9.zip 

*** Issue 4808 has been marked as a duplicate of this issue. ***

http://gool.googlecode.com/files/optlink80011_bug.rar
8.00.11 crash .

Please open a new bug unless you have a particularly good reason to believe the crash you're seeing is directly tied to the same cause as this bug. Given that this bug is closed, it's almost a certainty that your issue will be overlooked / lost.

(In reply to comment #10)
> http://gool.googlecode.com/files/optlink80011_bug.rar
> 8.00.11 crash .
Resubmitted as bugzilla 5670