www.digitalmars.com         C & C++   DMDScript  

digitalmars.D.learn - Unexpected OPTLINK termination

reply "cal" <callumenator gmail.com> writes:
I just ran into (I think) this bug 
http://d.puremagic.com/issues/show_bug.cgi?id=8377 using dmd 
2.060. Seeing that there was a fix in git, I cloned the latest 
git dmd, built, and dropped the .exe into my existing dmd 2.060 
installation directory.

I still get this unexpected optlink termination (although the 
location of the error is different - EIP=00424D76 - don't know if 
that is significant or not).

Thing is, I only hit the error when I compile a Debug target. In 
release mode there is no error.

I don't think I can narrow down this problem easily, I suspect 
the problem is in a 11k lines module of template heavy code 
generated by Pegged. I am just wondering if anyone who saw this 
problem pre the git commit also has trouble with debug builds, or 
if this is a different bug.

Cheers
cal
Aug 19 2012
next sibling parent Philippe Sigaud <philippe.sigaud gmail.com> writes:
On Mon, Aug 20, 2012 at 7:33 AM, cal <callumenator gmail.com> wrote:

 I don't think I can narrow down this problem easily, I suspect the problem
 is in a 11k lines module of template heavy code generated by Pegged. I am
 just wondering if anyone who saw this problem pre the git commit also has
 trouble with debug builds, or if this is a different bug.

I also had lots of trouble with Pegged and optlink since 2.060. Are you using rdmd? Because I have no linking problems with dmd, only rdmd, strangely enough. Maybe it's due to rdmd reusing cached information that should be updated more aggressively.
Aug 20 2012
prev sibling parent "cal" <callumenator gmail.com> writes:
On Monday, 20 August 2012 at 13:50:27 UTC, Philippe Sigaud wrote:
 On Mon, Aug 20, 2012 at 7:33 AM, cal <callumenator gmail.com> 
 wrote:

 I don't think I can narrow down this problem easily, I suspect 
 the problem
 is in a 11k lines module of template heavy code generated by 
 Pegged. I am
 just wondering if anyone who saw this problem pre the git 
 commit also has
 trouble with debug builds, or if this is a different bug.

I also had lots of trouble with Pegged and optlink since 2.060. Are you using rdmd? Because I have no linking problems with dmd, only rdmd, strangely enough. Maybe it's due to rdmd reusing cached information that should be updated more aggressively.

I tried both dmd and rdmd, both give the same result, optlink crash in debug build. I'm going to give dustmite a go, see if I can reduce it at all.
Aug 20 2012