www.digitalmars.com         C & C++   DMDScript  

digitalmars.D.bugs - [Issue 11716] New: ICE on ztc\cgcv.c 2197

reply d-bugmail puremagic.com writes:
https://d.puremagic.com/issues/show_bug.cgi?id=11716

           Summary: ICE on ztc\cgcv.c 2197
           Product: D
           Version: D2
          Platform: x86
        OS/Version: Windows
            Status: NEW
          Severity: normal
          Priority: P2
         Component: DMD
        AssignedTo: nobody puremagic.com
        ReportedBy: babele.dunnit gmail.com



PST ---
Hi,

I am on Windows XP, 32 bits, using DMD 2.064. I just finished a HUGE reduction
of our codebase with DustMite and eventually found that compiling this snippet:

--------------

immutable IPL_DEPTH_8U = 8;
enum Depth { U8 = IPL_DEPTH_8U}

class Crash {
this( Depth ){
}
}

--------------

with "-g -debug" DMD switches will crash DMD in ztc\cgcv.c 2197, i.e. while
generating output debug symbols. Compiling release or just plain versions is
fine, as it is on Linux/OsX and/or 64 bits. The bug seems definitely related to
Windows 32 bits architecture.

Thank you,
Aaron Brancotti
SRLabs Italy

-- 
Configure issuemail: https://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
Dec 10 2013
parent d-bugmail puremagic.com writes:
https://d.puremagic.com/issues/show_bug.cgi?id=11716




PST ---
Sorry, I must correct and add something: the snippet will not "crash" the
compiler, I meant "it will trigger the assertion" at ztc\cgcv.c:2197. But, I
also cloned from GitHub and recompiled DMD with VS10 (as always, on a XP 32-bit
machine) and that DMD/VS10 version WILL crash a bit above, also in cgcv.c, in
"unsigned cv4_typidx(type *t)" function, in "case TYenum", because t is null


Thanks again,
Aaron

-- 
Configure issuemail: https://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
Dec 10 2013