100x -O0 Compile Time Regression {3.2,3.3} -> {3.4,3.5}

[Mike Stump]

On Thursday, February 12, 2004, at 01:50 PM, John Vickers wrote:
> We've got some generated C++ which compiles in about 10s (on a P4 
> 2800) at -O0 with gcc-3.2 or 3.3.2, but takes about 1000s with 3.4.0 
> or 3.5.0.

:-(

> Any suggestions as to where to start digging ?  I can supply 
> preprocessed source to anyone interested.

Well, if you want to help track it down, you can binary search 
compilers using the reghunt script from the web site.  Once you narrow 
it down to the patch(es) that cause the problems, we can ask that 
person to look into it.

What the script essentially does is akin to cvs update -D '64 days ago' 
. && make && test, and then loop, adding or subtracting 32, then 16, 
then 8, then 4, then 2, then 1, and presto, the patch that caused it.

If you don't mind filing a bug report, please do, attach the file to 
the bug report.  Let me know which one it is, and I will help ensure 
that the problem is fixed.

Thanks for reporting this.