correctness and comprehensiveness of compilers [was: Re: C compiletime]

[Karel Gardas]

On Wed, 18 Jun 2003, Vladimir Makarov wrote:

> Karel Gardas wrote:
> >
> > On Wed, 18 Jun 2003, Scott Robert Ladd wrote:
> >
> > Just one example: all last three releases of Intel C++ compiler for Linux
> > (I mean 6.0, 7.0 and 7.1) miscompile MICO (http://www.mico.org) such that
> > its IDL compiler is either crashing (6.0 + 7.0) or loops infinitely (7.1).
> >
>
>   I don't know about MICO test.

FYI: MICO's IDL compiler is quite good test for testing C++ compilers and
their code generators. The code itself is correct - it's perfectly
compilable by gcc2.95.4, gcc3.2.3, gcc3.3, como4.3.0.1, como4.3.1beta1,
but the same code fails (crash) when compiled by some other buggy
compilers - like Intel's and also some other vendors...

BTW: During the gcc3.0.x development we have submited some bug reports
because of MICO's IDL compiler miscompilation and as a result, the first
gcc 3.x compiler suitable to compile MICO right was gcc3.1 :-)

The code itself is not much interesting - just few no much complicated
templates. The only thing which might be not so common is usage of quite
large virtual inheritance graphs (that's CORBA) and usage of some quite
large machine generated (from IDL compilation) files.

> But it is amazing for me how such respectable company as Intel could
> make a release (7.1) which can not compile correctly some SPECINT2000
> tests (vpr and gcc) with -O2 and -O3 for P4 (their mainstream
> processor).  I found that a week ago using the demo CD which we got on
> the gcc summit.

So, gcc rocks! :-)

Here, it's also good to thank you all, who work on developing/testing gcc
compiler, such that it's sound basis for other free-software development.

Thanks,

Karel
--
Karel Gardas                  kgardas@objectsecurity.com
ObjectSecurity Ltd.           http://www.objectsecurity.com