Re: crash in 1.0.3a fortran compilation

[Mumit Khan <khan at xraylith dot wisc dot edu>]

On Thu, 23 Jul 1998, Greg Lindahl wrote:

> That's fine. Is there any reason that there shouldn't be a second
> test-suite which compiles gobs of code in an attempt to ferret out
> goofs. There are plenty of people like me with tons of machines who'd
> be happy to run the "fat suite". There's a lot more to testing than
> compiling small, easily understandable programs.

Possibly because these codes are large and most people want to download
them along with the compiler.

This is what I do whenever I feel ambitious or am bored after getting a
new "stable" snapshot:
  
  C++:
    - build our old C++ code. Not too template heavy, but *large*.
    - build my new'ish code, that's very template heavy (lots of STL,
      uses expression templates ala Blitz). It doesn't work with egcs
      1.1 snapshots. This is mostly to see if I'm close to saving lots
      of $$ in reducing the number EDG-based compilers that we have.
    - build Octave (GNU matlab like program), which exercises both C++
      and FORTRAN.
   
  FORTRAN:
    - build our legacy FORTRAN code. Most of this is rather old crufty
      code.
    - build netlib stuff -- ie., what is not already in Octave
      distribution.

Since these are all in my source repository, it's pretty easy for me.

Unfortunately, by the time I'm done running these tests, I've usually
found something else to do, and never get around to reporting any 
actual new bugs that I may find ;-)

Mumit