Re: failure due to compiler?

> resler@liberty.mas.vcu.edu (Dan Resler) writes:
>
> >I am looking for pointers or references to descriptions where software
> >has compiled without error and later failed due to compiler-introduced
> >errors in the resultant program. In other words, the source was deemed
> >to be error free but the compiler botched the code generation. Any
> >help would be appreciated; thanks.


When porting the YAFL compiler to a mainframe environment, we ended up
with a real problem when the C compiler had to compile an overly complex
expression (+10 levels of nesting of C functions, up to 40 lines of
dense, very dense code). Instead of stopping saying that the code was
too complex, it simply generated crap instead. We had to disassemble
the generated code to make sure that something wrong was happening
(In practice, it seemed to mess with the various C functions that had
to be called, some were called more often then mentioned in the source
code, some were not called at all...)


Finally, we had to change the code generation scheme so that it could
split complex expression using temporary variables to solve the problem.


Amazingly, we have not been able to reproduce the problem on a sample.
It did fail repeatedly when dealing with the huge C programs generated by
the YAFL compiler, but it seemed to work well with hand-coded stuff, no
matter how complex we made it.


Unfortunately, we had no time to investigate any further...


Cheers,


--

Return to the comp.compilers page.
Search the comp.compilers archives again.