This is the mail archive of the gcc@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: Ada bootstrap failure due to loop optimization and builtin_stack_alloc


    because rewriting every use of the variable sized array via indirection
    makes debugging impossible, basically.  

It doesn't make it "impossible", it just means that there's a level of
indirection there that the programmer won't expect to see in the debugger.

If you use a REFERENCE_TYPE instead of a POINTER_TYPE, it's actually a
little cleaner.

Ada does this in cases where there's no choice (e.g., a global of
variable size) and you can indeed debug it, but see the indirection.

    But yes, as noted further in the thread, this seems like a best solution
    regardless.

I agree with that: it's not that common a case to justify adding a
large level of complication to GIMPLE to deal with it.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]