Re: [rfa] struct dictionary

[David Carlton <carlton@math.stanford.edu>]

On Tue, 10 Jun 2003 14:43:11 -0400, Elena Zannoni <ezannoni@redhat.com> said:

>> >> -static struct block *new_block (int);
>> >> +static struct block *new_block (int function);
>> 
>> > can this take an enum parameter with some meaningful names?
>> 
>> Well, I'm just using it as a boolean, so it seems to me that doing an
>> enum would be a little funny.  How about I rename the argument to
>> function_p, or is_function, or something like that?  But if you'd
>> prefer an enum, I could do something like
>> 
>> enum block_type { FUNCTION, NON_FUNCTION };
>> 
>> instead.

> For me the real problem is to understand the code that does the
> call. I don't know what 1 or 0 means at that point w/o going looking
> for the function definition. And I lose my train of thoughts.

That makes sense: an enum it is, then.  (And in a year or two, we can
switch GDB over to C++, and have an actual boolean type, but never
mind that.)

>> >> @@ -493,6 +495,11 @@ dump_symtab (struct objfile *objfile, st
>> >> fprintf_filtered (outfile, " under ");
>> >> gdb_print_host_address (BLOCK_SUPERBLOCK (b), outfile);
>> >> }
>> >> +#if 0
>> >> +	  /* NOTE: carlton/2003-04-28: If we really want to be able to
>> >> +	     print out something here, we'll need to add an extra
>> >> +	     dictionary method just for that purpose.  */
>> >> +
>> 
>> > Hmmm, I think we should. We don't want to change gdb's behavior.
>> 
>> Well, it's for a maint command, so we can really do whatever we want
>> here.  But if you think that's an important part of the info, I'll add
>> the extra method.

> thanks yes.

Will do, then.

>> I'll make those changes and commit it tomorrow afternoon (unless
>> exam grading takes longer than I expect), if nobody complains; and
>> I'll be spending all day Thursday doing GDB stuff as well, so I'll be
>> available if I accidentally break anything...

> ok, great.  BTW, i have started setting up the same test framework
> that Michael Chastain uses for his tests. I have done a preliminary
> run, but I screwed up the combinations of gcc and binutils to test
> with, so I'll get back to that tomorrow, and I should be able to help
> with testing for gdb6.

Cool!  I miss his test runs.

David Carlton
carlton@math.stanford.edu