Re: PATCH: gdb/corefile.c (0401 snap on HP-UX).

[Daniel Berlin <dan at cgsoftware dot com>]

jtc@redback.com (J.T. Conklin) writes:

> >>>>> "RDBrown" == RDBrown  <RDBrown@mira.net> writes:
> RDBrown> gdb/corefile.c was failing to compile with a missing
> RDBrown> declaration for `symfile_objfile'. This allows gdb to build,
> RDBrown> but still hangs when starting the child process.
> 
> Our discussions about removing unnecessary header files and a include
> file coventions didn't cover situations like this.  While I expected
> some fallout from removing "unnecessary" headers, I thought they would
> occur because macroes defined in a target's tm-*.h or nm-*.h headers
> would require declarations from one of GDB's headers (like regcache.h,
> etc.).  This is a bit different, because host specific behavior is in
> a common source file.
> 
> 
> I'm not sure that I like this patch, but I'm not sure I can come up
> with a resonable alternative in the short term.  Perhaps the code to
> find the symbol file cooresponding to the core file (this is the bit
> that needs symfile_objfile) should be replaced with a config specific
> hook.  But because it is also wrapped with HPUXHPPA, I'm not opposed
> to checking this in now, and fixing it up later.

Please don't.

The idea is not to *add* things wrapped with HPUXHPPA, rather, remove them.
Pretending they will be fixed up later isn't a good idea. They haven't
been thus far. There is no indication this will change.

If you must add those headers, add them unconditionally. Let's not add
another HPUXHPPA ifdef.


> 
> Would anyone else care to share their thoughts?  If there are no
> objections, I'll get this patch committed soon.
> 
>         --jtc
> 
> -- 
> J.T. Conklin
> RedBack Networks

-- 
I bought a dog the other day...  I named him Stay.  It's fun to
call him...  "Come here, Stay!  Come here, Stay!"  He went
insane.  Now he just ignores me and keeps typing.  He's an East
German Shepherd.