Re: [patch/testcase] fix gdb.objc/basicclass.exp

[Doug Evans <dje at transmeta dot com>]

Kei Sakamoto writes:
 > > On Fri, Jul 04, 2003 at 12:09:36PM +0900, Kei Sakamoto wrote:
 > > > The attached patch fixes gdb.objc.basicclass.exp.
 > > >
 > > > The current test always fail with sim because of libobjc's error.
 > > >
 > > > (gdb) PASS: gdb.objc/basicclass.exp: continue until method breakpoint
 > > > run
 > > > The program being debugged has been started already.
 > > > Start it from the beginning? (y or n) y
 > > > Starting program:
 > > > /project/m32r-tools/insight-20030703/build/gdb/testsuite/gdb.objc/basicclass
 > > > Virtual memory exhausted
 > > >
 > > > *** EXIT code 4242
 > > >
 > > > Program exited with code 0157255.
 > > > [Switching to process 0]
 > > > (gdb) FAIL: gdb.objc/basicclass.exp: resetting breakpoints when rerunning
 > > >
 > > > I assume this is not libobjc's bug. Please tell me if my
 > > > assumption is wrong.
 > > >
 > > > Anyway, I inserted load command before run command. I checked
 > > > this patch with m32r-elf, arm-elf, and sh-elf.
 > >
 > > You missed the patch - would you mind posting it?
 > 
 > I'm sorry about that. This is the patch.

I'm curious.  Why is this now necessary?
Way back when run did an implicit load, among other things
(at least for sims I remember).

One can certainly debate what semantics "run" should have.
I like "run" behaving the same across targets (unix, sim, whatever).

"run" might not make much sense to target boards
(including sims emulating target boards).
In such cases I'd argue we shouldn't be using it, rather than
trying to make it work.