[RFA/commit/Ada] Wrong value printed by info locals for dynamic object.

[Tom Tromey]

>>>>> "Joel" == Joel Brobecker <brobecker@adacore.com> writes:

>> return val_print (value_type (val), value_contents_all (val),
>> value_embedded_offset (val), value_address (val),
>> stream, recurse, options, language);

>> Short of creating a language-specific value_contents routine, I think
>> that this was the best/easiest compromise... Thoughts on that?

I suppose a new language method is also possible.
That may be overkill for a single specialization.

Joel> Well, stupid me (again!): I just realized that I checked in a couple
Joel> of patches for which I said yesterday that I'd wait a few days. Sorry
Joel> about that.  Hopefully this is not a problem, but I'm happy to revert
Joel> and/or adjust based on feedback. My entire week is dedicated to the
Joel> FSF tree, so I have time...

IMO, don't revert, just address comments if any.
I think reverting should generally be reserved for extreme scenarios,
say when some fundamental problem is discovered.

Tom