This is the mail archive of the
gdb-patches@sources.redhat.com
mailing list for the GDB project.
Re: PATCH: Remove unnecessary zero-initializations
Michael Snyder wrote:
Felix Lee wrote:
Daniel Jacobowitz <drow@mvista.com>:
Certainly it's a style issue. However, it's an awkward style issue and
anyone implementing a target should be looking over the complete list
of methods anyway.
Of course. It's not about writing the target in the first
place, it's about reading it and maintaining it later. Say,
a year from now, someone adds a new method but doesn't spend
all the effort necessary to make all N targets work
correctly with the change, which is reasonable since it's
not sensible to insist that everyone be familiar with the
issues of all N targets before doing any work.
Absence of an initializer is a simple indication that
someone should look at it and make sure it's ok.
No. New target methods need to be optional, else someone
will be required to go back and implement them for all old targets.
The fact that they don't need to be initialized to zero is intentional,
so that a new method does not require any modification to old targets.
If there is no initialization, and indeed no mention of the new
method at all in an old target module, you can safely conclude
that the module doesn't implement the method.
That's right. The goal was to make it follow an object-oriented style,
where when you instantiate something, you only fill in the elements
that don't have a default value.
While there's something to be said for the opposite style, on
the theory that it means you have to consider whether to implement
each method, it breaks down when the superclass (aka target
vector struct) changes, because you then have to go and modify
all the instances, including those for targets that you don't have
access to (and *nobody* has h/w for all the targets supported by
GDB!), so you're in the undesirable situation of modifying code
that you can't test.
Stan