In message <xofzopgxo5y.fsf@blubb.pdc.kth.se>, Johan Danielsson writes:
>John Hawkinson <jhawk@mit.edu> writes:
>
>> 2)	There should be a "step over" command (like :e in adb) that sets
>> 	a breakpoint at the next instruction if the current instruction
>> 	is a call (or jump?).
>
>Yes please!
>
>Any hope of a gdb-like interface (for those of us that never
>understood the beauty of adb)?

Can you state your request in the form of functional requirements?

gdb is quite a bit more verbose than adb or ddb, and personally I
think it's interface and style are overly bloated for a kernel
debugger. Perhaps more importantly, I'm certainly not going to do the
work to make adb look like gdb, or to try to "port" gdb into the
kernel, or anything like that. Remember, of course, that we do have
kgdb if you are truly attached to this sort of thing.

When I think of the "beauty of adb", it's that you can save a lot of
typing[*] ;-) I don't think that ddb's interface is particularly cryptic,
though it is different from gdb and adb. What would you like to see
changed?

In terms of user-friendlyness, I suppose we could add additional
helptext so that "help" told you what commands did rather than simply
what their names were. But presumably the ddb manpage is adequate. I
think it's certainly far more verbose than the adb documentation
(which "implies" or "leaves to the user to figure out" many useful
things, like that $c takes an address ("0x12345$c") or that :c takes a
count of breakpoints to ignore). It's certainly more concise
than the gdb documentation.

I would presume that if you're unfamiliar with ddb and are going to
be using it, printing out a copy of the manpage should be sufficient
to address the sorts of things that having on-line help would. That
doesn't completely address it, but it certainly seems to make it
a non-priority and perhaps a non-issue.

Can you list what you want in the debugger?

--jhawk

[*] ok, I guess I take a child-like delight in the crypticness of
adb at times...