Re: [PATCH 0/1] Integrate GNU poke in GDB

[Tom Tromey <tom@tromey.com>]

>>>>> "Jose" == Jose E Marchesi via Gdb-patches <gdb-patches@sourceware.org> writes:

Jose> poke STR
Jose> poke-add-type EXPR
Jose> poke-add-types REGEXP
Jose> poke-dump-types

>> It's maybe more gdb-ish to make one command and use subcommands.

Jose> What would be the gdb-ish way:

Jose> a) poke STR
Jose>    poke add-type EXPR
Jose>    poke add-types REGEXP
Jose>    poke dump-types

Jose> or

Jose> b) poke STR
Jose>    poke add type EXPR
Jose>    poke add types REGEXP
Jose>    poke dump types

Jose> Because a) will be problematic: `add-types' can be a valid Poke
Jose> expression if both `add' and `types' are defined as variables.

Jose> Doing b) would be ok I think.

Either is fine, though (b) is more of a pain to implement.  If the plain
prefix command is ambiguous, you can have it work like "set", where "set X"
evaluates X, but "set var" is a subcommand to avoid the ambiguous case.

Jose> But I am not really looking forward to write Python bindings for libpoke
Jose> (or Python for anything for that matter) and even if I could recruit
Jose> someone to do that work, the stuff would need to be maintained ... poke
Jose> depending on Python, supporting future Python versions and what not,
Jose> argh no no no no :)

Just to be clear, gdb often has a high burden to get a patch in.

For example, do you have a gdb copyright assignment in place?  We can't
accept your patch until that's done.

Jose> PS: If GDB supported writing plugins in C or C++ like bash does, I
Jose>     wouldn't blink twice before adopting your suggestion!

I would be opposed to this.  This would require stabilizing some gdb
API, which I think would be difficult and probably counter-productive.

Tom