Re: [PATCHv2 0/5] New Python hook for missing debug information

[Andrew Burgess <aburgess@redhat.com>]

Tom Tromey <tom@tromey.com> writes:

>>>>>> "Andrew" == Andrew Burgess <aburgess@redhat.com> writes:
>
> Andrew> This series adds a new Python hook to GDB, this hook allows a user to
> Andrew> register a handler object which will be called whenever GDB fails to
> Andrew> locate any debug information for an objfile it is loading.  These
> Andrew> handler objects can (potentially) perform host or project specific
> Andrew> actions to help GDB locate the missing debug information.
>
> Do you have some specific scenario in mind for this?  I'm wondering when
> this would be useful, and when Objfile.add_separate_debug_file wouldn't
> be sufficient.

Fedora and RHEL GDB has, for many years now, carried a local patch that
will suggest to the user a suitable RPM to install should GDB encounter
some missing debug information.

To some degree, maybe even mostly, this functionality has been replaced
by debuginfod, but there are still some situations where we're finding
this useful.

So, I'm trying to replace the RH local (C++ patch) with a Python
extension.

For RPM suggestion, Python only really need to be notified when GDB
encounters some missing debug info, and initially I did implement this
as a new event.

But when I looked at it, I didn't think this added much complexity to
core GDB, and given we don't really like to change the Python API, I
figure I might as well go with the fuller featured choice.

>
> Anyway the patches all seem fine to me.
> Approved-By: Tom Tromey <tom@tromey.com>

Thanks,
Andrew