Re: Intervention needed?

[Optikos <optikos@verizon.net>]

On Wednesday, April 3, 2019 at 1:15:24 PM UTC-4, Dmitry A. Kazakov wrote:
> On 2019-04-03 18:16, Simon Wright wrote:
> > "Dmitry A. Kazakov" writes:
> > 
> >>> Which feature(s) [that other debuggers have] do you think are
> >>> missing from GDB when debugging Ada?
> >>
> >> Other debuggers work, GDB does not. If you have an Ada project of a
> >> moderate size GDB stops working.
> > 
> > How big is "moderate"?
> 
> In none of my projects GDB works. I never tried to figure out if that is
> related to the number of compilation units or number of library projects 
> involved.
> 
> When you click Debug->Initialize->your-main-program in GPS and debugger 
> does not start you know you reached the point.
> 
> P.S. It never worked reliable in GPS and I bet it never will.

Thank you, Dmitry.  That sufficiently answers my list of questions above.

I am going to assume that bringing GNAT (whose upper annotated-AST stage is written in Ada) up inside GDB would be big enough to observe the defect.

Perhaps the reason that it has never been debugged is the chain of
1) the large-sized Ada executable must be brought up in
2) a(n inner) GDB-under-test session that itself is in
3) another (outer) GDB session.

Taming this chain might require using a few more features than usual (e.g., remote debugging through pseudoterminal; attach outer-GDB to running inner-GDB-under-test via running process in /proc).