This is the mail archive of the mailing list for the GDB project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: set multiple-symbol ask/cancel not working

> Tom> (gdb) b pck.adb:normal_Menu
> Tom> [0] cancel
> Tom> [1] all
> Tom> [2] pck.adb:pck.normal_menu:4
> Tom> [3] pck.adb:pck.normal_menu:8
> I am not at all sure about doing this.  It means that the breakpoint
> will stop working if the line number changes.

This is what is already happening today. One of the future improvements
we discussed was transitioning to a canonical format that used the
argument types, much like C++ does.  But this is a big project on its
own, and we're not sure yet whether the debug info is sufficient for
this, and how much code needs to be implemented in order to match
the relevant semantic rules of the Ada language.

> Maybe this is ok, since it is what the user asked for in
> multiple-symbols=ask mode -- not the default, so he essentially asked
> twice.

Just to be certain, I assume that if the user selects "all", then
we still get one multi-location breakpoint?

> Another option would be to only apply this treatment to Ada.

I am cool with that. Might be a little ugly, but I don't think
we should penalize the other languages just because we're not ready
to do so in Ada.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]