Re: [PATCH v2 1/2] compile: set debug compile: Display GCC driver filename

[Phil Muldoon <pmuldoon@redhat.com>]

On 27/04/15 21:36, Jan Kratochvil wrote:
> On Mon, 27 Apr 2015 20:53:46 +0200, Pedro Alves wrote:
>> "wrong" is subjective.  This split is a conscious design decision,
>> that has advantages like insulating the debugger from compiler ICEs.
>
> I haven't looked it up now but IIRC that is not a "conscious design decision"
> but just a workaround of buggy GCC which cannot recover/re-run from
> compilation errors in the same instance.  The ICE resistance was only
> a side-effect.
>

It's not buggy. We are pushing GCC in new ways. It's a side effect of
change. GCC was designed, and has only run, AFAIK, in this way with
the recent changes of libcc1. The state of GCC has never needed to be
preserved and/or reset (say with GDB and cleanups) as if it
encountered a problem it just exited. Change is good.

But this is not the GCC list. And we probably should not further
exhaust GDB'ers patience with GCC internals. If a plan is set, then
lets just go with it and see if it works?

Otherwise we can just explore until one works ;)

Cheers

Phil