Re: [PATCH][Ada] Add option to gnatlink to fix gnat.dg on some platforms

[Dave Korn <dave dot korn dot cygwin at googlemail dot com>]

Laurent GUERBY wrote:
> On Thu, 2009-05-07 at 12:45 +0100, Dave Korn wrote:
>> Laurent GUERBY wrote:
>>
>>> What we want if a way to turn off the following behaviour:
>>>
>>> "If the compiler executable is different from the default one (gcc or
>>> <prefix>-gcc), then the back-end switches in the ALI file are not used
>>> to compile the binder generated source."
>>   Perhaps it would work if we changed its idea of the "default" compiler to
>> include any found through a -B option?  That would resolve the bootstrapping
>> issue (well, I assume there are -B options passed during building the ada
>> stuff, I haven't actually verified that statement).
> 
> May be I'm missing something but the issue is not bootstrap but check
> here.

  Sorry, I misread the situation; I thought it was the same problem that
caused the system compiler to be used during the build.  In any case, -B
options will still be present during testing.  (And we would want the back-end
switches to be respected yet the system compiler not to be used during
bootstrap too).  But ...

>>   Or even unconditionally disable the behaviour if any -B option is present.

> I used a safe approach of having a small source code change and no
> existing behaviour change. As I said I don't know why the existing
> behaviour is in place and I've not the slightest idea on how it's
> possible to validate a change of existing semantics in released tools.

  ... Yeah, I guess it's not safe without someone being able to tell us why it
does this in the first place.  Shame, it's always nicer if we can get the
compiler to DTRT with the least number of flags and switches.

    cheers,
      DaveK