Stefan Teleman wrote:
>
>
> Garrett D'Amore wrote:
>> On 10/24/08 09:16, Stefan Teleman wrote:
>>>
>>> Garrett D'Amore wrote:
>>>
>>>> Your confusing "build dependencies" with "delivery".
>>> No, i am not. Please read my explanation again. I believe it was
>>> pretty clear the first time around.
>>
>> Not to me it wasn't. The statement you made was that you were
>> delivering binutils *first*, because of the fact that in order to
>> build gcc with --configure, you needed to supply paths to gas, etc.
>>
>> What I said is, that problem is a build dependency problem. You
>> could deliver the final bits together (which IMO is the right thing
>> to do if this case is primarily being done for benefit of gcc.)
>
> No, i cannot.
>
> Since you seem convinced of the opposite, please explain to PSARC
> exactly how is GCC going to find its assembler executable, at
> run-time, after pkgadd, when the path to the assembler executable was
> hardcoded at build time to:
>
> /builds2/steleman/ws/sfwnv-gcc4/proto/root_i386/usr/gcc4/bin/gas
>
> and this path will not exist on any Solaris distro that I can speak of.
That hard coding of the path is a *build time* artifact. If you can't
figure out how to fix that, then you should probably abandon the effort
and let someone else take this task on who can.
(Hint: you're build system can temporarily contain binaries and objects
that are not part of any official distribution yet. I.e. you could
install binutils manually in /usr/bin during the time that you do the
build.)
-- Garrett
