Re: [Boost-build] Process Jam Log compile failure on release branch

From: Rene Rivera (grafikrobot_at_[hidden])
Date: 2008-07-03 23:24:12

K. Noel Belcourt wrote:
> 
> On Jul 1, 2008, at 11:48 AM, Rene Rivera wrote:
> 
>> K. Noel Belcourt wrote:
>>> I've restarted the release branch tests for darwin (intel and ppc).  The
>>> PPC tests ran okay but I got this error trying to compile
>>> process_jam_log.cpp on the Intel-based machine.  If I just remove the
>>> -isysroot option, things compile okay.  Not sure who added this option
>>> for Intel Mac, nor when it was added.
>>
>> That would be me, I think. But the strange part is that it shouldn't be
>> there by default. It should add those only when one asks for a specific
>> macosx-version. And I just checked, and I don't get the option when
>> doing just "bjam". Do you have the bjam invocation that it used for pjl?
> 
> Any idea on this Rene?

I'm trying to figure it out.

> It seems to me that sdk-version is always being 
> defined in the rule init-available-sdk-versions in darwin.jam.

It shouldn't be if it doesn't find an SDK. And if it does, it should be 
defined as a conditional flag that only appears when one asks for a 
specific version.

But are you saying that the 'sdk-version' variable is always defined? 
Perhaps to an empty string?

> Seems 
> like you shouldn't search for SDKs in this rule unless the user asks you 
> to, rather than always searching for them.

That's the goal. And what I thought I implemented. And how it works in 
my setup.

-- 
-- Grafik - Don't Assume Anything
-- Redshift Software, Inc. - http://redshift-software.com
-- rrivera/acm.org (msn) - grafik/redshift-software.com
-- 102708583/icq - grafikrobot/aim,yahoo,skype,efnet,gmail