Re: [Boost-testing] Error Reasons in Regression Tests

Subject: Re: [Boost-testing] Error Reasons in Regression Tests
From: Belcourt, Kenneth (kbelco_at_[hidden])
Date: 2011-07-26 13:29:50

On Jul 26, 2011, at 11:23 AM, Belcourt, Kenneth wrote:

> Hi Artyom,
> 
> On Jul 21, 2011, at 2:06 PM, Artyom Beilis wrote:
> 
>> BTW is there any chance to get a configuration log on Darwin I can't find
>> a reason why iconv is not found on Darwin?
> 
> I placed some ECHO statements inside the Locale library configure, configure-full, and configure-flags rules.  I attach the output.
> 
> Here's the bjam command line I used (straight from the nightly test log file).
> 
> "/Volumes/scratch/kbelco/boost/tools_bjam/bin.macosxx86_64/bjam"
> "-sBOOST_BUILD_PATH=/Volumes/scratch/kbelco/boost:/Volumes/scratch/kbelco/boost/tools_bb"
> "-sBOOST_ROOT=/Volumes/scratch/kbelco/boost/boost"
> "--boost=/Volumes/scratch/kbelco/boost/boost"
> "--boost-build=/Volumes/scratch/kbelco/boost/tools_bb"
> "--debug-configuration" 
> -l300
> darwin-4.4
> -d+2
> preserve-test-targets=off
> --dump-tests
> -j8
> "--build-dir=/Volumes/scratch/kbelco/boost/results" 
> 
>> Can there be by any chance multiple installations of iconv (multiple iconv.h and/or libiconv.dylib)
> 
> Sure, there's always that chance but I think there's a more fundamental problem though I have no idea what it is.

Oh, I should mention that none of these environment variables are set.

ICONV_PATH, ICU_PATH, ICU_LINK

-- Noel