Skip site navigation (1)Skip section navigation (2)
Date:      Sun, 01 Apr 2012 13:05:22 +0200
From:      Matthias Andree <mandree@FreeBSD.org>
To:        freebsd-ports@freebsd.org
Subject:   Re: rawtherapee 4.0.7 build failure on 9-STABLE amd64
Message-ID:  <4F783672.9040107@FreeBSD.org>
In-Reply-To: <4F76C3B7.5010305@yandex.ru>
References:  <4F76ABE6.206@FreeBSD.org> <4F76C3B7.5010305@yandex.ru>

next in thread | previous in thread | raw e-mail | index | archive | help
Am 31.03.2012 10:43, schrieb Ruslan Mahmatkhanov:
> Matthias Andree wrote on 31.03.2012 11:01:
>> Ruslan,
>>
>> I cannot build rawtherapee 4.0.7 on 9-STABLE amd64,
>> re-running it after failure with "make MAKE_JOBS_UNSAFE=yes" yields:
> [...]
>> math.h line 239 declares "double log2(double);", so I presume that
>> patch-rtengine_improcfun.h proves harmful here; the "#ifndef log2(x)" is
>> bogus if log2 isn't declared as macro (and is actually what breaks the
>> compile), and the macro is #define log2(x) ((log(x)/log(2)).
>>
>> It might be more useful to do this instead:
>>
>> static const  double lnOf2 = log(2);
>> static inline double log2(double x) { return log(x)/lnOf2; }
>>
>> or possibly checking math.h for log2() before adding the patch.
>>
>> I'm afraid I won't have the time to do that now; could you?
>>
>> Thanks.
>>
>> Best,
>> Matthias
>>
>>
> 
> Should be fixed now. That was a last minute change when I realize that
> it doesn't build on 8.2. I didn't rechecked it on 9/10 after the fix was
> applied. Apologies and thanks for the hands up.

Thanks, works for me now.  I'd only used MAKE_JOBS_UNSAFE to be sure
there weren't any artifacts, and I get sequential logging.

> Btw, it builds fine to me on -current amd64 with MAKE_JOBS_SAFE=yes,
> should I add this knob to Makefile?

Assuming it uses cmake as a proven and established build utility, I'd
say 'go for it'.

Thanks for the prompt fix!

Best,
Matthias



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?4F783672.9040107>