Re: Effects of Fink / MacPorts [was: PCRE library requirement]

[Jarno Rajahalme]

On Feb 1, 2011, at 19:43 , ext Lukas Reichlin wrote:

> 
> On 01.02.2011, at 17:06, bpabbott wrote:
> 
>> On Feb 01, 2011, at 10:32 AM, Jarno Rajahalme <address@hidden> wrote:
>> 
>>> 
>>> On Feb 1, 2011, at 15:08 , ext Ben Abbott wrote:
>>>> 
>>>> I've been considering switching to MacPorts since it doesn't allow mixing 
>>>> of compiler versions (all dependencies are built with the same compiler). 
>>>> I don't think this was always the case. My impression is this changed with 
>>>> the release of Snow Leopard (just guessing really).
>>>> 
>>> 
>>> I don't think this is the case. Portfiles can not depend on a specific 
>>> variant of a dependency. You have any source for this?
>>> 
>>> On naming of compilers: I like the way macports uses it's own names (e.g. 
>>> gfortran-mp-4.5), so that by default the system's own compiler is used, 
>>> unless a specific other compiler is called for explicitly.
>>> 
>>> On the "need to port": I think of macports more like a "package management 
>>> system". The portfile contains all the info needed to download, verify, 
>>> build (e.g., if specific configure options are needed), and install the 
>>> package, with version control and management of dependencies. The few 
>>> patches where needed are a nice bonus.
>>> 
>>> Jarno
>> 
>> Looking at the port-file for octave-devel, I noticed the gcc variants. Look 
>> at lines 139-146.
>> 
>>      http://trac.macports.org/browser/trunk/dports/math/octave-devel/Portfile
>> 
>> Am I misunderstanding what those mean? Also does MacPorts supply a gfortran 
>> patched to work correctly with Apple's gcc-4.2.1, or did you use the AT&T 
>> version (you've likely mentioned this, but I've forgotten).

The purpose of the different gcc variants is to specify to toolchain to build 
the "port" in question. IMO it has no bearing on how the dependencies are 
built. There has been some discussion in past on syncing some of the math 
portfiles to use the same default gcc variant.

> 
> I am the guy who sent the updated Portfile to MacPorts, so take it with a 
> grain of salt. I don't know what lines 133 - 137 are good for.
> 

Apparently to negate the effect of a user wanting to override the default 
variant, i.e. to make sure one of the variants is set.

  Jarno