py25-numpy keeps wanting to install g95 on upgrade

[Ryan Schmidt]

On Apr 23, 2008, at 9:16 AM, Daniel Horwood wrote:

> When I initially installed py25-numpy I used its +gcc43 variant  
> (which removes py25-numpy's g95 dependency), and it installed just  
> fine. Today, however, when I tried to upgrade gnome-menus, I had to  
> Ctrl-C when I saw that macports had started downloading files for  
> the g95 port. It took me a while to trace down, but it's definitely  
> something to do with py25-numpy.
>
> The receipt in /opt/local/var/macports/receipts/py25-numpy/1.0.4_7 
> +gcc43/  correctly lists the variant as +gcc43, so I'm not sure  
> what's going wrong.
>
> Here are the bottom few lines of output from '$ sudo port -dv  
> upgrade py25-numpy':
> ===
> DEBUG: No need to upgrade! openssl 0.9.8g_0 >= openssl 0.9.8g_0
> DEBUG: No need to upgrade! py25-hashlib 2.5.2_0 >= py25-hashlib  
> 2.5.2_0
> DEBUG: Found port in file:///opt/local/var/macports/sources/ 
> rsync.macports.org/release/ports/lang/g95
> DEBUG: g95 is *not* installed by MacPorts
> DEBUG: Changing to port directory: /opt/local/var/macports/sources/ 
> rsync.macports.org/release/ports/lang/g95
> DEBUG: Requested variant darwin is not provided by port g95.
> DEBUG: Requested variant i386 is not provided by port g95.
> DEBUG: Requested variant macosx is not provided by port g95.
> DEBUG: Executing variant darwin_9 provides darwin_9
> DEBUG: Searching for dependency: g95
> DEBUG: Didn't find receipt, going to depspec regex for: g95
> DEBUG: Executing org.macports.main (g95)
> --->  Fetching g95
> DEBUG: Executing proc-pre-org.macports.fetch-fetch-0
> DEBUG: Executing org.macports.fetch (g95)
> --->  gcc-core-4.0.3.tar.gz doesn't seem to exist in /opt/local/var/ 
> macports/distfiles/g95
> --->  Attempting to fetch gcc-core-4.0.3.tar.gz from ftp:// 
> gcc.gnu.org/pub/gcc/releases/gcc-4.0.3/
> ===
>
> Steps to reproduce:
> 1. sudo port install py25-numpy +gcc43
> 2. sudo port upgrade py25-numpy
>
> My system: macports 1.6; OSX 10.5.2 (Intel)
>
> I'll file a bug once I confirm this isn't user error, but I have a  
> feeling that it's the funky way that the py25-numpy sets +g95 as a  
> default variant within a conditional block.

I don't consider the way in which py25-numpy sets the default variant  
in a conditional to be funky. It's a strategy I employ in my ports  
too. I recommend it.

Th´re may be a MacPorts base bug here. What you describe sounds  
vaguely familiar. I can't test it out right now though as my CPU is  
busy with other things.