svn commit: r351587 - in head: . Mk Mk/Scripts

[Gerald Pfeifer]

On Mon, 28 Apr 2014, Gerald Pfeifer wrote:
>   ===> Warning: Test was done with PREFIX != LOCALBASE
>   ===> Warning: The port may not be properly installing into PREFIX
> 
> ...feels a bit aggressive.  In my experience, if that != case works,
> it is more likely that the == case works than the other way around.
> 
>>      - There are likely still some false-positives. I fixed as many as I could
>>        find. Please let me know of others.
> 
> I have no idea what to do about
> 
>   ===> Checking for directories owned by MTREEs
>   Error: Owned by MTREE: @dirrmtry info
>   Error: Owned by MTREE: @unexec rmdir "/scratch2/tmp/gerald/prefix/gcc%%SUFFIX%%" >/dev/null 2>&1 || :
> 
> that I now see for lang/gcc49.  
> 
> "@dirrmtry info" is not part of pkg-plist.  That other lines is, yet
> how can this be owned by MTREE?

Any suggestions?  Checks like this being added, as hard errors,
make my ports work harder -- and more risky! -- since I now have 
to ignore errors to be able to succeed tests and commit.

(And I was wrong originally: the @unexec rmdir is not part of 
pkg-plist either.)

Gerald