PortIndex2MySQL run failure on Sunday 2016-03-06 at 07:00:01

[Mojca Miklavec]

On 6 March 2016 at 19:18, Ryan Schmidt wrote:
>> On Mar 6, 2016, at 12:15 PM, Mojca Miklavec wrote:
>>
>> Are you able to help me out to come up with a proper patch or should
>> we simply replace
>>
>>        if {[active_variants gtk3 quartz ""]} {
>>            default_variants-append +quartz
>>        } else {
>>            default_variants-append +x11
>>        }
>>
>> by
>>
>>        default_variants-append +x11
>>
>> ?
>>
>> We would probably need to add
>> an additional "if gtk is active" before "if {[active_variants gtk3
>> quartz ""]}", right?
>
> I haven't looked at your code but the usual thing to write would be:
>
>        if {[!variant_isset quartz]} {
>            default_variants-append +x11
>        }
>
> https://trac.macports.org/wiki/PortfileRecipes#default_variants

OK, I usually write *all* the variants in the list even though it
doesn't make any difference and I agree that it doesn't make much
sense to check if both variants are undefined.

But the major reason for this weird construct that broke the indexing
is that the variants between wxgtk-3.0 and gtk3 need to match with
each other. If the user already has gtk3 +quartz active, trying to
install wxgtk-3.0 +x11 will fail anyway, so it's usually better to
automatically switch to wxgtk-3.0 +quartz (if we ignore for a moment
that it's not quite ready yet).

Mojca