[129391] trunk/dports/perl/p5-cgi-speedycgi/Portfile

[Mojca Miklavec]

On Fri, Dec 12, 2014 at 10:38 PM, David Evans wrote:
> On 12/12/14 11:54 AM, Ryan Schmidt wrote:
>>> On Dec 12, 2014, at 8:49 AM, mojca at macports.org wrote:
>>>
>>> @@ -1,10 +1,13 @@
>>>   # -*- coding: utf-8; mode: tcl; tab-width: 4; indent-tabs-mode: nil;
>>> c-basic-offset: 4 -*- vim:fenc=utf-8:ft=tcl:et:sw=4:ts=4:sts=4
>>>   # $Id$
>>>   +# Port is broken:
>>> +# - http://trac.macports.org/ticket/33479
>>> +
>>>   PortSystem          1.0
>>>   PortGroup           perl5 1.0
>>>   -perl5.branches      5.10 5.12 5.14 5.16
>>> +perl5.branches      5.10 5.12 5.14 5.16 5.18 5.20
>>
>> I don't understand this change. If we're adding perl branches without
>> testing that they work, then let's just do that for all the remaining perl
>> module ports all at once, as was suggested in months past. I had wanted to
>> avoid that, because I like our standing policy of verifying that a port
>> builds before committing a change. If we're not going to do that here, then
>> there's no reason to do each port separately.
>
> I agree with Ryan.  While it would be easy to do a bulk change on the
> remaining ports that do not support 5.18 5.20, if a known issue exists with
> a particular port then it should be fixed before making the update.

I'm sorry for this change, but I have another question.

Assuming that all the working ports have been taken care of: what
would you do with completely broken ports (that is: ports where
p5.16-foo is just as broken as p5.20-foo)? Obviously we are not
removing the existing broken variants. (I removed broken variants
5.8-5.14 in p5-wx for example, but there 5.16 worked just fine.)

Mojca