Too many compilers needed for C++11

[Mojca Miklavec]

On 7 December 2017 at 10:28, Joshua Root wrote:
> On 2017-12-7 17:39 , Ryan Schmidt wrote:
>>
>> On Dec 7, 2017, at 00:37, Joshua Root wrote:
>>
>>> On 2017-12-7 17:29 , Ryan Schmidt wrote:
>>>> (I don't know why MacPorts base considers this a warning, rather than an
>>>> immediate-exit error.)
>>> Because you wouldn't want things like portindex or 'port info' to fail
>>> because of it.
>>
>> Surely there's a way to arrange it so that port info and portindex succeed but the compile is stopped as early as possible. For example, set the compiler to /usr/bin/false, or a MacPorts-internal binary that prints a helpful error message.
>
> Surely. What's wanted is much the same as the proposed support for
> indicating that ports don't build on particular platforms. (You want to
> realise it isn't going to work before you install any dependencies.)
>
> When the current warning was added, it was more of an afterthought, like
> the default case of a switch statement that should theoretically cover
> all possibilities. I don't think anybody expected that so many ports
> would end up in this invalid state.

Just a small additional piece of information.

Nowadays many ports claim
    Warning: All compilers are either blacklisted or unavailable;
defaulting to first fallback option
but do install nevertheless at the end. I would say that this is a bug
in either the portfiles or portgroups that's worth fixing anyway, but
I just wanted to say that if you make this an error, suddenly a huge
number of ports that can easily be built will simply stop working.

One of annoying things is that I'm getting this warning when I run
"port livecheck maintainer:..." and I have absolutely no clue where it
comes from.

Mojca