Is this just the way it is??

[doug at safeport.com]

  On Fri, 24 Jan 2020, Lowell Gilbert wrote:

> Doug Denault <doug at safeport.com> writes:
>
>> So is this just the way things are? Relative to python, I'm 99% sure
>> python36-3.6.9_1 works just as well as python36-3.6.9_3 or
>> python37-3.7.6. I thought (hoped??) the with recent change to
>> package/ports would result them not being this tied to sub-sub version
>> changes.
>>
>> I'm just going for a yes or no. In the past you had a python 3 and 2.7
>> if you needed it. No so now I guess??
>
> You shouldn't need the python36 versions at all.  If you follow
> the UPDATING directions, you *should* end up without them.

ah - /usr/ports/UPDATING  -- thank you

Don't know that I have gone past /usr/src/UPDATING, silly me :(

    For ports users wanting to keep version 3.6 as default,
    add DEFAULT_VERSIONS+= python=3.6 python3=3.6 to make.conf

and (maybe)

    20170602:
      AFFECTS: users of Qt 5 in presence of binutils
              :
     ...lib/libQt5Core.so:(.dynamic+0x27ac8): multiple definition of `__bss_start at Qt_5'
     ...lib/libQt5Core.so:(.dynamic+0x27ac8): first defined here
     ...lib/libQt5Core.so:(.dynamic+0x27ac8): multiple definition of `_edata at Qt_5'
     ...lib/libQt5Core.so:(.dynamic+0x27ac8): first defined here
     ...lib/libQt5Core.so:(.dynamic+0x2b2d0): multiple definition of `_end at Qt_5'

So python for sure - thanks again. With some of the sub-sub version stuff can 
the port/package makers not cover this with the way the Makefile-s define 
requirements? Qt is a pretty basic component. It just went 4-->5 I do not 
remember having this issue during the "4" days.

Sad to say if I knew about /usr/ports/UPDATING I forgot. Call it a senior 
moment. Thanks guys

Doug