[rescue] Python > 3.6 on Solaris 10

[Joshua D. Boyd]

> On Oct 17, 2021, at 11:34 AM, Doug McIntyre <merlyn at geeks.org> wrote:
>
> On Sun, Oct 17, 2021 at 01:07:49PM +0100, Mike Meredith wrote:
>> On Sat, 16 Oct 2021 17:39:28 +0000, Jonathan Chapman wrote:
>>> I don't know...I think I'd take Perl's legacy cruft and peculiarities
>>> over the pain that was the Python 2 -> 3 transition :P
>>
>> Disclaimer: I do write simple scripts in Python these days - mostly
>> because I prefer to steal other people's code than write my own, and no
>> I'm not a fan of the white space blocking.
>
> I do use python for sysadmin scripting now-a-days vs. perl in the past..
>
>> But I don't recall Python 2->3 being that bad in comparison to
>> Perl4->5. Perhaps that was because I was a clueless newbie for one and
>> a clueless old fart for the other?
>
> No, it wasn't bad. The primary issue was just the shear number of module
> (and other) programmers that denied that python3 was coming until it was at
a
> point that python2 was deprecated before almost anyone finally started
looking into
> the few simple edits to make their code python3 compatible. 80% of
> Python3 compatibility was the change of print from a statement to a
function.
> But yeah, people sat on it for so long moaning about it rather than doing
it,
> and then they did it, and finally there was critical mass to drag the rest.
>
> Perl 4->5 was far worse. I also remember earlier in the 5.x train of
> having to fix data access code as they deprecated older constructs and
> broke my code several times.
> Won't even mention that perl6 is a totally different language now, new name
and project..
> Or all the incompatible changes in PHP 5.3 -> 5.4 -> 5.6 -> 7.0

Ibm not saying that Perl wasnbt worse, but my recall is that while Python
3 was released in 2008, the early recommendation was that you either go 100%
Python3 (obviously not practical right away), or you maintain Python2 code and
generate Python3 builds from 2to3, which just wasnbt a good experience at
all.

It wasnbt until several years later that the idea of predominately trying to
write code compatible with 2 or 3 came to dominate, and that was a mess if you
needed to support python 2.5 or python 3.0 in addition to other versions.  If
you could support 2.6, 2.7, and 3.2[1], then it was finally practical to do a
single code base, and that wasnbt until 2011.

Thus (again, to my recall at least), by the time Python 3 support wasnbt
that bad (outside large legacy or C extension packages), there were already
3-4 years of bad taste about Python3 support and a lot of people who were
determined to dig their heels in about it.