[#480] Remove Python2 compatibility from code base

[d-w-moore]

With this change, we remove all code meant explicitly to deal with the possibility of the PRC's running under Python 2. Especially the awkward (and sometimes voluminous) parts that interfere with a ready understanding of the code base as a whole.

[trel]

looking good so far. please assert here how testing goes.

[d-w-moore]

looking good so far. please assert here how testing goes.

have already done some piecemeal testing with good results. will run a full test later.

[korydraughn]

Excellent! Keep pushing.

[d-w-moore]

There may still be a few points I haven't thought of, but presently the code is exclusively Python3-compatible (python2 -c"import irods" fails !!) and passes all tests in the suite.

[d-w-moore]

.... now with minimal squashing

[korydraughn]

Amazing work!

We'll start the review process.

[korydraughn]

Looks good to me.

Let's get another set of eyes on this before squashing/merging.

Good stuff!

[korydraughn]

There may still be a few points I haven't thought of, but presently the code is exclusively Python3-compatible (python2 -c"import irods" fails !!) and passes all tests in the suite.

How much does python2 -c"import irods" actually catch? Do you know?

I'm just curious.

[d-w-moore]

There may still be a few points I haven't thought of, but presently the code is exclusively Python3-compatible (python2 -c"import irods" fails !!) and passes all tests in the suite.

How much does python2 -c"import irods" actually catch? Do you know?

I'm just curious.

Here's what happens if you try importing the PRC at tip of this branch currently:

Python 2.7.18 (default, Jan 31 2024, 16:23:13) 
[GCC 9.4.0] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> import irods
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/home/daniel/ppy3/prc/irods/__init__.py", line 85, in <module>
    from . import client_configuration
  File "/home/daniel/ppy3/prc/irods/client_configuration/__init__.py", line 40
    class ConnectionsProperties(iRODSConfiguration, metaclass = iRODSConfigAliasMetaclass):
                                                              ^
SyntaxError: invalid syntax
>>> irods
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
NameError: name 'irods' is not defined
>>> 

As a result of the errors evoked, not the irods namespace, nor module, nor any submodules will be imported (as you might well intuit by that last command). Which means all code in the PRC is unavailable to the interpreter under Python2

[d-w-moore]

Amazing work!

We'll start the review process.

Great!...
Btw, I didn't mean to put the" laugh" emoji there. Joy is more what I wanted to express.

[trel]

Can we... be kinder to anyone who might stumble into this by accident...

like...

import sys

if sys.version_info[0] < 3:
    raise RuntimeError("This library is only supported on Python 3 and above.")

But also check the minor version while we're there...

[d-w-moore]

Can we... be kinder to anyone who might stumble into this by accident...

like...

import sys

if sys.version_info[0] < 3:
    raise RuntimeError("This library is only supported on Python 3 and above.")

But also check the minor version while we're there...

Good point. We'll put that into irods/__init__.py . It will work fine even if sub-namespaces are (naively) preferentially imported, because in order to import those, Python needs to import the parent modules first. (In other words, import irods.manager.data_object_manager has to import irods as the first step.

[trel]

nice and clean.

[FifthPotato]

Seems pretty complete to me. Couldn't find anything wrong, so I think it looks good!

[trel]

codacy has some opinions... but then... are we ready to squash?

[alanking]

I second a squash

[alanking]

Created #644 for f-strings in logging statements.

Created #645 for naked asserts.

Created #646 for import ... as issue.

"Possibly unused variable" in a test seems... not worth the review/test time, but I say we address that in this PR if we address it at all (irods/test/collection_test.py line 171).

[trel]

please grep for six and remove all instances found for this PR.

[d-w-moore]

please grep for six and remove all instances found for this PR.

Done. All gone!

$ grep -rwI six|grep -v '\.git/'
irods/test/message_test.py:        ivp.svalue = ['five', 'six']
irods/test/message_test.py:        expected = "<GenQueryInp_PI><maxRows>4</maxRows><continueInx>3</continueInx><partialStartIndex>2</partialStartIndex><options>1</options><KeyValPair_PI><ssLen>2</ssLen><keyWord>one</keyWord><keyWord>two</keyWord><svalue>three</svalue><svalue>four</svalue></KeyValPair_PI><InxIvalPair_PI><iiLen>2</iiLen><inx>4</inx><inx>5</inx><ivalue>1</ivalue><ivalue>2</ivalue></InxIvalPair_PI><InxValPair_PI><isLen>2</isLen><inx>1</inx><inx>2</inx><svalue>five</svalue><svalue>six</svalue></InxValPair_PI></GenQueryInp_PI>"
irods/test/message_test.py:        self.assertEqual(gq2.InxValPair_PI.svalue, ["five", "six"])

[trel]

alright - squash this. let's get it in.

[d-w-moore]

"Possibly unused variable" in a test seems... not worth the review/test time, but I say we address that in this PR if we address it at all (irods/test/collection_test.py line 171).

Is used in a format string.

[d-w-moore]

Tests pass , awaiting approval

[alanking]

Yes, please squash.

[d-w-moore]

alright - squash this. let's get it in.

ok - sorry didn't see this til now

[d-w-moore]

squash done

[alanking]

You said tests are passing, everything checks out... I say # it.

[d-w-moore]

#'d

[trel]

a note that we've also marked this as ready for Python 3.13.