TypeError: must be str, not bytes

[freephile]

meza --version

Traceback (most recent call last):
  File "/usr/bin/meza", line 1269, in <module>
    main(sys.argv[1:])
  File "/usr/bin/meza", line 64, in main
    print( "Meza " + version.strip() )
TypeError: must be str, not bytes

in the code for meza.py

version = subprocess.check_output( ["git", "--git-dir={}/meza/.git".format(install_dir), "describe", "--tags" ] )
commit = subprocess.check_output( ["git", "--git-dir={}/meza/.git".format(install_dir), "rev-parse", "HEAD" ] )
print( "Meza " + format(version.strip()) )
print( "Commit " + format(commit.strip()) )

The version and commit are byte objects. We must use the format method to convert them to string before we can concatenate them to the string values (e.g. 'Meza ')

[freephile]

Initially, I would get the error. After using format, I receive the proper version output (almost):

python3.11 meza/src/scripts/meza.py --version
Traceback (most recent call last):
  File "/opt/meza/src/scripts/meza.py", line 1269, in <module>
    main(sys.argv[1:])
  File "/opt/meza/src/scripts/meza.py", line 64, in main
    print( "Meza " + version.strip() )
           ~~~~~~~~^~~~~~~~~~~~~~~~~
TypeError: can only concatenate str (not "bytes") to str
[root@localhost opt]# python3.11 meza/src/scripts/meza.py --version
Meza b'39.0.0-16-gcd9359f'
Commit b'cd9359f36e816e0f9192c28833bb7eea1d19ddc4'
Mediawiki EZ Admin

It appears that some extraneous 'b' is being output. Is that 'b' for byte?

Edit: YES, it is the Python representation of a byte sequence. The change in handling of bytes vs strings in Python 3 (where it's now more strict) is the cause of this (compatibility) issue.

[freephile]

It doesn't seem to matter whether I use Python 3.6.8 or Python 3.11, I still get the 'b' (byte) output even when trying to format() or use text=True in subprocess.check_output()

Even trying 'f strings' still gives the same output

print( f"Meza {version}" )
print( f"Commit {commit}" )
print( "Mediawiki EZ Admin" )

Meza b'39.0.0-16-gcd9359f\n'
Commit b'cd9359f36e816e0f9192c28833bb7eea1d19ddc4\n'
Mediawiki EZ Admin

[freephile]

Recap

Here's what you get with an unmodified meza.py on Python 3.6.8

[root@localhost opt]# python3 ./meza/src/scripts/meza.py --version
Traceback (most recent call last):
  File "./meza/src/scripts/meza.py", line 1269, in <module>
    main(sys.argv[1:])
  File "./meza/src/scripts/meza.py", line 64, in main
    print( "Meza " + version.strip() )
TypeError: must be str, not bytes

And Python 3.11 gives pretty much the same error:

[root@localhost opt]# python3.11 ./meza/src/scripts/meza.py --version
Traceback (most recent call last):
  File "/opt/./meza/src/scripts/meza.py", line 1269, in <module>
    main(sys.argv[1:])
  File "/opt/./meza/src/scripts/meza.py", line 64, in main
    print( "Meza " + version.strip() )
           ~~~~~~~~^~~~~~~~~~~~~~~~~
TypeError: can only concatenate str (not "bytes") to str

The best I've come up with so far is to wrap a format call in the concatenation:

print( "Meza " + format(version.strip()) )
print( "Commit " + format(commit.strip()) )
print( "Mediawiki EZ Admin" )
print( "" )
sys.exit(0)

produces (for both Python 3.6.8 and 3.11)

Meza b'39.0.0-16-gcd9359f'
Commit b'cd9359f36e816e0f9192c28833bb7eea1d19ddc4'
Mediawiki EZ Admin

In Python 3.11, we can add the text=True argument to subprocessl.check_output and it works! without even using format() in the concatenation:

import subprocess
version = subprocess.check_output( ["git", "--git-dir={}/meza/.git".format(install_dir), "describe", "--tags" ], text=True )
commit = subprocess.check_output( ["git", "--git-dir={}/meza/.git".format(install_dir), "rev-parse", "HEAD" ], text=True )
print( "Meza " + version.strip() )
print( "Commit " + commit.strip() )
print( "Mediawiki EZ Admin" )
print( "" )
output = subprocess.check_output(["python", "--version"], text=True)
print(output)

(note how the system Python version is 3.6.8, but we run the code with python3.11 ./meza/src/scripts/meza.py --version)

Meza 39.0.0-16-gcd9359f
Commit cd9359f36e816e0f9192c28833bb7eea1d19ddc4
Mediawiki EZ Admin

Python 3.6.8

But that code produces an error in Python 3.6.8

python3 ./meza/src/scripts/meza.py --version
Traceback (most recent call last):
  File "./meza/src/scripts/meza.py", line 1271, in <module>
    main(sys.argv[1:])
  File "./meza/src/scripts/meza.py", line 62, in main
    version = subprocess.check_output( ["git", "--git-dir={}/meza/.git".format(install_dir), "describe", "--tags" ], text=True )
  File "/usr/lib64/python3.6/subprocess.py", line 356, in check_output
    **kwargs).stdout
  File "/usr/lib64/python3.6/subprocess.py", line 423, in run
    with Popen(*popenargs, **kwargs) as process:
TypeError: __init__() got an unexpected keyword argument 'text'

[freephile]

So, the open question is

1. whether Python must be upgraded;
2. or is there a better approach to formatting and concatenating bytes and strings;
3. or is the 'ugly' output good enough?.

Upgrading Python is discussed here

[freephile]

Responses:

1. Python does not need to be upgraded... at least not to solve this problem. (The discussion about Upgrading Python addresses the big picture.)
2. There is a "better approach" to formatting and concatenating bytes and strings in Python 3.x. Python 3.x is more strict about mixing types and so in this case we need to use decode(). Documentation for bytes.decode in Python 3.6.x
   See best practices for Ansible and Python
3. No, incorrect representation of the desired data is unacceptable.

[rjt]

Progress:)