minor code smells in static code analysis

[zacharyburnett]

when running SonarQube scanning on asdf in the bench repository, it found some minor code smells:

duplicates in regex character class

https://plsonarqube.stsci.edu/project/issues?tags=regex&branch=sonarscan&id=spacetelescope-bench&open=AYttO5qgrmNdxaPHqW54

asdf/asdf/_jsonschema/tests/_suite.py

Line 176 in ecb4d4f

delimiters = r"[\W\- ]+"

variable name shadows builtin

https://plsonarqube.stsci.edu/project/issues?tags=pitfall&branch=sonarscan&id=spacetelescope-bench

asdf/asdf/_jsonschema/_legacy_validators.py

Line 138 in ecb4d4f

cmp = "less than"

asdf/asdf/_jsonschema/_legacy_validators.py

Line 154 in ecb4d4f

cmp = "greater than"

asdf/asdf/_jsonschema/tests/test_validators.py

Line 332 in ecb4d4f

type = {"type": [{"minimum": 3}]}

asdf/asdf/_tests/test_yaml.py

Line 284 in ecb4d4f

buffer = io.BytesIO()

asdf/asdf/commands/edit.py

Line 101 in ecb4d4f

buffer = reader.read()

asdf/asdf/compression.py

Line 265 in ecb4d4f

buffer = np.empty((data_size,), np.uint8)

asdf/asdf/generic_io.py

Line 888 in ecb4d4f

buffer = self._buffer

asdf/asdf/generic_io.py

Line 897 in ecb4d4f

buffer = self._buffer

asdf/asdf/generic_io.py

Line 902 in ecb4d4f

buffer = self._buffer[:size]

[braingram]

Do these cause any issue for the benchmarking?

Looking over the "smells" I see:

• several in the copied/vendorized jsonschema
• uses of the word 'buffer'. I guess this is too generic of a word for sonarqubes liking?

[zacharyburnett]

Do these cause any issue for the benchmarking?

Looking over the "smells" I see:

• several in the copied/vendorized jsonschema
• uses of the word 'buffer'. I guess this is too generic of a word for sonarqubes liking?

There weren't any issues with benchmarking; it had a bunch of even more minor (minor-er?) suggestions that I ignored, but in this case it said cmp, type, and buffer are apparently already built-in functions / object names, so that might cause insidious errors

[braingram]

If these aren't an issue for the benchmarking I'm inclined to leave the code as-is. The cmp and type uses are within the vendorized jsonschema which will hopefully be replaced with a dependency if jsonschema re-adds the features required by asdf.

As far as I'm aware,buffer is a builtin only in python 2, perhaps sonarqube isn't scanning for only python 3 issues?

[zacharyburnett]

ok, sounds good to me then! I'll mark the issues as "ignore" in sonarqube