Validate GLE license is correctly detected

[pombredanne]

The license @ https://sources.debian.org/data/main/g/gle-graphics/4.3.3-3/doc/LICENSE.txt should be scanned and we should validate that everything is detected correctly.
Reported by @pabs3

[shricodev]

@pombredanne Sir, I tried running the scan with these parameters:
./scancode --json-pp gle-result.txt --license --package --url --email --copyright LICENSE.txt

and these look good to me, all the license seems to be detected properly :P
BTW is there something to look into particularly?
The output was large so I saved it here: https://github.com/OctoPie23/OctoPie23/blob/main/temp/gle-result.txt

[AyanSinhaMahapatra]

@OctoPie23 btw, you would want to use --license-text --license-text-diagnostics also when you're checking for license detection issues.

Thanks, Looks good to me, btw @pombredanne see these interesting case:

{
          "license_expression": "gpl-1.0-plus",
          "detection_log": [
            "possible-false-positive",
            "not-license-clues-as-more-detections-present"
          ],
          "matches": [
            {
              "score": 50.0,
              "start_line": 1439,
              "end_line": 1439,
              "matched_length": 1,
              "match_coverage": 100.0,
              "matcher": "2-aho",
              "license_expression": "gpl-1.0-plus",
              "rule_identifier": "gpl_bare_word_only.RULE",
              "rule_url": "https://github.com/nexB/scancode-toolkit/tree/develop/src/licensedcode/data/rules/gpl_bare_word_only.RULE",
              "matched_text": "gpl."
            }
          ]
        },
        {
          "license_expression": "gpl-2.0 AND agpl-3.0-plus",
          "detection_log": [
            "possible-false-positive",
            "not-license-clues-as-more-detections-present"
          ],
          "matches": [
            {
              "score": 98.0,
              "start_line": 1514,
              "end_line": 1514,
              "matched_length": 2,
              "match_coverage": 100.0,
              "matcher": "2-aho",
              "license_expression": "gpl-2.0",
              "rule_identifier": "gpl-2.0_238.RULE",
              "rule_url": "https://github.com/nexB/scancode-toolkit/tree/develop/src/licensedcode/data/rules/gpl-2.0_238.RULE",
              "matched_text": "GPL v2 ***"
            },
            {
              "score": 90.0,
              "start_line": 1516,
              "end_line": 1516,
              "matched_length": 2,
              "match_coverage": 100.0,
              "matcher": "2-aho",
              "license_expression": "agpl-3.0-plus",
              "rule_identifier": "agpl-3.0-plus_143.RULE",
              "rule_url": "https://github.com/nexB/scancode-toolkit/tree/develop/src/licensedcode/data/rules/agpl-3.0-plus_143.RULE",
              "matched_text": "GPL GhostScript"
            }
          ]
        },

These are marked as possible-false-positive since these are either triggered by 1/2 words of gpl like text, but as there are other concrete GPL/AGPL detections in other parts of the file, these are not discarded as false positives/added to license clues instead, i.e. not-license-clues-as-more-detections-present. What do you think?

[shricodev]

@AyanSinhaMahapatra Sir, One thing I noticed when running the --license-text is that when
there is complete match of the license, it just clutters the output file.

for reference: https://github.com/OctoPie23/OctoPie23/blob/main/temp/gle-result-with-license-text.txt#L2005

Because of the no line wrap of Github it would not be much visible. Instead in such full match cases we could add a link to the license instead of writing it completely in the matched_text.

Something like this:
matched_text: Complete Match of <License Name> <Link>

I can open up an issue if this seems to be a good idea to you and possibly start working on it ;)
This is just a very rough idea that came into my mind. Feel free to ignore it if it makes no sense :P

[AyanSinhaMahapatra]

there is complete match of the license, it just clutters the output file.

IMHO, that is why this is a CLI option which can be enabled, and not a default option.
Sometimes we do need to look at this data, for better understanding of the issues, even if
it clutters the output somewhat. We also have this project: aboutcode-org/scancode.io#450 which helps in providing an UI for review.

[shricodev]

@AyanSinhaMahapatra Thanks for clearing my doubt. Got it :)