pombredanne
commented
5 months ago Thanks for the report!
How do I need to interpret the log output? Is the token
of the URL optional here?
The
Why is the match for src/licensedcode/data/rules/cc-by-3.0_1.RULE with 100% cov not considered?
It is shorter and contained in another match. Let me explain what happens.
You get these three matches:
-
Match 1: LicenseMatch: 'cc-by-3.0', lines=(0, 0), matcher='3-seq', rid=cc-by-3.0_23.RULE, sc=54.55, cov=54.55, len=6, hilen=3, rlen=11, qreg=(87, 92), ireg=(1, 8)
-
Match 2: LicenseMatch: 'cc-by-nc-sa-3.0', lines=(0, 0), matcher='3-seq', rid=cc-by-nc-sa-3.0_23.RULE, sc=54.55, cov=54.55, len=12, hilen=4, rlen=22, qreg=(85, 96), ireg=(3, 21)
-
Match 3: LicenseMatch: 'cc-by-sa-3.0', lines=(0, 0), matcher='3-seq', rid=cc-by-sa-3.0_1.RULE, sc=100.0, cov=100.0, len=8, hilen=2, rlen=8, qreg=(88, 95), ireg=(0, 7)
The Match 1 attributes are:
- matcher='3-seq': match was done using the sequence matcher, e.g., a diff
- rid=cc-by-3.0_23.RULE: name of the rule file
- sc=54.55 : overall score
- cov=54.55: coverage as the proportion of the rules words matched, here 54.55% of the words were matched
- len=6 : length of the match as a number of matched words
- hilen=3 : number of matched words that are important aka. "legalese"
- rlen=11 : rule words length
- qreg=(87, 92) : matched query region as start and end word positions
- ireg=(1, 8) : matched index (e.g., the rule) region as start and end word positions
Here is what happens:
- Match 3 is entirely contained (words-wise) in the longer Match 2 and therefore discarded
- Match 1 and 2 have the same score and significant overlap, yet Match 2 is longer by one word and has one more legalese words and hence wins
- You end up with Match 2.
FWIW, this is something @AyanSinhaMahapatra is working on to massively fix likely with https://github.com/nexB/scancode-toolkit/pull/3254
In the meantime the fix is going to be like in https://github.com/nexB/scancode-toolkit/pull/3644 The faulty rule is https://github.com/nexB/scancode-toolkit/tree/develop/src/licensedcode/data/rules/cc-by-nc-sa-3.0_23.RULE and the cure will be to add required phrases and may be a new rule.
-
add curly braces in https://github.com/nexB/scancode-toolkit/blob/develop/src/licensedcode/data/rules/cc-by-nc-sa-3.0_23.RULE
licensed under the Creative Commons License, {{Attribution-NonCommercial-ShareAlike 3.0 }} unported. {{http://creativecommons.org/licenses/by-nc-sa/3.0/legalcode}} -
create a new rule for good measure for
cc-by-sa-3.0with this text.Creative Commons license {{ http://creativecommons.org/licenses/by-sa/3.0/legalcode }}
jloehel
Description
This code snippet:
from https://github.com/Xilinx/u-boot-xlnx/blob/3290b109bfa70d65ed4ce49ed84afed0ed4335e0/drivers/video/console_truetype.c#L39 gets detected as
CC-BY-NC-SA-3.0instead ofCC-BY-SA-3.0How To Reproduce
I have tested it also with the patches from https://github.com/nexB/scancode-toolkit/pull/3644, but it does not fix the issue.
Logs
Log Snippet with
SCANCODE_DEBUG_LICENSE=True:How do I need to interpret the log output? Is the token
<nc>of the URL optional here? Why is the match forsrc/licensedcode/data/rules/cc-by-3.0_1.RULEwith 100% cov not considered?System configuration