Copyright detection range is sometime approximative

[tdruez]

input.txt:

Files: src/macos/src/macos_main.cpp
Copyright: 2001, Eric Lavigne
License: curl

Files: docs/examples/rtsp.c
Copyright: 2011, Jim Hollinger

scancode -ce --json-pp - input.txt

Current results:

"copyrights": [
    {
      "value": "Copyright 2001, Eric Lavigne",
      "start_line": 2,
      "end_line": 3
    },
    {
      "value": "Copyright 2011, Jim Hollinger",
      "start_line": 6,
      "end_line": 6
    }
],

Expected results:

"copyrights": [
    {
      "value": "Copyright 2001, Eric Lavigne",
      "start_line": 2,
      "end_line": 2
    },
    {
      "value": "Copyright 2011, Jim Hollinger",
      "start_line": 6,
      "end_line": 6
    }
],

The first copyright is detected on line 2 to 3, while only present on line 2, where the second copyright is properly detected on line 6 only.

[sritasngh]

@tdruez can I work on this issue?

[sritasngh]

input.txt:

`Files: src/macos/src/macos_main.cpp
Copyright: 2001, Eric Lavigne
License: curl`

$ SCANCODE_DEBUG_COPYRIGHT=yes scancode --copyright --json-pp - input.txt

Result:

> DEBUG:cluecode.copyrights:# candidate_lines: evaluating line:(1, 'Files: src/macos/src/macos_main.cpp\n')
> DEBUG:cluecode.copyrights:# candidate_lines: evaluating line:(2, 'Copyright: 2001, Eric Lavigne\n')
> DEBUG:cluecode.copyrights:   candidate_lines: line is candidate
> DEBUG:cluecode.copyrights:# candidate_lines: evaluating line:(3, 'License: curl\n')
> DEBUG:cluecode.copyrights:   candidate_lines: line is in copyright
> DEBUG:cluecode.copyrights:candidate_lines: finally yielding candidates
>     [(2, 'Copyright: 2001, Eric Lavigne\n'), (3, 'License: curl\n')]

@pombredanne @tdruez copyright scan result is giving incorrect range because DEBUG:cluecode.copyrights: candidate_lines: line is in copyright. Could it be a good approach if I limit candidate _lines to only single line when it get

DEBUG:cluecode.copyrights:# candidate_lines: evaluating line:(2, 'Copyright: 2001, Eric Lavigne\n')
DEBUG:cluecode.copyrights: candidate_lines: line is candidate

It will also help to debug this #2445.

[pombredanne]

Could it be a good approach if I limit candidate _lines to only single line when it get

That would truncate several copyrights. Instead I think a better approach is to post-process the detected values against the original line range and find the correct start and end line (and even better as a bonus the offset or the start and end column in each line)

[pombredanne]

can I work on this issue?

Sure thing, and you do not need to ask for permission! Yet that's always a good idea to ask like you did here so we can avoid stepping on each other's toes. 👍

[pombredanne]

The latest https://github.com/nexB/pygmars/ now supports Token with line numbers and offsets. This is being integrated in copyright detection

[pombredanne]

This is now all fixed in the develop branch and we get these results:

        copyrights:
            -   value: Copyright 2001, Eric Lavigne
                start_line: 2
                end_line: 2
            -   value: Copyright 2011, Jim Hollinger
                start_line: 6
                end_line: 6
        holders:
            -   value: Eric Lavigne
                start_line: 2
                end_line: 2
            -   value: Jim Hollinger
                start_line: 6
                end_line: 6
        authors: []