aboutcode-org / aboutcode-org/scancode-toolkit

Consider collecting the "License-Filename" tag

未关闭
#2,852 0 条评论 0 个 reaction 已指派 0 人 在 GitHub 查看
license scan package-formats
主要语言
Python
星标
2.6k
派生
791
平均合并
1 天 12 小时
30 天内合并 PR
5

描述

Seen in https://git.sr.ht/~etalab/sill.etalab.gouv.fr/tree/master/item/src/clj/sillweb/config.clj#L3

> ;; Copyright (c) 2019-2022 DINUM, Bastien Guerry
> ;; SPDX-License-Identifier: EPL-2.0
> ;; License-Filename: LICENSES/EPL-2.0.txt

We have one example of such rule https://github.com/nexB/scancode-toolkit/blob/ebb55c341bc6ba644d2be0e531ffbfc2948bca47/src/licensedcode/data/rules/gpl-3.0_115.RULE

There are enough in the wild: https://github.com/search?p=2&q="License-Filename"&type=Code

It likely may need to be a new detector for licenses that would:

1. report these tags with a synthetic new license rule that would have the detected filename as a `referenced_filenames`
2. then use this in the regular process to follow file references

This is often used with an SPDX-License-Identifier tag, so we may also be consider these tow together.
Along the way it could be interesting to consider other structured tag-like conventions used to create structured metadata in source code such as dunder variables in Python javadoc, jsdoc and similar. We deal with these with rules today, but that's not optimal

贡献指南

打开贡献指南

评估

这个 Issue 还没有评估数据。

把新 issue 发到你的邮箱

精选适合新手参与的 GitHub issue 摘要。