add --copyright-style option to reuse lint to enforce checking for a consistent copyright style

[kbroch-rivosinc]

It would be useful if the same functionality https://github.com/fsfe/reuse-tool/pull/248 that was created for reuse addheader --copyright-style could be used to check for that copyright style consistency with reuse lint.

[mxmehl]

Thank you for your suggestion. However, my inclination is not to implement this feature.

Let me explain why: our strong conviction is that REUSE compliance is binary: you are either compliant or not. Therefore, there are no flags to influence the strictness of a lint.[^1]

The flag you are proposing introduces a differentiation. In your repo or your configuration, REUSE compliance would be different than from a check I would run on my machine (e.g. when doing a pull request and confirming compliance).

This rather calls for a project-specific convention or style you want to enforce. Which is fine but for which the REUSE tool is the wrong way. There are a number of projects that are REUSE compliance but in some way mandate a certain order, format, placement, or content of the copyright line. They do this with own scripts which is the best solution as these requirements are custom. I would recommend the same procedure for your project as it gives you the freedom to define the format you want, given that it's still compliant with REUSE's supported annotation styles.

[^1]: The one exception is whether a lint shall optionally include submodules, which is a whole different topic and trade-off.

[kbroch-rivosinc]

Thanks for the reply and I appreciate your perspective. It is along those same lines that brought me to this request as I would rather the REUSE v3.0 spec only allow one way to denote copyright tags, namely SPDX-FileCopyrightText. My assumption was that proposing a spec. change for (say v3.1) wouldn't be considered.

[mxmehl]

Mandating only SPDX-FileCopyrightText was discussed a few times before, @silverhook is a strong proponent. However, we so far decided not to take that route (yet) as there are hundreds of REUSE compliant projects that would become incompliant.

My personal question is whether regulating this would actually bring us closer to our aim (making licensing and copyright easy and understandable for humans and machines alike). Sure, we could avoid some false-positives, but is the unavoidable conversion battle the most effective way to spend REUSE's energy and resources?