Add support for ignoring files using REUSE.ignore

[gergap]

I'm missing support to ignore certain files or folders for the license check in reuse-tool. In our repos (commercial product) there are certain files that will not be distributed, hence they don't need a license and should not appear in SPDX listing to avoid confusion. The existing solutions are not ideal:

• .gitignore: I don't want to ignore those files in Git, just for SPDX generation
• REUSE.toml: you can specify a license, but cannot ignore the file

I introduced a new file REUSE.ignore, which can be used to extend the IGNORE_FILE/DIR_PATTERNS in the code. This file is read at startup if it exists.

Example patterns:

^LICENSE-3RD-PARTY.TXT
^hooks/
^km/
^gitlab/
^src/test/afl/

Full regex syntax can be used. If the patterns ends with '/' it will be added to the _IGNORE_DIR_PATTERNS variable, otherwise _IGNORE_FILE_PATTERNS will be used. Therefor the trailing slash gets replaced with a '$' to terminate the expression and avoid wrong substring matches. More generic patterns are of course still possible using ^dir.*/ which will be expanded to ^dir.*$.

[carmenbianca]

Hi @gergap, and thanks for the contribution. Unfortunately, I have to close this. This oft-requested feature is very specifically out-of-scope. And even if it weren't out-of-scope, the specification would need to be updated before the change is integrated, because the tool should reflect the specification.

For your case, you can use .reuse/dep5 (or REUSE.toml when it is released) to achieve some of what you want. Set your company as copyright holder, and apply the license LicenseRef-Proprietary or some such, making sure to create the file LICENSES/LicenseRef-Proprietary.txt.

Alternatively, do your linting and/or SBOM generation on a filtered repository. Create some workflow that deletes the files which will not be distributed, and then run the linting/SBOM generation on the remaining subset of files.

I hope that this is moderately helpful.