some included libs are missing provenances and licenses

[joelb123]

It's good practice to indicate versions and origins (repository URLs and web sites) of included libraries. Several libraries are missing license files, which may violate the usage terms in many cases.

[martin-steinegger]

Thank you for looking into it. What libraries are missing a license?
Debian made a list of libs and copyrights:
https://salsa.debian.org/med-team/mmseqs2/-/blob/master/debian/copyright

[milot-mirdita]

Which library is missing licenses? I think everything not developed by us should be prefixed by the respective license. I would strongly recommend to not split away any library we bundle (except maybe zstd, as Debian is doing). We occasionally change stuff (e.g. I rewrote microtar completely, to fit our needs, maybe I should move it to commons now).\

We could add a module (mmseqs licenses?) that lists all libraries used and their licenses, or maintain our own version of debian's copyright file.

[joelb123]

Adding a module doesn't meet most common license terms that require redistribution of copyrights. You know your libraries better than I do, but the directories in lib/ that are missing licenses and readmes are alp/, cacode/, gzstream/, ksw2/, simd/, and xxhash/. Libs that you have edited, like microtar, should have a readme indicating you have done so.

[milot-mirdita]

I added missing the missing licenses. lib/simd is developed by us, it's just at this location for convenience as we occasionally sync it with HHsuite (https://github.com/soedinglab/hh-suite).

[joelb123]

Thanks, closing this.