Stuart Prescott <stu...@debian.org> writes: > This is certainly true for validating parsers, which will need > modification to stop them complaining about the missing standalone > License stanza; that's a relatively simple modification to not complain > if the licence key is within the predefined list from > /usr/share/common-licenses. Validating parsers we have in the archive > are lintian and cme; non-validating parsers such as debian.copyright > from python-debian and that in sources.debian.org require no > modification.
Agreed. It's a bit of a hack and kind of ugly, but one could just whitelist those inside validating parsers. I can imagine some non-validating parsers that would still care (a parser that for some reason wants to associate a long paragraph with every short token, for instance), but I don't know if any of those exist. > copyright-format/1.0 disallows misuse of licence tokens to point at > things that are not *exactly* the well-known licence and consequently, > there is no useful licensing information in the standalone License > paragraph. The BSD license's unfortunate inclusion in common-licenses complicates this a touch, but we probably should make an exception for it and not allow people to omit the license for it because of the change in copyright holder across various instances of the license. (The file name in common-licenses also doesn't match our license identifier.) There's a broader question about how we want to handle SPDX's concept of replaceable text in licenses. SPDX has done a lot more formal work on this than we have and annotates their license list with the text that's allowed to change from license to license while still being considered the same license. But SPDX doesn't have the problem of trying to replace verbatim copies of the license text. It's only metadata. So we'll have to figure out what we want to do about those cases. I think the BSD license is the only license being discussed in this bug with that problem, though, so we can postpone that problem for later by just omitting BSD from the list of licenses whose License extended paragraph can be omitted. > I'm quite happy to accept this proposal (without the brackets) as a > single change. It's minimal work for the parser and is an incremental > improvement to the format; because it relaxes a requirement, we could > even view it as backwards compatible and not increment the version > (although that has potential for confusion in the output of validating > parsers). I think we need to bump the version number for this, unless we're going to decide that the version number was a bad idea and we're not going to change it except for major changes. This is a backward-incompatible change for at least validating parsers, which seems like it's the basic purpose of a version number. > (The brackets, however, remain unnecessary.) Yup, agreed. -- Russ Allbery (r...@debian.org) <http://www.eyrie.org/~eagle/>
