GPL-2.0(*) templates: var text to handle variation in FSF address formatting?

[pmonks]

In this GPL-2.0 license text, the FSF's address has been split across multiple lines, and 2 comma characters (,) removed, which means that matching of this text via the various GPL-2.0(*) templates fails. This was originally discovered in this Spdx-Java-Library issue.

In that issue, @goneall suggested that variations in this section of the license text could be handled in the GPL-2.0(*) templates via the addition of <var> elements.

[jlovejoy]

hi @pmonks - so... white space (ie., split across multiple lines) shouldn't matter as per B.4 https://spdx.github.io/spdx-spec/v2.3/license-matching-guidelines-and-templates/

However, punctuation does matter (see B.6). So, would this mean we'd be adding markup just for the commas?

[goneall]

@jlovejoy Just to clarify - the issue is with the matching regular expression in the license XML - this will need to be fixed for matches to occur. In other words, there is an <alt tag with a match regular expression that assumes the text it is matching is in the same line - which is a bad assumption.

[jlovejoy]

@goneall hmmm... so that could be a problem with any license that has an tag, then?

[goneall]

so that could be a problem with any license that has an tag, then?

@jlovejoy it has to do with that is in the match="..." string - I don't think this will be an issue for most of the licenses since they typically use .+ which will match line feeds.

[jlovejoy]

thanks @goneall - so... is there a PR needed here?

[goneall]

so... is there a PR needed here?

Yes - a change to the regex to allow for line feeds. Personally, I would just do a .{X,Y} where X is the minimum number of characters for a match and Y is the maximum number of character for a match rather than the specific text.

[jlovejoy]

similar to #2279 - will let @goneall approve!

[goneall]

Resolved with #2279