Replace Placeholders in Licenses

[JakobDev]

Some Licenses have Placeholders e.g. the BSD Licenses start with this line:

Copyright (c) <year> <owner>. 

This PR replaces the placeholders with the correct value, when creating a new Repo. Tested with the BSD, GPL, AGPL and MIT Licenses, which are the most used.

[delvh]

Hmm, I'm not sure if it is a good idea to manually overwrite licenses.
While this change does not seem malicious, I don't think it is a good idea to manipulate a legally binding document in any way that the original author did not intend.
If the owner of the license did not update their license to throw out the placeholder, is it really a good idea that we do it for them?
Also, doing that probably gives them the false impression that they already filled out the fields, while actually the license is still not filled out when cloning the repo…

[JakobDev]

I don't think it is a good idea to manipulate a legally binding document

GitLab does already replace the placeholders in Licenses for Years now. And I think, GitLab has Lawyers who say this is not an Issue.

in any way that the original author did not intend.

I think the original Author intended this, or else he would not have used Placeholders.

If the owner of the license did not update their license to throw out the placeholder, is it really a good idea that we do it for them?

Most people do not edit the License. They just select License foo for the Project and that's all. They expect, everything is done with that and not, that they have to edit the License and remove Placeholders.

Also, doing that probably gives them the false impression that they already filled out the fields, while actually the license is still not filled out when cloning the repo…

I don't think that. Most Forks are just for making PRs. If someone makes a hard Fork, he probably takes a look at the License anyway.

[KN4CK3R]

Gitlab matches more placeholders:

  PROJECT_TEMPLATE_REGEX =
    %r{[\<\{\[]
      (project|description|
      one\sline\s.+\swhat\sit\sdoes\.) # matching the start and end is enough here
    [\>\}\]]}xi.freeze
  YEAR_TEMPLATE_REGEX = /[<{\[](year|yyyy)[>}\]]/i.freeze
  FULLNAME_TEMPLATE_REGEX =
    %r{[\<\{\[]
      (fullname|name\sof\s(author|copyright\sowner))
    [\>\}\]]}xi.freeze

[zeripath]

hmm... I guess licenses tend to be small but this could be significantly improved.

for example if the license does not contain a '<' there is no point doing any of the replacements. So that could be a very simple improvement.

If one were then to use a strings.Replacer, whilst this causes a copy of the bytes to be made, the original []byte could then be used as the basis of a buffer to write into.

However, I'll approve because I think the perf issues here are small enough.

[KN4CK3R]

The placeholders do not cover all the cases:

[YEAR]

gitea/options/license/MirOS

Line 3 in 8400146

Copyright [YEAR] [NAME] [EMAIL]

{YEAR}

gitea/options/license/NASA-1.3

Line 45 in 8400146

[The following copyright notice will be used if created by a contractor pursuant to Government Agency contract and rights obtained from creator by assignment. Government Agency will insert the year and its Agency designation and remove the bracketed language.] Copyright (c) {YEAR} United States Government as represented by ___ ____. All Rights Reserved.

[KN4CK3R]

Why not use a regex? You are still missing placeholders.

[name of copyright owner], [yyyy] (Apache-2.0)
[name of copyright holder] (MulanPSL-2.0)
<COPYRIGHT HOLDERS> (BSD-2)
<copyright holders> (ECL-1.0)
<AUTHOR> (AAL)
<author's name or designee> (OPUBL-1.0)
[one or more legally recognised persons or entities offering the Work under the terms and conditions of this Licence] (CC-BY-NC-SA-2.0-UK)
...

[wolfogre]

I got inspired by this PR, and I think we could handle this in a more maintainable way: #24354.

What do you think? @JakobDev

[wxiaoguang]

ping ~~~

Personally I think #24354 seems handling more cases and could be covered by tests

[JakobDev]

I'm closing this PR in favour of the new one