Please clean up the license.

[kou1okada]

Welcome a new maintainer of the transcode-open/apt-cyg.
I'm happy to get a chance to merge our code to the trunk line.
But, ...

The original apt-cyg has been licensed under the GPLv2.
In this repository, the apt-cyg was re-published under the MIT license by Stephen Jungels who is the original author of the apt-cyg.
But why do you re-introduce the GPLv2 code into this repository from the original apt-cyg?
It produces serious conflicts of the license.
You can merge our code, but we can't merge your's.

I think that the initial revision to merge forks must be fc54d5c, not the clone from the original apt-cyg on the Google Code (45de191).
And it must be merged back to the original apt-cyg on the Google Code.
It's clean way for the MIT license.
Maybe, any other way remains the affect of the GPLv2.

I'll be happy, if you clean up the license problem.
Thank you.

[skl]

@kou1okada you are referring to the header comments block in the code? That has never been altered and the LICENSE file does still state MIT. The original Google Code project is in fact still GPL v2 however I do not have control of that as I am only a committer. I'm happy to licence the header comments block as per the LICENSE file that Stephen Jungels created, however the Google Project will still state GPL v2 unless Stephen changes it.

[kou1okada]

Hmm,,,
But ｔhe header comments block in this repository was already altered from the GPLv2 to the MIT license by Stephen Jungels, when he republished here the apt-cyg.
Actually, the trunk line of this repository was clean from GPLv2, before you imported codes from the original repository on the Google Code.
As long as the header comments block has the GPLv2 term, LICENSE.txt file will not produce any effect.
We must patch to descendants of fc54d5c, if we want to keep this repository under the MIT license.
Because anyone, exclude Stephen Jungels, can not remove again the GPLv2 term from there.

Thanks.

[ghost]

I understand your concern about the licensing, but if thats the case even the first commit is suspect, see

e8891b5

How are we supposed to resolve that? Rebase the whole repo?

[kou1okada]

Sorry, I couldn't understand the meaning about "first commit".
e8891b5 was brought by Stephen Lang from r12 of the original repository on Google Code at 2014-02-20. It was not exist on this repository before 2014-02-20. acabffd seems the initial commit, but it is not the initial commit of this repository.
The initial commit of this repository is ab235c0 that was committed by Stephen Jungels at 2013-07-29.

The points of this license problem is below:

• Stephen Jungels created this repository under the MIT license.
• Stephen Lang reintroduced the GPLv2 term to the head of ./apt-cyg.
• Nobody can remove it.

I'm sorry, that I have no idea to solve this problem without doing git reset --hard to fc54d5c for master.
Fortunately, the critical patch merged to the current master (that re-introduced the GPLv2) is only fc54d5c...46fb5d3. It provides multi-architecture support and .xz package support.
But fc54d5c...3d67e12 is better solution to provide these supports, I think.
Unfortunately, some patches will be revoked by above solution, but it have only minimal impact on the function of apt-cyg.

Thanks.

[ghost]

This should now be fixed. I ran these commands

git reset --hard 46fb5d3
git merge -X ours origin~21
git cherry master origin | awk '$0=$2' |
  git cherry-pick --stdin --keep-redundant-commits -X theirs

It ended up regressing the README somewhat, but that can be easily fixed. Let me
know if the result is okay. I have a backup pre force push if it comes to that.

[skl]

I was originally trying to keep both the Google Code SVN project and this Git project in sync using git svn, that may help explain has this all came about in the first place. As to that is still possible going forward - I'm unsure as the repos are to be licensed differently. What do you think @svnpenn ?

[ghost]

If MIT will help with collaboration I vote for that, but personally I dont care as long as the code works.

[skl]

Agreed.
The version function still echo's "GPL" by the way.

p.s. Thanks for helping out with the repo.

[ghost]

I am innocent on that, it was even on first commit 2af0c86

[skl]

Not trying to blame, just bringing up as it seemed relevant to this discussion is all.

[kou1okada]

I appreciate you to clean up about license problem.
But I'm sorry that I brought some troubles to @skl.
Of course, I know GPL string in version function and I commented to there: 2af0c86#diff-f648982bd734e49811c931f511cb33e2R68.
But, unfortunately, I couldn't get answer from Stephen Jungels.

Thanks.

[ghost]

@kou1okada ok then, can we consider this issue closed?

[kou1okada]

Yes, please.
Thank you for clearing my concern.

[ghost]

@kou1okada @skl 062883c