Reviewing the licence of the Cinelerra source tree.
We need to review the Cinelerra-CV sources that muzzol has pushed to bzr. The branch can be seen at https://code.launchpad.net/~cinelerra-ubuntu/+branch/cinelerra-cv/2.1.0+svn20070109, and there are instructions there for pulling it from http://bazaar.launchpad.net/~cinelerra-ubuntu/cinelerra-cv/2.1.0+svn20070109.
- Bob wants to use Cinelerra and Ubuntu to edit his video, and doesn't want to use a non-free equivalent. However, he is unable to, nor does he want to, build Cinelerra himself.
- We need to get Cinelerra into the repository. This cannot happen unless all licence issues are sorted, and thus the purpose of this spec.
The task could be done in either of two ways, both of which I will outline below.
1. We each create, on Launchpad, a branch of ~ubuntustudio-dev/ubuntustudio/cinelerra (thus becoming ~<user>/...). Each branch has the file, say, ./list in it, where each reviewer keeps a note of the licence of every source file they have reviewed. When the reviews have finished, all the personal branches are merged back into ~ubuntustudio-dev.
2. We each pull from and push to the ~ubuntustudio-dev branch. Each reviewer has their own file in which they update their list of files and licences. This file could be called something like ./list.reviewer, or ./list<number>. When the reviews are complete, these files can be merged into one file, ./LICENCES, or debian/copyright.
As the source tree is 59MiB, the first method seems like too much uploading - to create new branches for each of us, and then make a merge at the end. It would almost double the amount of up/down traffic we need. As a result, I suggest we choose to follow method 2.
However, either way we will need to decide how to split the 5200 files in the source tree between the team. It does indeed look a daunting task, but if divided up properly, it can be made much easier. It may be possible to have a system of locking, where we each check each others' LICENCES files, where it would be specified which files have been and will be done.We could also have a page like UbuntuStudio/ToPackage, where we divide up and allocate the sources. Either way, we cannot waste time stepping on toes.
(The below was taken in part from my post to the UbuntuStudio mailing list).
Due the the upstream SVN server, I have been unsuccessful in having Launchpad import the repository automatically. However, I have done so manually, and you can find the clean and current-ish trunk at ~cinelerra-ubuntu/cinelerra-cv/trunk, and my cleaned-up version of muzzol's branch at ~cinelerra-ubuntu/cinelerra-cv/2.1.0+svn20070109. When Launchpad is able to do imports automatically, I will be the first (maybe second) to hear, and will set that up again.
We need to decide firmly how we are not going to step on each others' toes. I propose we create a list of all the files, and split it into, say, 50 parts - or files of 100 lines. Each reviewer takes one or more of these split files, and marks down when they have reviewed all the files in their allocated list.
I have created the file lists, and they can be found and assigned at [Cinelerra/LicenceReviewStatus]. As the lists are in order, we can simply merge them all back together when we are done. Recording the information on licence and author should be done in the following way, with the same format for each line in the file: <filename>, <licence>, <authors list>.
Remember, try and always push to ~cinelerra-ubuntu/cinelerra-cv/2.1.0+svn20070109 (every member of ubuntustudio-dev is a member of cinelerra-ubuntu, but I'd still like a distinction between those that maintain Cinelerra - cinelerra-ubuntu - and those that maintain UbuntuStudio - ubuntustudio-dev - even if we are the same at first), and keep the filenames in a standard format, ie list<number>, like they were found. Push them up as they were downloaded, with the additions of the licence and author information as described above (comma delimited). Push them into the root directory/review of the sources dir (ie, cinelerra/review). Create it if it doesn't exist when you pull your tree, and do not take an already assigned list.
Once we have this structure done we should post to the Cinelerra ML. Invite people to help.