Details
-
Type:
Story
-
Status: Done
-
Resolution: Done
-
Fix Version/s: None
-
Component/s: Developer Infrastructure
-
Labels:None
-
Story Points:7.5
-
Epic Link:
-
Team:SQuaRE
Description
Refactor how we manage copyright and license information in stack repositories
- Identify a list of packages to process
- build and test an automated system of
- adding a global COPYRIGHT to each repo. Content will be “Copyright YYYY-YYYY The LSST DM Developers”. Years will be determined by git history.
- adding a GPLv3 license file to each repo.
- changing the boilerplate in all files to say "See the COPYRIGHT and LICENSE files in the top-level directory of this package for notices and licensing terms.” Use https://gist.github.com/ktlim/fdaea18ab3d39afdfa8e
- automatically branch, commit, merge and push
And deploy this automated system.
Attachments
Issue Links
Activity
Hold on, I incorrectly implemented C++ comment conversion. Will correct.
C++ migration works now!
And Kian-Tat Lim, I get that the code is a bit inefficient. As you mention, my implementation of the replacement algorithm does have the nice feature of killing the extra space after a # in the Python license blocks. For a run-once utility that has a bit of extra functionality for a bit worse runtime performance, I consider this a fair tradeoff 
I actually like the consistency of having "#<sp>" empty comment lines in Python. But I'm not wedded to it, and I guess it's not PEP8 compliant.
PR merged by Frossie Economou
It's not a huge deal, but I think my original replacement script is somewhat more efficient because it only does regexp matches until the "see" line is found; the entire rest of the file is just copied verbatim. I think you're matching every line against the comment regexp and then doing tests on every content-containing comment thereafter. Is that just about replacing "#<sp>" with "#"?