On 6/1/15, 4:31 AM, "Nicholas Kwiatkowski" <nicho...@spoon.as> wrote:
>The problem is the project was abandoned, and somebody other than the >original owner did the migration from Google Code to GitHub (this is in >the >last month -- which is why this took a bit longer). There are about 8 >forks of that "original" github project, all which include different >patches to make the engine work a bit better with certain edge cases. We >can't download the original from GC anymore since the migration took >place. > >If we were to bring the source and it's dependencies into the project via >a >git clone, do we then add those directories into the .gitignore to avoid >people adding them into our tree? I couldn't see any examples of where we >did this in any of our projects (we do bring in external bytecode via >swc's, but I haven't seen source). Hmm. I guess you could be right. I haven’t ever studied how we build batik-all-flex.jar in detail. I know we have a fork of batik with some mods in modules/thirdparty/batik, but that’s for code already in the ASF. IMO, if we don’t actually check in source from GH into our repo, then we aren’t treating the source as ASF source and can skip all the donation steps. Really, all of our third-party downloads, unless MPL, should be from source not binary like jars. So I don’t know if ‘git clone’ is the only way to download the source (I think for Google Closure Compiler we grab zips), but I would add the folders to .gitignore. IMO, you don’t need the true “original”, just pick one of the GH projects. Thanks, -Alex
