I’ve noticed that Blender’s Github server has files that aren’t in the actual build version of Blender on git://git.blender.org/blender.git
For example I have a file called voronoi.h which is also on the Blender Github server, but the Github server also has a file called svm_voronoi.h which I do not have.
While the contents in both files are similar, albeit not identical, in this case, there are also files like drawnode.c which do not have a corresponding file i the build version.
The closest I came to a file like drawnode.c is drawnode.cc, which however consists of only a subset of drawnode.c.
What’s going on here?
drawnode.c was renamed to drawnode.cc and I suspect something similar happened to voronoi.h
Looks like the syncing to github isn’t done correctly and deleted files linger…
There’s nothing wrong with the sync, his links to svm_voronoi.h are pointing at the tree of commit 594f47ecd2d5367ca936cf6fc6ec8168c2b360d0 (oct 2020) which indeed still had those files.
if you go to https://github.com/blender/blender/tree/master/intern/cycles/kernel/svm you’ll see the current state which doesn’t
So what exactly is the issue here? Why is Github showing me files that are no longer there?
you’re asking github to show you the blender source from October 2020, when that file existed.
this links the October 2020 state of the source
this shows the current state
if you were to run
git checkout 594f47ecd2d5367ca936cf6fc6ec8168c2b360d0 in your source folder, it would also revert back to the oct 2020 state and you’ll have that file…
there is no issue…
That makes sense. I’m using Github to search for code pieces as Visual Studio becomes very slow with projects the size of Blender. I guess Github also searches for occurrences in past states of the repository.
Is there a way to have Github only search for occurrences in the current state of the repository?
No idea, can’t say i have ever used github for that purpose. I generally use ctrl-shift-f in Visual Studio to search for things, first time after you create a project it takes about a minute, subsequent searches complete virtually instantly.