Open agitter opened 5 years ago
I think we did it per reviewer request upon publication.I think the issue is that if we remove it, git will still insist on downloading it as part of the history. Would adding it to .gitignore
get around this problem?
Okay, that sounds familiar. I also recall that we struggled to remove the history of these large files tracked with git-lfs (#23).
I was thinking that if we remove the dist/
subdirectory and then clone with a shallow clone (e.g. depth of 1) the deleted files would no longer be downloaded. I haven't tested that, and I'm not certain that meets any use case. The developers may want the full history, and users can download the *.tar.gz
files instead of cloning.
edit:
Would adding it to
.gitignore
get around this problem?
Yes, possibly we could delete the large files, add the subdirectory to .gitignore
, and then attach the example data as a release asset every time we build a new release. The code is attached automatically.
This sounds reasonable, if it works...
Making in a fresh clone in #32 reminded me that the contents of the
dist/
subdirectory are tracked by git. I don't recall the history of why we decided to do that, but it no longer seems necessary. We could ignore this directory and have users downloadOmicsIntegrator-x.y.z.tar.gz
and the latest version ofOmicsIntegrator-Examples.tar.gz
as assets in each release.