Dmitriy Kalinin

@caphrim007 currently that's intentional behaviour. > Due to "reasons" (network hiccup, fat fingered path, path not existing at a particular version, etc) the paths include invalid paths. vendir produces a...

@yeahdongcn yeah this was meant to be an implementation that covers majority of use cases and remains simplest. i think we would be open to a PR to enhance it...

@LukeWinikates any thoughts on picking git (by a tag) vs githubRelease fetch types to fetch source code?

> "error: no files in release, try a git resource instead`" would be a good thing to do, I think. Happy to submit a PR for that, if it would...

may be also would be good to scope `{os}`/`{arch}` to something like `{vendir.os}`/`{vendir.arch}`, since in the example of githubRelease+http, tag is derived from that "local" context of githubRelease. probably also...

interesting questions on this topic: - vendir.lock.yml relationship to what os/arch is set - how are values os/arch are set: auto, manually - should vendir have rename functionality (e.g. ytt-darwin-amd64...

nah dont think so. this is more about overriding directories wholesale (with an updated dependency) whereas #29 is more about modifying downloaded dependency. there is some relationship but prolly a...

> We don't want to modify any file downloaded what i meant is that we see directory represented via a source as a single asset. so any modification to that...

> Vendir should ignore anything here. Not sync it. Not delete it. what would happen when upstream deleted a directory that happens to currently contain *.tf.json? > but to solve...

and to be truthful to the example description, it would be this right? ``` ignorePaths: &ignore - **/backend.tf.json - catalog/** ``` to cover all nested backend.tf.json. what should happen if...