odo doesn't properly re-sync all files if a new deployment is created on 'odo push'

[johnmcollier]

/kind bug
/area devfile

I'm seeing an issue on master with odo push when we re-sync the files of a project after a component restart (such as from changing the env in the devfile). The push will succeed and it will say the source code synced fine, but if I exec into the container (or look at its logs), I'll see that only the devfile (and whatever files were changed locally or created as part of the build), were synced, and the rest of the source code was not synced

Files synced on first odo push (full logs: first-push.txt):

I0826 21:22:20.287844   92492 adapter.go:211] Copying files /Users/johncollier/ododev/brokesync/nodejs/.gitignore
/Users/johncollier/ododev/brokesync/nodejs/LICENSE /Users/johncollier/ododev/brokesync/nodejs/README.md 
/Users/johncollier/ododev/brokesync/nodejs/devfile.yaml /Users/johncollier/ododev/brokesync/nodejs/package-lock.json
/Users/johncollier/ododev/brokesync/nodejs/package.json /Users/johncollier/ododev/brokesync/nodejs/server.js 
/Users/johncollier/ododev/brokesync/nodejs/test /Users/johncollier/ododev/brokesync/nodejs/test/test.js to pod

Files synced on odo push with redeploy (full logs: push-redeploy.txt):

I0826 21:29:25.688476   93487 adapter.go:211] Copying files /Users/johncollier/ododev/brokesync/nodejs/devfile.yaml 
to pod

If I use the -f flag on odo push, sync is still broken. Additionally, the issue can be seen if you do odo delete, followed by odo push again.

It seems we're not properly detecting the scenarios where we need to bypass the odo-file-index.json file and do a full sync (such as component redeploy, or force push)

[johnmcollier]

Marking this as high priority since sync is pretty much broken if a container needs to be recreated, and that this is a regression from v1.2.5 (which properly detected when/when not to do a full sync)

/priority high

[johnmcollier]

Closing as a dupe of #3838