The strangest thing just happened on 2 separate node red instances on 2 machines...
Machine 1: I had made a bunch of edits yesterday to a flow. Today I did a reboot of that machine and all my changes were gone after it started back up. I found the changes in another flows file in ~/,node-red:
Oct 21 13:19 flows_machinename.local.json Dec 11 15:13 flows_machinename.localdomain.json
I was able to recover the changes by copying the localdomain one to the local one.
machine 2: Then, I was seeing an old bug from machine 2 - one I had fixed a long time ago. When I went to look at the flow, it looked fine - the bug wasn't there. I disabled the node that was sending me the it was getting so I could just view the debug panel without my phone continuously vibrating. But after deploy, the buggy behavior went away. I re-enabled and redeployed and everything was fine. No more bugs.
I remember what was causing that old bug. It was a -c option in a specific exec node. I even reproduced it on the command line and confirmed it's the same output the node was texting me. It seems like there was a stale version of the flow still in memory that somehow got enabled?? But I never rebooted that machine...
I don't understand what the heck just happened. It seems as if both machines obtained stale/old flows. Machine 1 (on node-red version 1.2.7) actually showed the stale flow in the editor. Machine 2 (on node-red version 1.0.3) was running the old flow but showing me the latest one.
I can imagine how machine 1 ended up with a stale flow because of the reboot, but I have no idea how machine 2 ended up with a stale flow. Is this the same bug with 2 manifestations? And is this a known issue that an update will resolve?