runsascoded / gsmo Goto Github PK

gsmo run will fail if an existing gsmo jupyter server is running; in principle, running inside the existing container should work and is the right thing to do in that situation

currently there is one fixed container name that is used each time gsmo is run, and it is rm'd if it already exists

it might be nice to keep around the last few containers, for posterity; each gsmo run could rename them all each time it ran, or each new container could get an autoinc idx…

general test of submodule usage would also be good

…and merge generated commits back upstream

sometimes (always?) a whole PapermillExecutionException ends up serialized onto the first line of the commit msg, which doesn't interact nicely with git tooling that expects a short "subject" line at the top of large/multi-line commit msgs

should pull out an error summary and use that as commit subject