search

ReactionMechanismGenerator / RMG-tests

Continous Integration Testing Platform for RMG-Py
3 stars 14 forks source link

Better Logging of error causes #47

Closed goldmanm closed 1 year ago

goldmanm commented 7 years ago

I ran a local job and I think it had an error, since it didn't create all the files in data_dir. However, the slurm.out file didn't state where the error occurred. It would be helpful for the slurm output to mention where the error likely occurred. Increased logging may be helpful. If there is already a way to get logging at the source of the error, that would be helpful.

KEHANG commented 7 years ago

@goldmanm I've looked at your RMG-tests directory. I think I know the cause. If you fetch the newest RMG-tests, it should work

goldmanm commented 7 years ago

fixed

goldmanm commented 7 years ago

actually, while the original cause might be fixed, I am not sure if better bug detection and logging has been implemented.

goldmanm commented 7 years ago

One additional thing that could be beneficial when logging is to state which branch is throwing an exception (the master or new branch). I got confused on a recent run. When the master branch was throwing an exception and crashing the RMG-tests job, I thought it was the new branch that was failing. Maybe I just missed where the log mentioned which RMG version was crashing.

JacksonBurns commented 1 year ago

The changes described in #86 always treat the main branch as "stable" and clearly indicate that the proposed changes on another branch (the "dynamic" changes) are responsible for regression failures. Closing as completed.