Exit on error?

[reinvantveer]

For the sake of discussion: a concern for me at the moment is the strategy for error handling - it appears to me that it is following a "keep calm and carry on" kind of thing 😉 . But what if you'd want to exit on error? The error itself (for example in postgis.tx_execute() gets assigned to self.e, but at what point can or should halt this error further execution?

The use case for this is to further support automated workflows - without having to resort to logfile analysis. You could scan logs for Errors of any kind in the logs, but it would be rather nice to fail the import when for some reason or other, the receiving end of the transformation isn't available or busy or something.

Any thoughts/pointers on this?

[sebastic]

Stopping on error is the right thing to do in my opinion, the current behaviour is very good.

[reinvantveer]

I'm not sure that the current behaviour exits on error though @sebastic can you confirm?

[sebastic]

It does not, I ran into it with NLExtract, see: nlextract/NLExtract#331 (comment)