@wilkerlucio: Is there an easy way to limit the length of Pathom exception logs? We have well over 200 resolvers now and the exceptions are unwieldy. (And are crapping up production logs as well.)
hello, I'm interested to understand how it gets there, its for sure something I'm looking forward to improve (the error handling/reporting)
can you tell me what you see there, that we might be able to avoid/eliminate?
It’s dumping huge paths that seem to involve every resolver.
There will be the call stack, the exception message, then an enormous Pathom path.
is there many optionals that might be accumulating errors from possible paths that failed? or its more straithforward?
I just generated a simple one by querying a non-existent key…
And admittedly it’s not overly long. But most of the text is generated on the :com.wsscode.pathom3.connect.planner/graph key.
that's a good point to note, because we can change these to not be exceptions, or make them simpler exceptions
Lots of lists and maps that would be very hard to make use of, and instead make for a lot of scrolling.
yeah, I've been thinking about it, removing the graph from the exception data, and make accessing it indirect (so they dont bloat the exceptions)
The other verbose piece is having to trace the exception message itself back to your own code, skipping the Pathom namespaces.
That might not be in your hands to overcome.
can you make a demo repro of a bad one similar to what you see, so I can work on top of it here?
Might be kind of hard… seems like a function of the number of resolvers…
How about I private message you with an exception?
This is one of the shorter ones I’ve seen.
Mmm, it’s too big to send over Slack.
If you send me an email address, or something, I can get you the full log. Otherwise I’ll think on a way to make a repo.