Matthias Oertel

I cleaned up the code a bit and added a bench case that actually has to remove whitespace. There the difference is less stark, but still pretty big.

I'll remove the trim_original and then we'll see whether the PR run succeed. I think it was just flakey CI.

Apart from that from my side the PR would be done. There is nothing more I was thinking of including.

That's it from my side. Let me know if there is more that needs to be done.

Might be related: https://github.com/prisma/prisma/issues/12524

Related: https://github.com/prisma/prisma/issues/14543

We're about to look into this but can't yet say whether it's a big issue or a quick fix since we haven't reproduce it yet. I'll post an update as...

I tried reproducing the issue by running the same request you did on the same schema. For me, the created MongoDb query contains nowhere near 50 aggregation pipeline stages. Can...

This is the update, its very large, but does not use the aggregation pipeline. Can you also try running that on Atlas directly? ``` db.Standing.updateMany({ _id: { $in: [ ObjectId("632a0fd5392630a54d36fcba"),...

Ok, I ran all the request we trigger directly against Atlas and they are succeeding. So I can't reproduce this, I'm sorry.