Corwin Joy

@Tishj Thanks for getting back to me! I'm not a big fan of this approach because I believe it will just mask the error (making it harder to reproduce) and...

@Tishj It looks like the PR you already submitted fixes this: https://github.com/duckdb/duckdb/pull/10512 See https://github.com/duckdb/duckdb/issues/10699 for more details. A future PR may want to improve the error handling / message for...

Here is the draft PR with a more complete test example: https://github.com/duckdb/duckdb/pull/10700

1. In this draft PR I was deliberately throwing an `int`, because when I was debugging this issue before it looked like the `py::error_already_set` (the translation of a Python exception...

As per above, it seems that the latest version of main does fix this issue.

This is going to be hard to diagnose (and fix) but I'm guessing what is happening is something like the following: 1. The query launches a subprocess to compute the...

@soerenwolfers @m4rs-mt I have created a branch with unit tests to replicate this problem. (Based on the PR I posted above for Windows). You can see the new tests here:...

@soerenwolfers The test shown here runs the query and detects when duckdb does not react to an interrupt signal promptly. This lack of response to an interrupt signal is what...

@soerenwolfers The reason that your python example doesn't have any problems is that python has built in interrupt handlers which will trap interrupt signals and handle them. In contrast, when...

@soerenwolfers One other possibility here is that your kernel is running out of memory with this large query and crashing. So, when I look at the default duckdb settings in...