Fixes the segfault on my machine but feel free to test if we think there's anything dubious here. I opted to remove the autoconf substitution here since this seems pretty straightforward.

Could we do jit_stencils*.h to catch both files with one pattern?

Also, does the PCbuild logic need to be updated? Or is it good?

I'm no make expert, but based on my testing and what I've read, it seems like the wildcard keyword is required for proper glob expansion in Makefiles. It doesn't expand out in the same way as it would in a shell with * (see failed commits above 🫠 ). Windows should be good since we explicitly enumerate the stencils by platform.

That said...I...I am truly perplexed as I can no longer reproduce the segfault. The first commit from last night fixed the issue locally, so I know that this did solve the issue I was seeing, but...????

@brandtbucher Can you still repro?

I know what the issue is here, will fix later.

I am truly perplexed as I can no longer reproduce the segfault.

I have the exact same problem, it suddenly stopped crashing and I haven't figured out a way to get into the segfaulting state ever since. So it seems normal for this abnormal issue.