Error message when two tasks try to claim the same interrupt isn't very helpful

In development, I accidentally had two tasks trying to claim the same hardware interrupt:

[tasks.one]
interrupts = {"usart1.irq" = 0b01}
# ...

[tasks.two]
interrupts = {"usart1.irq" = 0b10}
# ...

This correctly does not build, but we should fail more nicely than we do:

   Compiling kern v0.1.0 (/home/john/github/hubris/sys/kern)
error: failed to run custom build command for `kern v0.1.0 (/home/john/github/hubris/sys/kern)`

Caused by:
  process didn't exit successfully: `/home/john/github/hubris/target/release/build/kern-c72a9bb98816b09a/build-script-build` (exit status: 101)
  --- stdout
  cargo:rustc-cfg=armv7m
  cargo:rerun-if-env-changed=HUBRIS_SECURE
  cargo:rerun-if-env-changed=HUBRIS_IMAGE_ID
  cargo:rerun-if-env-changed=HUBRIS_KCONFIG

  --- stderr
  thread 'main' panicked at 'called `Result::unwrap()` on an `Err` value: Cannot build a perfect hash with duplicate keys', sys/kern/build.rs:304:22
  note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace
Error: command failed, see output for details

Reproduce the duplicate interrupt configuration from the issue and inspect sys/kern/build.rs around line 304, where the duplicate-key failure is reported. Improve the diagnostic so it identifies the conflicting task or interrupt configuration instead of exposing the perfect-hash panic, then verify the build fails with the clearer message.