Compilation of app_uart fails on strip rust_begin_unwind

[kristofferkoch]

Thanks for making zinc! I find the platform and ideas really exciting, and I'm working on trying to understand how zinc is put together and how to make nice things.

However, when compiling with rake PLATFORM=lpc17xx build_all, compilation fails at app_uart with:

rustc --opt-level 2 -Z no-landing-pads -g --cfg cfg_mcu_has_spi --cfg mcu_lpc17xx --cfg arch_cortex_m3 --target thumbv7m-linux-eabi -Ctarget-cpu=cortex-m3 -C relocation_model=static  -C lto  --emit obj  -L /home/kko/projects/zinc/build  -o /home/kko/projects/zinc/build/intermediate/uart/app_uart.o  /home/kko/projects/zinc/apps/app_uart.rs 
arm-none-eabi-strip -N rust_stack_exhausted -N rust_begin_unwind -N rust_eh_personality /home/kko/projects/zinc/build/intermediate/uart/app_uart.o
arm-none-eabi-strip: not stripping symbol `rust_begin_unwind' because it is named in a relocation

I'm running on debian/unstable with latest rust.

[farcaller]

libcore/failure.rs is too heavyweight and pulls in fmt.

[farcaller]

(it happens in UART as this is apparently the only demo with array range checks and division, that can fail in runtime)

[farcaller]

Seems that it's not something we can fix, filed rust-lang/rust#17886.

[farcaller]

For the reference, this is where things go wrong: https://gist.github.com/farcaller/282860631f1757e2a17d#file-main-c-L328

[kristofferkoch]

Thank you for your analysis, farcaller!

[farcaller]

I've submitted 05c8f3c as a temporary workaround.

[mcoffin]

This isn't really a thing anymore as fail was changed to panic upstream and some of how the lang items work was changed. Upstream bug was closed.