[avr-llvm-devel] AVR LLVM bug

[Dylan M. <dyl...@gm...>]

I am working on a fork of this repository (located here
<https://github.com/dylanmckay/avr-llvm>), which I am adding features,
fixing bugs, and implementing missing functionality.

I have run into a bug which causes compilation to fail if all optimisations
are turned off (see issue #15
<https://github.com/dylanmckay/avr-llvm/issues/15> on GitHub). You can work
around the bug by setting the register allocator to 'fast'.

I have found a mailing list
<https://groups.google.com/d/msg/llvm-dev/UeIaa0Z74lw/w4EAY1HGWgMJ> mentioning
the problem, but I can't find any real documentation on it.

Does anybody remember any details about it, or is there anywhere that I'm
missing where it may be documented?