Vlado Handziski wrote: > Yes, it looks bad. 3.2.3 is still the best option, and that is what we > are > using here at TU Berlin. This gen_lowpart error is not always occurring. It > is hard to detect what exactly is going wrong. For example, when compiling > nucleus for the eyesIFX platform (mainly different communication stack) with > the gcc-3.2.3 there is no error, but it breaks the telos builds.
Do you have any advice you could give me for the case when the build breaks? What do you do when it happens on your side? Is it only due to the code size, or is it due to specific coding constructs? Or does it just "go away" when you further change the code? The MSP430 platform is pretty useless if it's only possible to compile small programs. I hope gcc-3.4 will be better, but so far I was only able to have it segfault when compiling TinyOS applications. Thanks and best regards. -- Remy Remove underscore and suffix in reply address for a timely response.