Re: ARM toolchain recommendation
+++ Craig Hollabaugh [04-06-02 11:11 -0600]:
> On Wed, 2004-06-02 at 10:25, Wookey wrote:
> > Yes 3.2 was not good (on arm at least). 3.3 is much better and even mostly
> > works :-)
> Just curious, what kind of cross-compiler testing do you do to know that
> it "mostly works"?
Well, proper testing is obviously a big job, but I've found that if I can
compile a kernel, a complicated bootloader like bootldr, and some
non-trivial C and C++ apps with the same compiler and they all work then I
call that 'mostly working'. Running the built-in tests on gcc is probably
actually a better test but for various reasons I've never actually tried that.
> I've always wondered what the structured test is,
> particularly for a cross-compiler.
I think 'using the gcc test suite' is probably the right answer. IIRC Dan
Kegel's crosstool scripts help with the testing of this in a cross
What do others do?
Aleph One Ltd, Bottisham, CAMBRIDGE, CB5 9BA, UK Tel +44 (0) 1223 811679
work: http://www.aleph1.co.uk/ play: http://www.chaos.org.uk/~wookey/