Sometimes I build with -fno-inline to simplify debugging. (ARM GDB doesn't cope well with inlined functions.) However, with a recent mozilla-central (569a45bfb71c), I get the following error on several tests: Assertion failure: LIR AccSet error (end of writer pipeline): 'sti' AccSet is '.slots'; base = (addi1 = addi ldi1, immi1/*0x40000*/); disp = 0: 0 (/work/moz/mc/js/src/nanojit/LIR.cpp:3599) The problem goes away if I don't use -fno-inline.
Note that if NanoJIT is going away sooner than expected (to be replaced with type inference), this bug might not be worth pursuing, especially as it only affects my own debugging builds.
That is *strange*. I wonder if it's a GCC bug?
Nanojit has been dead for several years. Its Bugzilla component has been moved to the graveyard (bug 984276). I checked all the open bugs. They're all uninteresting, so I'm WONTFIXing them all. Apologies for the bugspam.