> Then this is clearly a gcc bug. You might want to verify that yourself, before making that claim.
I've already seen handfuls of bugs in other software that gcc-4.8.0 kicks up, that 4.7.x did not. - so unless you are _positive_ it's a gcc bug ~ i wouldn't be asserting that it is, since it's also a possibility that your code (like any code) may contain bugs.. just my 2 cents Jordan > On Sun, Apr 21, 2013 at 3:45 PM, jordan <[email protected]> wrote: >>> The danger of -O3 (long gone, to my knowledge) is to the compiled >>> code, not the compiler. An internal compiler error is just that (or >>> faulty hardware). >> >> James (the OP) never mentioned what version of gcc he is using, but if >> it is 4.8.0 (which i expect might be the case) then yes, NON- code >> fails to compile... and you are also wrong to suggest stricter and >> heavier optimization levels won't potentially break compilation or >> cause undesired behavior (ie: "danger"), it does and will, depending >> on a variety of factors. >> >> anyway, James is correct, I have verified exactly what he has written >> ~ using -O2 allows non-daw and friends to compile (with gcc-4.8.0). >> >> cheers >> >> jordan >> >>> On Sun, Apr 21, 2013 at 3:07 PM, James Morris <[email protected]> wrote: >>>> >>>> hi, >>>> >>>> just a note. i tried to build the non-* stuff and it failed with a >>>> internal compiler error. saw that you're using -O3 optimazation which i >>>> know is (from when i once used gentoo) frowned upon in some circles. i >>>> disabled it (via nano as i don't know waf well enough to do otherwise) >>>> and was able to build all the non-* stuff without further-ado. >>>> >>>> sorry i can't test any of this stuff out further, just wanted a quick >>>> peek. sadly no time for doing anything requiring deep involvement on pc >>>> these days. >>>> >>>> james. >>>> >>>> >>> >>> >> >> > >
