All the passes that run by default in -O2 or -O3 have been converted. Also the frounds pass. Otherwise, all the rest hasn't been, like safeHeap, outline, emterpretify, etc. However, it isn't urgent to convert those, as the compiler can interleave native and js passes, so even in -O2 -s SAFE_HEAP=1, it will run all the passes fast except for the one safeHeap pass. So you would still get almost all of the speedup. Also, it's easier to prototype new passes in JS, so I don't think we'll ever get rid of the JS optimizer.
Hmm, I believe there aren't any major changes like this currently planned. Overall the compiler is in pretty good shape now as far as compilation speed goes, I think. The next big code change will probably be updating to LLVM 3.5 and beyond. - Alon On Wed, Dec 17, 2014 at 2:39 PM, bobajeff <[email protected]> wrote: > > Wow, that was fast. Bravo. Are there any remaining passes that need to be > converted to C++? Also, are there any other major changes like this planned? > > -- > You received this message because you are subscribed to the Google Groups > "emscripten-discuss" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to [email protected]. > For more options, visit https://groups.google.com/d/optout. > -- You received this message because you are subscribed to the Google Groups "emscripten-discuss" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/d/optout.
