Parser bug: Invalid token, cannot triage about emscripten

[kripken]

The problem is the "module asm" lines, which are inline assembly. Emscripten can't translate CPU-specific code like that.

The source code you are building might have an option to not use inline assembly.

Using e.g. dragonegg this is done to all bc files. If you know a way how to prevent this embedding it would surely be appreciated.

[kripken]

What command are you using to compile with dragonegg? Might be related to optimization somehow.

Try to compile with emmaken.py with dragonegg, that might help since it will then use the emscripten headers and compile flags, something like

EMMAKEN_COMPILER=dragonegg python emmaken.py myfile.cpp -o myfile.o

I cannot use emmaken since I'm using SCons and doing a lot of configuring in SCons.

[kripken]

I believe you should be able to tell SCons to use a different compiler than the platform's default. Looks like this has the right syntax:

http://stackoverflow.com/questions/2944251/changing-default-compiler-in-linux-using-scons

So,

env = Environment()
env.Replace(CC = "PATH-TO-EMMAKEN")

Yeah I know and I do this quite extensively. I have already a python script in place so I'll have to see how I can access emmaken.py without making the whole thing any slower than it already is.
Which means I have to have a good look into emmaken.py, which has currently very low priority since I try to get the whole mess compiling again, first.