Thomas Fransham

Userdata is also in a separate chain from g->root, its head is instead g->mainthref->nextgc. I don't really know what purpose g->mainthref once served but its always assigned (&GG->L), I know...

One changes needed for 2.1 bytecode is calls on GC64 builds need an extra stack slot for the two slot call frames.

It crashing in the wrapper generated for the intrinsic i think. I remember rebasing on a branch with gc64 before fixing tons of merge conflicts but maybe I missed adding...

One thing to quickly do is check all the tests still pass with your build running tests/runtests.sh some of them were upgraded to test vectors in the JIT for the...

That branch won't build its just preparation work to combine my work with [corsixs branch](https://github.com/corsix/LuaJIT/commits/newgc). His branch kind of works but has some crash bugs.

It still missing lots of things and I got no reply from Mike Paul when I emailed him months ago

> Is readonlytables the right branch? I also see readonly_tables. I thought I hadn't pushed it from a separate repo before. > Any loose ends in the implementation that I...

This happens with plain LuaJIT as well, only -O-fold stops it out of all the optimization levels. It looks like the wrong string pointer is being used for pointer arithmetic...

As a workaround I just ended up modifying _premake_main.lua to wrap the entry point in a xpcall to get a full stack trace ``` lua local function main() p.callArray(m.elements) end...