Siddhesh Poyarekar

I'm afraid I've never looked at MIPS. The backtrace doesn't look like it has anything to do with luajit unless there is code in there that affects the `ngx_http_lua_pcre_malloc_done` call....

I've got the ppc64le port PR #77 for luajit2 which includes some fixes on top of the ppc64le patches that were in Debian, etc. months ago. Can you please try...

Thanks for confirming. It'll try to look into it this weekend.

I skimmed through the patches and overall they seem to be in the direction I wanted to take, thanks! I can do a more thorough review in some time if...

Also as a side note, I am beginning to wonder how much we need to bother with merge compatibility with the original LuaJIT project. It has been almost a year...

> Hi all, I have been away from the computer for some time. My turn to apologize; I'm still kinda away but stole some time for this :) > That...

My understanding is that it's not the goal (hence my request to remove the `dynamic` word from the PR title) of this PR to have the same binary work on...

That's what #75 does; you can test it in action in moonjit. I think @thibaultcha has a plan to stage the changes in though (I don't fully understand how things...

@thibaultcha I'm afraid that still does not help. I just skimmed through the patchset and building the entire binary with -msse4.2 is incorrect because it will not run on CPUs...

@thibaultcha I haven't had a chance to do a thorough review, but a quick skim gives me the impression that this should work correctly. I'll sync this into moonjit this...