Tom Tromey
Tom Tromey
I was going to debug this today, but when I reload the page the times all seem to be correct. So I'm not sure what is going on; they were...
The issue is there's no lower-level IR which can represent unboxed values.
I pushed a patch to my emacs branch adding a `:cl-block-p` property to the catch symbol, with a value of `t`. The compiler side has yet to be done.
I hadn't really planned to do that but I'm not opposed to it.
Yeah, that's totally doable. Though I have a more radical idea, which is to use [Emacs FFI](https://github.com/tromey/emacs-ffi) to pull in [gcc-jit](https://github.com/tromey/emacs-gcc-jit) and then lower the el-compilador IR to jitted code....
I'm completely fine with submitting it for inclusion. I have assignment papers; and I think probably the other contributions are small enough not to need them. John W. asked on...
> I would also have to think about how a ffi.el that has been installed from Melpa would build ffi-module.c and how to communicate to the user that they have...
Now that Emacs has bigints, this could be fixed.
Emacs isn't single-threaded any more. Errno could be handled specially by the library. That wouldn't be totally unreasonable. If other libraries export TLS symbols then we'd need to think of...
Hi! I'm currently using the GCC that comes with Fedora 21. That is GCC 4.9.2. But really pretty much any version should work. Dynamic modules are available only on the...