[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Emacs Lisp JIT Compiler
From: |
T.V Raman |
Subject: |
Re: Emacs Lisp JIT Compiler |
Date: |
Mon, 13 Aug 2018 08:23:02 -0700 |
User-agent: |
Gnus/5.13 (Gnus v5.13) Emacs/27.0.50 (gnu/linux) |
Tom Tromey <address@hidden> writes:
Good to know - -asked because of the reference in your original note to
"redefined functions" (speaking from memory)>>>>>> "T" == T V Raman
<address@hidden> writes:
>
> T> Will it work with adviced functions?
>
> Yes. The JIT works by adding another slot to the bytecode object.
> This slot holds the JIT-related data, like the compiled function.
> So, things like advice don't affect the way it operates.
>
> Writing this out made me remember another possible issue with the JIT:
> because it works based on the bytecode object, when a closure is made
> with make-byte-code, the JIT will compile each such closure separately.
> It would be possible to, instead, have it compile a closure once; but at
> some cost to the generated code.
>
> Tom
>
--
Id: kg:/m/0285kf1
- Re: Emacs Lisp JIT Compiler, (continued)
- Re: Emacs Lisp JIT Compiler, Eli Zaretskii, 2018/08/23
- Re: Emacs Lisp JIT Compiler, Tom Tromey, 2018/08/24
- Re: Emacs Lisp JIT Compiler, Eli Zaretskii, 2018/08/24
- Re: Emacs Lisp JIT Compiler, Tom Tromey, 2018/08/24
- Re: Emacs Lisp JIT Compiler, Eli Zaretskii, 2018/08/25
Re: Emacs Lisp JIT Compiler, Tom Tromey, 2018/08/15
Re: Emacs Lisp JIT Compiler, T.V Raman, 2018/08/13
Re: Emacs Lisp JIT Compiler, Eli Zaretskii, 2018/08/13
Re: Emacs Lisp JIT Compiler, Richard Stallman, 2018/08/13
Re: Emacs Lisp JIT Compiler, Clément Pit-Claudel, 2018/08/14