[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: please review new branch feature/byte-unwind-protect
From: |
Tom Tromey |
Subject: |
Re: please review new branch feature/byte-unwind-protect |
Date: |
Tue, 23 Jan 2018 08:50:33 -0700 |
User-agent: |
Gnus/5.13 (Gnus v5.13) Emacs/26.0.91 (gnu/linux) |
>>>>> "John" == John Wiegley <address@hidden> writes:
>>>>> "TT" == Tom Tromey <address@hidden> writes:
TT> I recently wrote a couple of patches to add two new bytecodes to Emacs.
TT> These make it possible to compile unwind-protect without the need to
TT> introduce a closure for the unwind forms.
John> Hi Tom,
John> Thanks for these changes. I just want to ask: what motivates them? Is it
John> efficiency? Does it enable something you can't achieve otherwise? If the
John> former, has it been measured? What is the corresponding cost of having
this
John> change?
It's partly for efficiency -- I didn't benchmark it though. I suppose I
can. It's pretty normal to see unwind-protect in the expansion of
macros, though; there are plenty in subr.el. This enables better code
generation in the jit I've written (I plan to post about that when it's
ready but I can do it sooner if you want).
Also this seems like an un-finished to-do item from the lexical binding
conversion.
I don't think there's a cost to this change. It uses two byte codes,
but there are still some to spare, so I think it isn't a big deal.
Tom