>>>>> On Thu, 10 Nov 2022 13:18:56 +0200, Eli Zaretskii <eliz@gnu.org> said:
>> From: Mattias Engdegård <mattiase@acm.org>
>> Date: Thu, 10 Nov 2022 10:52:40 +0100
>> Cc: casouri@gmail.com, emacs-devel@gnu.org
>>
>> 9 nov. 2022 kl. 14.06 skrev Eli Zaretskii <eliz@gnu.org>:
>>
>> >> You are right to worry about it, but there is actually no need for
>> >> concern here: it's a tail call (and the extra argument is last) so
>> >> it should compile to an unconditional jump (and setting a register).
>> >
>> > That's not what I see here, even with -O2.
>>
>> Trying to read your mind, are you using 32-bit x86? Even that
>> shouldn't be disastrous; it's just some stack manipulation. Hardly
>> noticeable given that it's a regexp match that comes next.
Eli> I agree that the effect is probably minuscule, but it's still there.
Eli> And on some architectures it could be more than that.
Eli> Why not have two separate functions for these two jobs? The code is
Eli> short, and some duplication is not an issue in this case.
Or one generic function and two #defines that pass in the appropriate
canonicalize table value.
Robert
--
RetroSearch is an open source project built by @garambo | Open a GitHub Issue
Search and Browse the WWW like it's 1997 | Search results from DuckDuckGo
HTML:
3.2
| Encoding:
UTF-8
| Version:
0.7.4