Re: non-functional backout for review
- From: Andrey Chernov <ache@xxxxxxxxxxxx>
- Date: Thu, 1 Nov 2007 06:19:22 +0300
On Thu, Nov 01, 2007 at 11:15:21AM +0800, David Xu wrote:
Andrey Chernov wrote:
Some people complaints that new check in __isctype() is not human readable
and requests for backout. Compiler gives almost identical code for old and
new excepting non-optimized case where non-human readable one wins. I am a
bit tired to change it forth and back, so ask for final consensus here.
Old one
return (_c < 0 || _c >= 128) ? 0 :
New one (requests for backout)
return (_c & ~0x7F) ? 0 :
I will use a version which compiler can generate optimized code.
the name __isctype already tells what the function is doing. :-)
Already done.
--
http://ache.pp.ru/
_______________________________________________
freebsd-current@xxxxxxxxxxx mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-current
To unsubscribe, send any mail to "freebsd-current-unsubscribe@xxxxxxxxxxx"
- References:
- non-functional backout for review
- From: Andrey Chernov
- Re: non-functional backout for review
- From: David Xu
- non-functional backout for review
- Prev by Date: Re: non-functional backout for review
- Next by Date: Re: csup broken on -currnet amd64?
- Previous by thread: Re: non-functional backout for review
- Next by thread: Re: non-functional backout for review
- Index(es):
Relevant Pages
|
