Message146755
> > Note we only really see the effect if we make sure that gcc
> > isn't emitting its "special" memcmp: that's why the -fno-builtin-memcmp
> > is SO important on gcc builds!
>
> I'd rather infer the opposite: given how GCC generates code, this patch
> is not worthwhile. Given that it actually slows down Python in the
> default system configuration, I'm -1 on applying it. This is really not
> a route we should take; it leads to maintenance pain.
I agree with Martin. This patch would be very nice if there wasn't the
memcmp() perf problem. A possible solution would be to write a simple
optimized memcmp()-alike for our own purposes.
But I'm not sure it's really worth the hassle: comparing long unicode
strings doesn't strike me as a very common operation. Finding a short
substring, conatenating strings together, are all much more common. |
|
Date |
User |
Action |
Args |
2011-10-31 23:47:44 | pitrou | set | recipients:
+ pitrou, loewis, rhettinger, scoder, vstinner, asmodai, ezio.melotti, petri.lehtinen, RichIsMyName |
2011-10-31 23:47:43 | pitrou | link | issue13279 messages |
2011-10-31 23:47:43 | pitrou | create | |
|