[darcs-users] darcs patch: fpstring.c: switch a memchr for memrchr (and 2 more)

[Gwern Branwen]

On 2008.04.28 14:46:48 -0700, Jason Dagit <dagit at codersbase.com> scribbled 2.0K characters:
>    On Mon, Apr 28, 2008 at 2:05 PM, <gwern0 at gmail.com> wrote:
>
>      Fri Apr 25 16:01:53 EDT 2008  gwern0 at gmail.com
>       * fpstring.c: switch a memchr for memrchr
>       See <http://bugs.darcs.net/issue814>; memrchr speeds up is_funky quite a bit and thus helps
>      whatsnew -s. It doesn't seem to break (any more) tests.
>
>    Now that we're looking at less of the data does memrchr really provide a speed up?  I'm also
>    concerned about what Don Stewart said about the portability of memrchr.  I don't see a
>    portability fix included in your patch bundle.  So, my recommendation would be to wait on the
>    memrchr addition.
>
>    The rest looks good.
>    Jason

Oh, you misunderstand - the first patch changes it to memchr/memrchr, but the second one restores the status quo and simultaneously moves the 4096 truncation up the stack to is_bin. So if the bundle is applied, it'll still be memchr/memchr, as portable as usual.

If the patch is a null-op, then you might be forgiven for wondering why I included it. Well, two main reasons: it's nice to have a record of other ideas and approaches (eg. if someone wants to see how much of a difference memchr/memrchr would make now, they can just rollback a few patches), it's a very small patch, and I was too lazy to do whatever magic would be necessary to make the two important patches apply straight to the Darcs darcs repo without depending on't.

--
gwern
ASIO A/C undercover Audiotel Global domestic Internet Sayeret Canine Cap-Stun
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
Url : http://lists.osuosl.org/pipermail/darcs-users/attachments/20080428/033e192e/attachment.pgp