On Wed, May 28, 2025 at 4:27 PM Patrick Palka wrote:
>
> >
> >
> > On Tue, May 20, 2025 at 6:32 PM Patrick Palka wrote:
> > On Tue, 20 May 2025, Tomasz Kaminski wrote:
> >
> > > I think I do not have any more suggestions for cases to check,
> so the impl LGTM.
> >
> > It's cool
>
>
> On Tue, May 20, 2025 at 6:32 PM Patrick Palka wrote:
> On Tue, 20 May 2025, Tomasz Kaminski wrote:
>
> > I think I do not have any more suggestions for cases to check, so the
> impl LGTM.
>
> It's cool how many optimizations we came up with for this algorithm :)
>
>
On Tue, May 20, 2025 at 6:32 PM Patrick Palka wrote:
> On Tue, 20 May 2025, Tomasz Kaminski wrote:
>
> > I think I do not have any more suggestions for cases to check, so the
> impl LGTM.
>
> It's cool how many optimizations we came up with for this algorithm :)
>
> >
> > On Tue, May 20, 2025 at
On Tue, 20 May 2025, Tomasz Kaminski wrote:
> I think I do not have any more suggestions for cases to check, so the impl
> LGTM.
It's cool how many optimizations we came up with for this algorithm :)
>
> On Tue, May 20, 2025 at 4:33 PM Patrick Palka wrote:
> Changes in v5:
> * d
I think I do not have any more suggestions for cases to check, so the impl
LGTM.
On Tue, May 20, 2025 at 4:33 PM Patrick Palka wrote:
> Changes in v5:
> * dispatch to starts_with for the both-bidi/common range case
>
> Changes in v4:
> * optimize the both-bidi/common ranges case, as suggeste
On Tue, 20 May 2025, Patrick Palka wrote:
> Changes in v5:
> * dispatch to starts_with for the both-bidi/common range case
Forgot to mention:
And for a bidi common haystack, prefer iterating forward instead of
backward if the needle is at least half the size of the haystack,
>
> Changes in v
Changes in v5:
* dispatch to starts_with for the both-bidi/common range case
Changes in v4:
* optimize the both-bidi/common ranges case, as suggested by
Tomasz
* add tests for that code path
Changes in v3:
* Use the forward_range code path for a (non-sized) bidirectional
haystack,
