[stdlib] String index interchange, etc.#9806
Conversation
|
@swift-ci Please test source compatibility |
Build comment file:Optimized (O) Regression (6)
Improvement (9)
No Changes (257)
Regression (4)
Improvement (8)
No Changes (260)
|
|
Wow, some of those speedups are really surprising. |
|
@swift-ci Please smoke benchmark |
|
@swift-ci Please test source compatibility |
|
All those "Hearts <3 ♥︎ 💘" and "Café 🍵"s will be lost in time, like tears in rain? |
|
@palimondo Gonna assume that's a joke that went over my head. If it's not, please try to explain. |
|
That was a Blade Runner reference... and I was wondering if the removed documentation that was demonstrating use of |
Build comment file:Build failed before running benchmark. |
|
Oh, removed for good because these types and therefore overloads are collapsing with others. However I believe the same strings exist on the overloads with which they are collapsing, unless @natecook1000 was exceptionally, gratuitously creative ;-) |
818bdc2 to
652c417
Compare
|
I take it back; @natecook1000 was creative with good reason and many of them do drop; He'll have to rewrite the docs for these things anyway. |
|
@swift-ci Please test source compatibility |
Build comment file:Build failed before running benchmark. |
652c417 to
5bcb52b
Compare
|
@swift-ci Please test |
|
@swift-ci Please benchmark |
|
@swift-ci Please test source compatibility |
|
Build failed |
|
Uh, right. The linux failure is because I took away some SPI that Foundation was using for the Linux case only. It's easy to bring back but I want to let the other tests finish so I didn't push it here. |
Build comment file:Optimized (O) Regression (18)
Improvement (4)
No Changes (250)
Regression (9)
Improvement (9)
No Changes (254)
|
|
|
||
| public init(encodedOffset o: Int) { | ||
| _compoundOffset = UInt64(o << _Self._strideBits) | ||
| /// Creates a new index at the specified UTF-16 offset. |
There was a problem hiding this comment.
FYI, I aim to decouple this stuff from UTF-16 eventually. That is really meant to be a code unit offset in some arbitrary encoding. I'd have labeled it utf16Offset otherwise.
|
@swift-ci Please benchmark |
|
@swift-ci Please test |
|
Build failed |
|
Build failed |
Build comment file:Optimized (O) Regression (10)
Improvement (21)
No Changes (241)
Regression (5)
Improvement (17)
No Changes (250)
|
|
@swift-ci Please benchmark |
|
@swift-ci Please test |
bc48148 to
4d97cef
Compare
|
Please test with following pull request: @swift-ci Please test source compatibility |
|
Please test with following pull request: @swift-ci Please test |
|
Build failed |
|
Build failed |
4d97cef to
d21259c
Compare
|
Please test with following pull request: @swift-ci Please test source compatibility |
|
Please test with following pull request: @swift-ci Please test |
|
Build failed |
This works either way I suppose, but the code is clearer now.
The fancy code didn't turn out to be any faster.
Only the last of each set of dups gets run anyway. See https://bugs.swift.org/browse/SR-5362
test the new APIs as well as the legacy failing conversions.
d21259c to
a2db605
Compare
|
Please test with following pull request: @swift-ci Please test |
I failed to merge the upstream changes to swift-corelibs-foundation at the same time as I merged that #9806, and it broke on linux. Going to get it right this time.
No description provided.