Fix internally tracked sibling positions earlier when hydrating

[WorldSEnder]

Previously, the fixup of next_siblings was done in the first "priority" render, now this happens immediately in the render that transitions from ComponentRenderState::Hydration to ComponentRenderState::Render.

We do so by threading through a reference to the sibling that needs a fix

Fixes #3913

Checklist

• I have reviewed my own code
• I have added tests

[github-actions]

Visit the preview URL for this PR (updated for commit de85d40):

https://yew-rs-api--pr3914-fix-early-rerender-daywzpi8.web.app

_{(expires Tue, 16 Sep 2025 10:11:03 GMT)}

_{🔥 via Firebase Hosting GitHub Action 🌎}

[github-actions]

Benchmark - core

Yew Master

vnode           fastest       │ slowest       │ median        │ mean          │ samples │ iters
╰─ vnode_clone  2.463 ns      │ 3.721 ns      │ 2.469 ns      │ 2.809 ns      │ 100     │ 1000000000

Pull Request

vnode           fastest       │ slowest       │ median        │ mean          │ samples │ iters
╰─ vnode_clone  2.462 ns      │ 3.733 ns      │ 2.466 ns      │ 2.666 ns      │ 100     │ 1000000000

[github-actions]

Benchmark - SSR

Yew Master

Details

Benchmark	Round	Min (ms)	Max (ms)	Mean (ms)	Standard Deviation
Baseline	10	310.657	318.801	312.297	2.579
Hello World	10	464.129	482.973	472.466	5.242
Function Router	10	1603.743	1651.915	1630.303	14.265
Concurrent Task	10	1006.374	1008.145	1007.113	0.599
Many Providers	10	1099.507	1149.708	1120.262	14.000

Pull Request

Details

Benchmark	Round	Min (ms)	Max (ms)	Mean (ms)	Standard Deviation
Baseline	10	312.367	313.603	312.939	0.497
Hello World	10	469.491	473.541	471.351	1.415
Function Router	10	1627.290	1640.471	1634.880	3.776
Concurrent Task	10	1006.600	1007.758	1007.074	0.447
Many Providers	10	1086.938	1133.868	1120.437	13.072

[github-actions]

Size Comparison

Details

examples	master (KB)	pull request (KB)	diff (KB)	diff (%)
async_clock	99.403	99.387	-0.017	-0.017%
boids	168.695	168.756	+0.061	+0.036%
communication_child_to_parent	91.689	91.724	+0.034	+0.037%
communication_grandchild_with_grandparent	102.709	102.796	+0.087	+0.085%
communication_grandparent_to_grandchild	97.718	97.805	+0.087	+0.089%
communication_parent_to_child	87.607	87.642	+0.034	+0.039%
contexts	103.773	103.939	+0.166	+0.160%
counter	85.107	85.091	-0.017	-0.020%
counter_functional	85.279	85.263	-0.017	-0.019%
dyn_create_destroy_apps	87.906	87.922	+0.016	+0.018%
file_upload	98.469	98.452	-0.017	-0.017%
function_delayed_input	91.740	91.724	-0.017	-0.018%
function_memory_game	170.643	170.835	+0.192	+0.113%
function_router	337.664	338.255	+0.591	+0.175%
function_todomvc	163.279	163.419	+0.140	+0.086%
futures	238.370	238.354	-0.017	-0.007%
game_of_life	104.434	104.417	-0.017	-0.016%
immutable	194.620	194.705	+0.085	+0.044%
inner_html	79.994	79.978	-0.017	-0.021%
js_callback	108.272	108.330	+0.058	+0.053%
keyed_list	196.418	196.452	+0.034	+0.017%
mount_point	83.318	83.302	-0.017	-0.020%
nested_list	115.036	115.123	+0.087	+0.076%
node_refs	90.795	90.829	+0.034	+0.038%
password_strength	1776.479	1776.513	+0.034	+0.002%
portals	93.260	93.294	+0.034	+0.037%
router	307.312	307.707	+0.395	+0.128%
suspense	112.228	112.338	+0.110	+0.098%
timer	88.680	88.663	-0.017	-0.019%
timer_functional	95.011	95.045	+0.034	+0.036%
todomvc	143.627	143.610	-0.017	-0.012%
two_apps	85.722	85.705	-0.017	-0.019%
web_worker_fib	134.766	134.829	+0.063	+0.047%
web_worker_prime	187.477	187.540	+0.063	+0.034%
webgl	82.934	82.917	-0.017	-0.020%

✅ None of the examples has changed their size significantly.

[WorldSEnder]

Small explanation why the previous approach was not always correctly working:
Both the initial hydrating render and the priority re-render work top-down. When trusting the latter to fix-up internal reference, these could still be wrong for child components that haven't gotten their priority re-render yet. This is only an issue if this first re-render causes elements to get moved around or added, which is usually not the case (but also not wrong per se, as by that point hydration is over).