Time#strftime: avoid repeated coderange scanning

Followup: a3f589640fd443deea230c27efd6bdfc92f5817f

Restarting scanning has a fixed cost. Since we build a string
from many small parts, that causes the scan to restart many times.

It's much faster to ignore the coderange, if it's needed later it
will be faster to scan it in one go.

```
compare-ruby: ruby 3.4.0dev (2024-09-04T11:54:06Z opt-strftime-growth 149480f0ff) +YJIT [arm64-darwin23]
built-ruby: ruby 3.4.0dev (2024-09-04T11:55:26Z opt-strftime-coder.. 29a5153cba) +YJIT [arm64-darwin23]
warming up...

|                            |compare-ruby|built-ruby|
|:---------------------------|-----------:|---------:|
|time.strftime("%FT%T")      |      1.870M|    2.225M|
|                            |           -|     1.19x|
|time.strftime("%FT%T.%3N")  |      1.731M|    2.068M|
|                            |           -|     1.19x|
|time.strftime("%FT%T.%6N")  |      1.744M|    2.073M|
|                            |           -|     1.19x|
```

Cumulative with a3f589640fd443deea230c27efd6bdfc92f5817f:

```
compare-ruby: ruby 3.4.0dev (2024-09-04T11:55:26Z opt-strftime-coder.. 29a5153cba) +YJIT [arm64-darwin23]
built-ruby: ruby 3.4.0dev (2024-09-04T12:57:30Z opt-strftime-coder.. 2b938d667a) +YJIT [arm64-darwin23]
warming up...

|                            |compare-ruby|built-ruby|
|:---------------------------|-----------:|---------:|
|time.strftime("%FT%T")      |      1.784M|    2.277M|
|                            |           -|     1.28x|
|time.strftime("%FT%T.%3N")  |      1.504M|    2.056M|
|                            |           -|     1.37x|
|time.strftime("%FT%T.%6N")  |      1.489M|    2.094M|
|                            |           -|     1.41x|
```

0 files changed, 0 insertions, 0 deletions