Fix conversion of non-utf8 sequences to AnyValue

[Nevay]

Converting to AnyValue:

String values which are not valid Unicode sequences SHOULD be converted to AnyValue's bytes_value with the bytes representing the string in the original order and format of the source string.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 84.62%. Comparing base (bb07aca) to head (a9b76b2).
Report is 1 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff            @@
##               main    #1253   +/-   ##
=========================================
  Coverage     84.62%   84.62%           
- Complexity     2136     2140    +4     
=========================================
  Files           284      284           
  Lines          6054     6062    +8     
=========================================
+ Hits           5123     5130    +7     
- Misses          931      932    +1     

Flag	Coverage Δ
8.0	84.57% <100.00%> (+<0.01%)	⬆️
8.1	84.60% <100.00%> (+<0.01%)	⬆️
8.2	84.60% <100.00%> (+<0.01%)	⬆️
8.3	84.60% <100.00%> (+<0.01%)	⬆️
8.4	84.60% <100.00%> (+<0.01%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Files	Coverage Δ
src/Contrib/Otlp/AttributesConverter.php	100.00% <100.00%> (ø)

... and 1 file with indirect coverage changes

---

Continue to review full report in Codecov by Sentry.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update bb07aca...a9b76b2. Read the comment docs.

[agoallikmaa]

Technically this condition might not be necessary since symfony/polyfill-mbstring is pulled in transitively from the SDK. Not sure if it's best practice to rely on that though.

[brettmc]

@Nevay should we rely on the polyfill, or leave it as-is. Otherwise, I'm happy to approve and merge.

[Nevay]

We should leave it as-is. The polyfill is slower and exports can contain a large number of attributes.
Creating and exporting spans with 10 string attributes (16 bytes ea.; 8 valid and 2 invalid utf8 sequences) to local collector w/ default batch size was around 10% slower with the polyfill compared to the preg_match() fallback (fallback was incorrect until latest push).

Benchmark results for creating AnyValue from a 16 bytes string (polyfill is esp. costly for invalid utf8 sequences due to its iconv fallback):

+------------------------------------+---------+---------+--------+---------+
| subject                            | memory  | mode    | rstdev | stdev   |
+------------------------------------+---------+---------+--------+---------+
| benchCheckEncoding (valid)         | 1.959mb | 0.245μs | ±1.95% | 0.005μs |
| benchCheckEncoding (invalid-begin) | 1.959mb | 0.248μs | ±1.68% | 0.004μs |
| benchCheckEncoding (invalid-end)   | 1.959mb | 0.250μs | ±1.94% | 0.005μs |
| benchPregMatch (valid)             | 1.959mb | 0.307μs | ±2.06% | 0.006μs |
| benchPregMatch (invalid-begin)     | 1.959mb | 0.262μs | ±1.53% | 0.004μs |
| benchPregMatch (invalid-end)       | 1.959mb | 0.270μs | ±1.62% | 0.004μs |
| benchIsUtf8 (valid)                | 1.959mb | 0.250μs | ±1.62% | 0.004μs |
| benchIsUtf8 (invalid-begin)        | 1.959mb | 0.252μs | ±1.64% | 0.004μs |
| benchIsUtf8 (invalid-end)          | 1.959mb | 0.253μs | ±1.31% | 0.003μs |
| benchPolyfill (valid)              | 1.959mb | 0.383μs | ±1.67% | 0.006μs |
| benchPolyfill (invalid-begin)      | 1.959mb | 0.830μs | ±1.51% | 0.013μs |
| benchPolyfill (invalid-end)        | 1.959mb | 0.945μs | ±1.51% | 0.014μs |
+------------------------------------+---------+---------+--------+---------+