Repository navigation
Force using the ROOT locale when formatting to avoid involving the default locale when numbers are converted to strings - #16025
Conversation
…locale when numbers are converted to strings Signed-off-by: Ludovic Orban <lorban@bitronix.be>
…locale when numbers are converted to strings Signed-off-by: Ludovic Orban <lorban@bitronix.be>
|
This PR seems to want to correct our |
|
@joakime what |
Signed-off-by: Ludovic Orban <lorban@bitronix.be>
It felt like we were missing the Examples: QuotedQualityCSV SampleStatistic |
Signed-off-by: Ludovic Orban <lorban@bitronix.be>
|
Guess the title of this PR is no longer accurate. |
ROOT locale when formatting to avoid involving the default locale when numbers are converted to strings
sbordet
left a comment
There was a problem hiding this comment.
Missing:
- ByteRange.toHeaderValue()
- ContentDecoder.Factories.put() -> Same problem but with DecimalFormat
- ShutdownService.listen()
Tests use Locale ar-EG, but Locale ar switched digits from arabic to roman in Java 25.
If ar-EG changes the same way, these tests would pass without testing anything.
Force native digits using the -u-nu- extension, for example hi-IN-u-nu-deva or ar-EG-u-nu-arab.
Signed-off-by: Ludovic Orban <lorban@bitronix.be>
All the places where
String.format("...%d...", ...)was used were modified to explicitly useLocale.ROOTto avoid involving the default locale.Fixes #16020