bugfix 319 fix invalid BigDecimal serialization because of locale#320
bugfix 319 fix invalid BigDecimal serialization because of locale#320Simulant87 wants to merge 2 commits intoeclipse-ee4j:masterfrom Simulant87:319-fix-invalid-big-decimal-localization
Conversation
Signed-off-by: Simulant <nfaupel.dev@gmail.com>
aguibert
left a comment
aguibert
left a comment
There was a problem hiding this comment.
Thanks for the contribution @Simulant87! Just one minor change requested and then we can merge
| final NumberFormat format = NumberFormat.getInstance(jsonbContext.getConfigProperties().getLocale(numberFormat.getLocale())); | ||
| final NumberFormat format = NumberFormat.getInstance((Locale.ENGLISH)); | ||
| ((DecimalFormat)format).applyPattern(numberFormat.getFormat()); | ||
| format.setParseIntegerOnly(integerOnly); |
There was a problem hiding this comment.
going with the same train of thought not allowing commas in JSON numbers, I think we should also call format.setGroupingUsed(false) here, because people may try to use a number format like ###,##0.00 which could output a number like 123,456.78 and should be serialized as 123456.78 according to RFC 7159
There was a problem hiding this comment.
I updated this and also the test to cover it. But there are other test failing:
org.eclipse.yasson.customization.NumberFormatTest#testDeserializer
org.eclipse.yasson.customization.NumberFormatTest#testDeserializeWithoutClassLevelFormatter
I am unsure if I should just adopt the expected behavior/input data to my new implementation to let them pass as this changes current behavior.
There was a problem hiding this comment.
while I think this behavior proposed in this PR is technically correct, some existing users may be depending on the behavior and it would be good to not break such users in a micro release.
Since we have a major/minor release of JSON-B on the horizon, I am thinking we should defer this PR to a later release. Is that OK with you @Simulant87 or do you urgently need this change?
There was a problem hiding this comment.
That is OK with me and I totally agree that because of the changed behavior it should be part of a major- and not part of a minor-release. I am not depending an a fix.
Signed-off-by: Simulant <nfaupel.dev@gmail.com>
2 participants
fix #319