• Significant Change
  • TextFormatDecodingError has a new error case to got with a recursion limit for TextFormat decoding (add safety found via fuzz testing for potential bogus input trying trigger stack overflow #1132), if you have any switch states on all the cases, this is a breaking change in that you must handle the new case.
• Fixes from fuzz testing:
  • Fix octal TextFormat decoding failure #1124
  • Avoid walking off the end of the buffer in two parsing cases. #1126
  • Add TextFormatDecodingOptions and implement a recursion limit. #1132
  • Don't walk off the end of the buffer during a unicode bytes decode. #1136
• Minor Updates
  • Change oneof enforcement to allow null (found via upstream conformance test requirements) #1135
  • Allow proto3_optional for extensions. #1138
  • Some edge case speed improvements:
    • Add modify operation to ExtensionFieldValueSet. #1137
    • Don't do characterwise-compares if not needed. #1145
    • Clear previous contents before decoding Any from TextFormat #1147