[JENA-2118] Change IO.asUTF8 to map bad characters to U+FFFD not throw an exception.

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: Jena 4.1.0
• Fix Version/s: Jena 4.2.0
• Component/s: Base, RIOT
• Assignee: Andy Seaborne
• Reporter: Andy Seaborne
• Created: 15/Jun/21 09:13
• Updated: 16/Sep/21 11:26
• Resolved: 22/Jun/21 20:17

Description

IO.asUTF8 uses an unmodified CharsetDecoder created from StandardCharsets.UTF_8 passed to JDK InputStreamReader. It has "unmappable" set to REPORT.

But if InputStreamReader is passed the charset, it creates then modifies to get a CharsetDecoder with unmappable set to "REPLACE" (i.e. use U+FFFD, the unicode "unmappable" character.)

Change the behaviour to be "REPLACE".

Issue Links

• is depended upon by: JENA-2120 Generate warning for raw character U+FFFD