On 29/09/11 08:29, Michael Brunnbauer wrote:
hi all
I get a RiotParseException with a big TDB and tdbdump from TDB-0.8.10 after
much of the data got written out. Do I have a corrupt TDB or is there a bug
somewhere ? Here is the full trace:
Exception in thread "main" org.openjena.riot.RiotParseException: [line: 1, col:
4 ] illegal escape sequence value: f (0x66)
The data seems to have a form feed character "\f" in it. Unusual but legal.
I've just checked the code and it looks as if the database is fine,
because it's (just) the decoding of nodes that is broken. I've just
fixed that in SVN.
Work arounds:
I can do a ARQ SNAPSHOT build (to go with TDB-0.8.11 SNAPSHOT).
Or you can take the code for ARQ 2.8.8 and patch that in
TokenizerText.readLiteralEscape ; add
case 'f': return '\f' ;
to the switch statement.
(The whole node writing-reading could do with a checkover in case it can
be made more efficiency but that's another matter)
...
at tdb.tdbdump.main(tdbdump.java:20)
cu,
brunni
Sorry about that,
Andy
