Closed GoogleCodeExporter closed 8 years ago
Just to recap Paul's analysis from the email so that we have all the info in
one place, he suspects that short reads from the input stream can leave the
buffer only partially filled, leaving a large number of NULs which corrupt the
character set analysis.
Original comment by tfmorris
on 13 Jun 2011 at 9:09
Fixed in r2102. Thanks for the patch. It was committed mostly as given except
I conditionalized the wrapping of the input stream to only do it if the input
stream doesn't support mark/reset.
p.s. It's a nit, but it'll make your patch less noisy if you turn off automatic
"cleanup" of formatting in your editor/IDE. Not a big deal for a little patch
like this, but could complicate things for a larger patch.
Original comment by tfmorris
on 14 Jun 2011 at 5:58
Original issue reported on code.google.com by
paulm%pa...@gtempaccount.com
on 13 Jun 2011 at 6:42Attachments: