Hi,
I'm testing a bit with tapestry 5.4 (latest alpha). As soon as I have
utf8 characters in my tml i get an exception concerning invalid utf8
bytes. This happens at least on windows.
I assume it's a bug in XMLTokenStream
(https://git-wip-us.apache.org/repos/asf?p=tapestry-5.git;a=blob_plain;f=tapestry-core/src/main/java/org/apache/tapestry5/internal/services/XMLTokenStream.java;hb=HEAD)
within openStream.
On Tapestry 5.3, both reader and writer use system encoding so encoding
will be intact.
On 5.4 the reader ist opened with "utf-8", the writer with system
encoding which might mess up resulting encoding. That was discussed a
while ago in
http://mail-archives.apache.org/mod_mbox/tapestry-dev/201201.mbox/%3ccafaqxjuw4agkr-9cbtpb4qvvgchvevhiwp0vdw7n6ewzu6k...@mail.gmail.com%3E
but never changed.
It would be great if this can be fixed. I could not test an in-place-fix
(using service overrides or so) because the class is directly
instanciated rather than injected. Basically the fix would be to change
the line
PrintWriter writer = new PrintWriter(bos);
to
PrintWriter writer = new PrintWriter(new OutputStreamWriter(bos,
"UTF8"));
(Is not 100% equal because the upper constructor wraps the OSW into a
BufferedWriter but this should not make sense when writing to a BAOS.
Thank you,
Michael.
