You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@poi.apache.org by Daniel Noll <da...@nuix.com> on 2008/11/14 05:46:17 UTC
HWPF: "Your document seemed to be mostly unicode..."
Hi all.
I'm testing out 3.2 (FWIW, so far 3.1 is much more stable)
I'm getting this error a *lot*:
"Your document seemed to be mostly unicode, but the section definition
was in bytes! Trying anyway, but things may well go wrong!"
On top of this, I have lost a lot of embedded images which used to be
present in 3.1. These warning messages tend to coincide with the
documents where the embedded images are lost, so I'm wondering if these
two phenomena are related?
I'm yet to find a failure on some test data which I can distribute, but
if/when I do so I will file a bug report noting the regression.
Daniel
--
Daniel Noll Forensic and eDiscovery Software
Senior Developer The world's most advanced
Nuix email data analysis
http://nuix.com/ and eDiscovery software
---------------------------------------------------------------------
To unsubscribe, e-mail: user-unsubscribe@poi.apache.org
For additional commands, e-mail: user-help@poi.apache.org
Re: HWPF: "Your document seemed to be mostly unicode..."
Posted by Daniel Noll <da...@nuix.com>.
Nick Burch wrote:
> I think we have one or two files in the svn testdata that trigger this
> warning, but not sure if any of them have images in. It might be worth
> seeing if you can add an image to one of them without word making bytes
> vs chars consistent, then that'd be a good base to work from. It's a
> very nasty bit of the spec though :(
I managed to find a test document which was redistributable, so I've
created Bug 46220.
Daniel
--
Daniel Noll Forensic and eDiscovery Software
Senior Developer The world's most advanced
Nuix email data analysis
http://nuix.com/ and eDiscovery software
---------------------------------------------------------------------
To unsubscribe, e-mail: user-unsubscribe@poi.apache.org
For additional commands, e-mail: user-help@poi.apache.org
Re: HWPF: "Your document seemed to be mostly unicode..."
Posted by Nick Burch <ni...@torchbox.com>.
7~On Fri, 14 Nov 2008, Daniel Noll wrote:
> "Your document seemed to be mostly unicode, but the section definition
> was in bytes! Trying anyway, but things may well go wrong!"
poi 3.1 would quite happily just read in garbage to the various text piece
tables, then do odd things. poi 3.2 will try harder to get it right
(The problem is that microsoft did a half-arsed job at supporting unicode
in the files. Even with a pure unicode file, half the offsets are stored
as bytes, and half are as characters, so much hilarity ensues. This error
is triggered when we're pretty sure one particular definition ought to be
in characters for your file, but the size of the offsets tends to hint
they're in bytes instead. If you fancy hurting your head a lot, then
improvements to the byte/character expectations code are gratefully
received!)
> I'm yet to find a failure on some test data which I can distribute, but
> if/when I do so I will file a bug report noting the regression.
I think we have one or two files in the svn testdata that trigger this
warning, but not sure if any of them have images in. It might be worth
seeing if you can add an image to one of them without word making bytes vs
chars consistent, then that'd be a good base to work from. It's a very
nasty bit of the spec though :(
Nick
---------------------------------------------------------------------
To unsubscribe, e-mail: user-unsubscribe@poi.apache.org
For additional commands, e-mail: user-help@poi.apache.org