[liblouis-liblouisxml] Re: Hyphenation, I think I have got it, but may have found more bugs

  • From: "John J. Boyer" <johnjboyer@xxxxxxxxxxxxx>
  • To: liblouis-liblouisxml@xxxxxxxxxxxxx
  • Date: Tue, 9 Jun 2009 06:31:59 -0500

Michael,

Again, thanks for finding the bugs! I am working on the lou_checkhyphens 
tool, so I'll soon be able tr track them down.

John

On Tue, Jun 09, 2009 at 11:54:55AM +0100, Michael Whapples wrote:
> Hello,
> After looking at the liblouisxml code a bit more and a bit of thinking I 
> believe I now have worked out the hyphenation problems I was having, it 
> seems to work both for mode=0 and mode=1 now. When I say working, it 
> gives what seems sensible although possibly at times I might personally 
> add another potential hyphenation point if I were doing it manually. My 
> work containing this working code has been committed to mercurial on 
> http://bitbucket.org/mwhapples/jlouis should anyone wish to look at it.
> 
> Unfortunately I think I have found a few more bugs (I suppose this is 
> what comes of trying to create full bindings for a library).
> 
> The first is that when I use mode=1 sometimes my application (Test.java) 
> was crashing with a load of glib errors and messages I didn't 
> understand. The other time I had come across these sorts of errors was 
> when I passed the wrong size arrays to liblouis in some of the other 
> function calls. So I guessed that may be it was something to do with the 
> difference in size of the translated text I passed in and the back 
> translated text lou_hyphenate would create to perform the hyphenation, 
> and sure enough when I set the array size of hyphens to twice the size 
> of inlen, all these errors have gone away. The docs say that hyphens 
> should be of size inlen.
> 
> Secondly I previously reported getting values other than '0' and '1' in 
> hyphens, I believe these may be occurring around contractions and only 
> seems to happen for mode=1. Could this be due to the values in each byte 
> for the back translated hyphens being combined (eg. added together) when 
> translation is done back for the calling app? If this is the case may be 
> that's why sometimes I would personally find one more potential 
> hyphenation point than liblouis.
> 
> Michael Whapples
> For a description of the software and to download it go to
> http://www.jjb-software.com

-- 
My websites:
GodTouches Digital Ministry, Inc. http://www.abilitiessoft.com/godtouches
Abilitiessoft, Inc. http://www.abilitiessoft.com
Location: Madison, WI, USA

For a description of the software and to download it go to
http://www.jjb-software.com

Other related posts: