When loading a .tex file, the filter removes the first character of each paragraph.
Also, the filter does not respect line breaks and the output of an untranslated document is totally broken.
ftp://www.ctan.org/tex-archive/macros/latex/base/small2e.tex
http://www.gnu.org/software/emacs/manual/texi/elisp.texi.tar.gz
The first example (small2e.tex) works fine here: no first character missing, and the output is fine.
Are you talking about paragraphs, or arbitrary line breaks? As far as I remember, arbitrary line breaks are not respected on purpose.
Browsing quickly the other files, the only one where I can see serious issues is texinfo.tex. The other "texi" files seem to work reasonably fine.
Are you sure you are using the LaTex filter? Just a routine check, as sometimes the list of filters can get messed up, and you end up using the wrong filter.
Didier
Based on a discussion in the development mailing, it's quite possible your issues were coming from the use of Java 8. As a workaround was implemented for this inconsistent Java behaviour, I consider the issue closed, unless there's a good reason to reopen.
Didier
Fixed in SVN (/trunk).
Didier
Fixed in the released version 3.1.6 of OmegaT.
Didier