Re: Regression testing of tetex

[Frank Küster <frank@kuesterei.ch>]

Norbert Preining <preining@logic.at> wrote:

> Hi all!
>
> On Die, 14 Mär 2006, Norbert Preining wrote:
>> 	http://tug.org/svn/texlive/trunk/Master/support/tests/
>
> The more I look at it the more I like it. Ok, it is not up-to-date, the
> knowngood are probably old, but alreday like this several tests run
> without problem, and there is a bit of infrastructure and examples how
> to extend/add tests.
>
> Ic checked in the test-bed item on the top of the svn tree. Maybe you
> take a look and we start hacking around it...

Do you think that using pdftoppm and comparing the pixel files is the
way to go?  I guess any internal structure in the PDF gets lost, like
hyperlinks, character information (whether searching for "fl" will find
a fl ligature), and other things that are in fact important for PDF
files. 

Therefore I think my approach of using real, uncompressed PDF files
without the changing font and document information might serve us better
in the long run.

Regards, Frank
-- 
Frank Küster
Single Molecule Spectroscopy, Protein Folding @ Inst. f. Biochemie, Univ. Zürich
Debian Developer (teTeX)