Regression testing of tetex

[Junichi Uekawa <dancer@netfort.gr.jp>]

Hi,

I'm mostly concerned with the lack of regression testing with tetex
packages; they are uploaded, but only mostly tested against
installation/removal, not document processing.

The processes that can obviously be automated are:

1. install-remove of packages 

2. making sure that documents (possibly in tetex-doc) can be processed
   without stopping with an error.

   Of course, doing this would help detect cases like 'dvipdfmx was
   broken, and gone unnoticed for several months'.

The process that cannot really be automated is:

3. making sure that the resulting text is same, or reasonable.

But if 2 was done regularly, and the output was there, there is less
overhead to perform manual review of documents.


This is what I've been meaning to look at for the last few months and
have not gone around to doing, maybe I'll have a chunk of time
available at Debconf to start off, but I've wanted to do a brain-dump
so that others know of what I'm thinking.


regards,
	junichi
-- 
dancer@{debian.org,netfort.gr.jp}   Debian Project