Is there some magical way to avoid fixing pdf parsing code every 2-3 months when ever something small changes? And where the list of small things that change is

Is there some magical way to avoid fixing pdf parsing code every 2-3 months when ever something small changes? And where the list of small things that change is quite large and random. It has become a full time job for some people I work with - decoding corporate/govt pdfs. They could be spending their time on more useful stuff instead of header format changed, data format changed, location moved etc etc so parser crashed.