PEP 594 cgi & cgitb removal

Robin Becker robin at reportlab.com
Wed May 22 03:51:30 EDT 2019


In PEP 594 t has been proposed that cgi & cgitb should be removed. I suspect I am not the only person in the world that likes 
using cgi and cgitb.

One of the nice features in cgitb is the ability to get a nice traceback with variable values etc etc etc. I have used the 
underlying mechanism to produce better traceback information on several occasions and not only in cgi applications.

I filed a bug against cgitb in 2004 with (apparently unacceptable patches) that is still unfixed.

Although cgi is claimed to be dead it is still one of the easiest ways to get a web service to operate; it is also stateless and 
robust. If cgi and similar are removed from the stdlib python will become significantly less charged. These batteries are not dead 
they are pining.

There is some discussion on the python-dev list of moving these allegedly dead packages to pypi, but of course that means issues 
of control, where do the sources reside and other politics.

Django has a similar feature to cgitb's output for tracebacks, but is too deeply embedded for use elsewhere; is there anything 
suitable elsewhere?
-- 
Robin Becker



More information about the Python-list mailing list