Re: Unit Testing Frameworks (was Re: Singletons)

From:

Balog Pal <pasa@lib.hu>

Newsgroups:

comp.lang.c++.moderated

Date:

Fri, 28 Dec 2012 01:52:00 CST

Message-ID:

<kbii77$10sv$1@news.ett.com.ua>

On 12/27/2012 11:11 PM, Dave Abrahams wrote:

on Thu Dec 27 2012, Ian Collins <ian-news-AT-this.is.invalid> wrote:

Unless certainly the test is not interested in the singleton
object's state, then it is even simpler:

NullLogger l;
Logger& GetLogger() { return l;}

and the cases skip the fist line.

I repeat my original question: what the heck can possibly be that
test-ruining boogie related to singleton usage?

I'd like to know the answer to that as well.

A logger singleton is probably not the greatest example for proving
this point, because its actions are usually benign and often
untested.

Logger was just a picked name for the example, and no assumptions
about the semantics were used. Those benign/untested things will more
likely use the simplified version, while the rest the natural one.

However, the use of a singleton may impede or prevent the logger
from being replaced for testing purposes, or prevent different
loggers from being used for individual tests.

How? That is the claim I keep in the FUD category until someone cares
to present the problem finally. Instead of just repeating "may".

I just showed exactly how you can use either the same or different
loggers for test cases -- via a single assignment to a pointer. Why do
you just shove it aside?

--
[ See http://www.gotw.ca/resources/clcm.htm for info about ]
[ comp.lang.c++.moderated. First time posters: Do this! ]