On Sun, Apr 22, 2001 at 01:54:31AM +0100, Michael G Schwern wrote: > At minimum, everything should at least have a "do I compile and export > the documented interface" test. > > Curmudgeon's anti-excuses follow. Pumpkin's anti-anti-excuses follow. > > > CGI::Carp > > > CGI::Cookie > > > CGI::Push > > > CGI::Switch > > > > I don't see us testing much CGI stuff. > > There are already six CGI related tests in the core. All these > modules can run intependent of a web server. If we're going to > distribute them, we're going to test them. Tests welcome. > > > File::stat > > > > Something very simple can be tested, but not much. > > Something very simple like the 287 lines of t/op/stat.t? These should If you had used 'less' instead of 'wc' :-) you would have noticed that most of it is $^O testing, due to the rather unportable nature of file system attributes. > all be mirrored as much as possible against File::stat. > > > > Filter::Simple > > > > The tests coming with Filter::Simple were too...simple. > > Better than nothing. When I did the integration I was going to write better tests, but the forgot. > > > I18N::Collate > > > > Heavily, heavily deprecated. > > It still has to work. Do not tempt me or I'll take it away :-) It's hard to test since it depends on locales. What this means is that you can test but you cannot know which is the correct result. > Can't we assume certain things will be there on certain OS's, such as > loopback? > Great, shall we remove the t/base tests then? ;) > > PerlIO claims it can do things, for example, the :crlf in the > SYNOPSIS. That's a starting point to test against. True, all those things should be tested. > > > filetest > > > > System dependent, requires filesystems with ACLs. > > We're smart. We can figure out if a filesystem has ACLs and only run > the test in that case. Sorry but you are too flippant, this is *really* hard. ACLs are perhaps the one most unportable thing in UNIXes you can think of. Without people who use this supplying tests that make sense on their filesystems and ACL schemes, we are out of luck on this. -- $jhi++; # http://www.iki.fi/jhi/ # There is this special biologist word we use for 'stable'. # It is 'dead'. -- Jack Cohen
