Re^2: Class::DBI -- does it get any easier?

Coverage numbers are not a perfect indicator. Class::DBI has one of the largest test suites around, and Tony always asks people to submit tests with their patches. It is a well-tested module.

Comment on Re^2: Class::DBI -- does it get any easier?

Replies are listed 'Best First'.
Re^3: Class::DBI -- does it get any easier? by stvn (Monsignor) on Nov 12, 2004 at 20:49 UTC
Coverage numbers are not a perfect indicator. No they are not, and it seems from looking closer at some of the source that a good amount of Class::DBI seems to be auto-generated (all that `__PACKAGE__->mk_something` stuff). Which makes it very difficult to check the test's code coverage. But that code is not even included by Devel::Cover in the coverage report anyway (see here). However, a good portion of the code that is checked by Devel::Cover is not tested. I will agree that Class::DBI is surely better tested than the Devel::Cover numbers give it credit for, but that is not to say that it couldn't use a little more. -stvn	[reply] [d/l]

Replies are listed 'Best First'.

Re^3: Class::DBI -- does it get any easier?
by stvn (Monsignor) on Nov 12, 2004 at 20:49 UTC

Coverage numbers are not a perfect indicator.

No they are not, and it seems from looking closer at some of the source that a good amount of Class::DBI seems to be auto-generated (all that __PACKAGE__->mk_something stuff). Which makes it very difficult to check the test's code coverage. But that code is not even included by Devel::Cover in the coverage report anyway (see here). However, a good portion of the code that is checked by Devel::Cover is not tested.

I will agree that Class::DBI is surely better tested than the Devel::Cover numbers give it credit for, but that is not to say that it couldn't use a little more.

-stvn

[reply]
[d/l]