Re: #67: atf doesn't conform to GNU coding standards; breaks hier(7) on FreeBSD

["atf trac notification" <nobody%julipedia.org@localhost>]

#67: atf doesn't conform to GNU coding standards; breaks hier(7) on FreeBSD
----------------------------------+-----------------------------------------
  Reporter:  gcooper@...          |       Owner:  jmmv  
      Type:  defect               |      Status:  closed
  Priority:  major                |   Milestone:        
 Component:  BUILD                |     Version:  HEAD  
Resolution:  wontfix              |    Keywords:        
----------------------------------+-----------------------------------------

Comment (by Ade Lovett <ade@...>):

 The main problem with choosing a generic subdirectory such as 'tests' is
 the implication that atf is going to be the ONLY automated test framework.
 If another one comes along (let's not get bogged down in whether that's
 likely or not) expecting to also use ${PREFIX}/tests/<foo>,<bar> but in a
 different way, then you run into the issue of which tests/ subdirectories
 are to be processed by atf, and which by a.n.other tool.

 We already have well-established conventions for
 ${PREFIX}/share/${PORTNAME}, ${PREFIX}/share/doc/${PORTNAME} etc., so it
 would seem logical to say something like:

 ${PREFIX}/tests - top-level unit-frame testing directory
 ${PREFIX}/tests/<foo> - top-level unit-frame testing directory for unit-
 frame application <foo>

 and then put everything below ${PREFIX}/tests/<foo>/

 One extra component in the path to the unit tests themselves isn't going
 to kill anyone, and makes the whole tests/ directory inherently more
 scalable when faced with multiple testing platforms.

-- 
Ticket URL: <http://www.julipedia.org/projects/atf/trac/ticket/67#comment:2>
Automated Testing Framework <http://www.NetBSD.org/~jmmv/atf/>
Automated Testing Framework