make check isn't quite right

Prof Brian D Ripley
Thu, 22 Apr 1999 07:45:53 +0100 (BST)

On Thu, 22 Apr 1999, Kurt Hornik wrote:

> >>>>> Robert Gentleman writes:
> > If I don't build the help pages at the start and then
> > do make check, 
> >   the examples for the base man pages get made but not
> > for any of the packages:
> >> example("smooth",package="eda")
> > Error: No help file found for'smooth'
> > Execution halted
> Your observation is correct.  What currently happens is something like
> 	make R
> 	for pkg in PKGS; do
> 	  maybe rebuild pkg examples/help pages
> 	  try to run the examples
> 	done
> This will fail if the tests for pkg1 assume that pkg2 help/examples must
> be there.
> We could easily change the `make R' to `make; make help; make examples'
> but I am not sure that this is the right thing.  Conceptually, `eda' is
> an add-on package and all that must be there is base (Rlang/Rstat), so I
> think the example should be changed.

I think 

-- the example is a nice one for human readers
-- it is a useful test on the codebase (at least one version I created
   of example failed it).

My problem is with

> > Error: No help file found for'smooth'
> > Execution halted

Now if we could keep running, we would get to see _all_ the problems
and be able to ignore those we knew about. 

Alternatively, I could fix the specific case, and make this a warning and
do-nothing condition.  I am inclined to do that anyway. The only
knock-on effect I can see is that some examples call other examples, and if
the included ones are missing then the main example main fail in an
unanticipated way.


Brian D. Ripley,        
Professor of Applied Statistics,
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272860 (secr)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595

r-devel mailing list -- Read
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: