On Wed, Jun 30, 2010 at 11:20 AM, Robert Miller <r...@rlmiller.org> wrote: > On Wed, Jun 30, 2010 at 10:01 AM, William Stein <wst...@gmail.com> wrote: >> On Wed, Jun 30, 2010 at 9:48 AM, Robert Miller <r...@rlmiller.org> wrote: >>> I was recently searching through the source and I noticed that some >>> loads/dumps doctests say #indirect doctest, and others don't. I know >>> that for __reduce__ methods, this is correct, but it struck me that >>> perhaps all loads/dumps doctests are indirect. I guess this is a >>> philosophical/moral question, but I wanted to get other people's >>> opinions... What do y'all think? >> >> Clearly, all but two of them are indirect (the two that test the loads >> and dumps functions!). > > How very correct. That was my thinking as well. Although I suppose > this is strictly pedantic, since the coverage/doctesting scripts don't > care about the distinction. Coverage just looks for a loads(dumps test > in each class, and doctesting doesn't care, so I don't see a point in > fixing this, other than to be consistent. > > NB There are many of these types of tests which are not marked as indirect. > > Should I open a ticket, or find something more productive to do?
I think all special methods should be implicitly indirect, which includes loads/dumps (usually doctested in __reduce__, though sometimes it's worth calling __reduce__ directly). The only reason #indirect is there is to silence the coverage script warnings that you don't appear to be using the function that you're supposedly testing. - Robert -- To post to this group, send an email to sage-devel@googlegroups.com To unsubscribe from this group, send an email to sage-devel+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-devel URL: http://www.sagemath.org
