On Tuesday, January 24, 2017 at 8:14:23 PM UTC+1, William wrote: > > On Tue, Jan 24, 2017 at 11:05 AM, Jeroen Demeyer <jdem...@cage.ugent.be > <javascript:>> wrote: > > On 2017-01-24 16:14, William Stein wrote: > >> > >> I would very likely just get it done keeping the > >> current behavior as much as possible. > > > > > > As I said, the current behaviour is not documented. So you would have to > > actually reverse-engineer the code to find out what the current > behaviour > > actually is. This makes little sense to me... > > That's not much different than saying "The current behavior is > documented. So you would have to actually read the documentation to > find out what the current behaviour actually is. This makes little > sense to me..." Except, at least the code really does define how > things work, whereas documentation is often wrong. > > Again, I 100% defer to **whoever does the actual porting work** to > make the decision of how they will do it. >
I understand this as practically saying that we should avoid breaking existing doctests as much as possible and try to be consistent in new doctests as much as possible. And how the new doctests would be written is determined when we actually write the code. So how about opening one meta ticket and lots of individual tickets for the cmp cases to tackle, and defer discussions to each ticket? Though I think, like Jeroen, setting forth some guiding principles outright would be useful to start. I guess a dictator who knows very well the coercion framework may do this... -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at https://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
