I appreciate everyone's input on this. I've decided I'm going to treat this as a minor version bump, for the following reasons:
* It's unclear that *any* breakage will occur.* If any breakage does occur, it should be trivial to fix in a backwards compatible way.
* And while breakage is generally a bad thing, in this case, it would likely be beneficial to the community if we got extra data out of the exercise by having a build breakage result from this change.
If I hear any reports of breakage as a result of this change, I'll try to remember to report them.
On Tue Dec 16 2014 at 11:48:41 AM Roman Cheplyaka <
roma@ro-che.info> wrote:
On 16/12/14 11:40, Erik Hesselink wrote:
> On Mon, Dec 15, 2014 at 11:39 PM, Roman Cheplyaka <roma@ro-che.info> wrote:
>> On 16/12/14 00:31, Johan Tibell wrote:
>>> So yes, if you use open imports and allow new minor versions, your code
>>> might break. This is expected.
>>
>> One similarly could argue that "if you use functions in polymorphic
>> contexts and allow new minor versions, your code might break".
>>
>> This isn't in the PVP, but it's exactly in the same spirit, IMO.
>
> I think this is a bad idea, since a "polymorphic context" (even if
> rigorously defined) is not clear from reading the code, you have to
> type check it in your head. The import style, on the other hand, is
> very simple to read and check if it matches the style of dependency
> version ranges.
Ok, this is a fair point.
Roman
_______________________________________________
Haskell-Cafe mailing list
Haskell-Cafe@haskell.org
http://www.haskell.org/mailman/listinfo/haskell-cafe