Breaking backwards compatiblity

[deadalnix]

Le 11/03/2012 22:59, David Nadlinger a écrit :
> On Friday, 9 March 2012 at 22:41:22 UTC, Timon Gehr wrote:
>> Most bug fixes are breaking changes. I don't think we are there yet.
>
> In my opinion, this is a very interesting and important observation –
> due to the powerful meta-programming and reflection capabilities, most
> of the time the question is not whether a change is backward compatibile
> or not, but rather how _likely_ it is to break code. There isn't really
> a good way to avoid that, even more so if your language allow testing
> whether a given piece of code compiles or not.
>
> A related problem is that we still don't quite have an appropriate
> language spec, so you can never be sure if you code is really »correct«
> or if you are relying on DMD implementation details – I'm sure everybody
> who had their meta-programming heavy code break due to a seemingly
> unrelated DMD bugfix knows what I'm trying to say…
>
> David

D is very tied to DMD's implementation. This may be not good, but it is 
too soon to maintain several compilers.