In rt.util.utf are a set of low level utf functions.  Phobos duplicates quite a few of them in std.utf.  Just like arrays and other built in data types, their support code should be exposed by the runtime, imho.

My current aim is to make sure the dmd unit tests stay minimal.  Importing phobos level files... isn't. :)

Thoughts?

Later,
Brad

On Jun 19, 2010, at 11:36 PM, Brad Roberts wrote:

> In rt.util.utf are a set of low level utf functions.  Phobos duplicates quite a few of them in std.utf.  Just like arrays and other built in data types, their support code should be exposed by the runtime, imho.
> 
> My current aim is to make sure the dmd unit tests stay minimal.  Importing phobos level files... isn't. :)

Yeah, that's why I cloned std.utf in the first place :-)  I've been thinking the same thing, but figured it would be nice to do a code review of the module at the same time and just haven't gotten to it.  Do you think the module should just be moved into core as-is or should it be worked over a bit?

On 6/20/2010 12:42 PM, Sean Kelly wrote:
> On Jun 19, 2010, at 11:36 PM, Brad Roberts wrote:
> 
>> In rt.util.utf are a set of low level utf functions.  Phobos duplicates quite a few of them in std.utf.  Just like arrays and other built in data types, their support code should be exposed by the runtime, imho.
>>
>> My current aim is to make sure the dmd unit tests stay minimal.  Importing phobos level files... isn't. :)
> 
> Yeah, that's why I cloned std.utf in the first place :-)  I've been thinking the same thing, but figured it would be nice to do a code review of the module at the same time and just haven't gotten to it.  Do you think the module should just be moved into core as-is or should it be worked over a bit?
> _______________________________________________

A review would be good, but I'd also be just fine with a straight move and a good hearted attempt to come back to it later. :)