• Dmitry Olshansky (7/7) Oct 11 2017 My motivation is, of course, std.regex.
• Steven Schveighoffer (5/14) Oct 11 2017 100% agree. Any CTFE'd result is implicitly unique. In fact CTFE
• Michael V. Franklin (4/11) Oct 11 2017 For whatever it's worth, I also agree.

Dmitry Olshansky <dmitry.olsh gmail.com> writes:

My motivation is, of course, std.regex.

static immutable r = regex("abc");

This is perfectly fine but because in general there is some 
amount of impurtiy in regex and it won't work. Yet regex is 
called at CTFE so there is no way it can escape mutable pointer 
to its result.

Thoughts?

Steven Schveighoffer <schveiguy yahoo.com> writes:

On 10/11/17 10:30 AM, Dmitry Olshansky wrote:
 My motivation is, of course, std.regex.
 
 static immutable r = regex("abc");
 
 This is perfectly fine but because in general there is some amount of 
 impurtiy in regex and it won't work. Yet regex is called at CTFE so 
 there is no way it can escape mutable pointer to its result.
 
 Thoughts?

100% agree. Any CTFE'd result is implicitly unique. In fact CTFE 
annoyingly makes everything unique, even where it doesn't have to!

Please file enhancement request if not already.

-Steve

Michael V. Franklin <slavo5150 yahoo.com> writes:

On Wednesday, 11 October 2017 at 14:30:54 UTC, Dmitry Olshansky 
wrote:
 My motivation is, of course, std.regex.

 static immutable r = regex("abc");

 This is perfectly fine but because in general there is some 
 amount of impurtiy in regex and it won't work. Yet regex is 
 called at CTFE so there is no way it can escape mutable pointer 
 to its result.

 Thoughts?

For whatever it's worth, I also agree.

Mike