Remove all uses of std.metastrings from Phobos #1171
Conversation
| @@ -2285,7 +2284,7 @@ unittest | |||
| auto p = toUTFz!P(s); | |||
| immutable len = zeroLen(p); | |||
| enforce(cmp(s, p[0 .. len]) == 0, | |||
| new AssertError(Format!("Unit test failed: %s %s", P.stringof, S.stringof), | |||
There was a problem hiding this comment.
This will have different semantics. You'll want to assign the result of the format call to an enum and then use that.
There was a problem hiding this comment.
Why does it matter? Sure, the string will be generated at runtime instead of compile time, but why do we care? I don't know why it was using Format in the first place (though I'm pretty sure that I wrote this code, so that oddity is my fault).
There was a problem hiding this comment.
I don't know. I'm just saying, since it seemed weird to change something that was compile-time to run-time.
There was a problem hiding this comment.
I'll admit that I didn't catch the change in semantics, but looking at the code, I really don't think that it matters. If anything, it never should have used Format in the first place.
|
Just a remark, not intended to block this pull, that one good thing about std.metastrings is that it's pure, whereas std.format.format isn't. Just noting that this could be a problem if we need to deal with pure code. |
|
LGTM. This should be pulled ASAP to fix http://d.puremagic.com/issues/show_bug.cgi?id=9587 |
|
LGTM. Please pull ASAP, as my compiles are now flooded with deprecation messages from std.typecons. |
Remove all uses of std.metastrings from Phobos
As it's now deprecated, nothing in Phobos (aside from std.metastrings itself) should be using it anymore.