Bug 6471 – std.metastrings.Format has recursive expansion problems

dmd32 2.054. compiling the attached file results in /usr/include/d/std/metastrings.d(72): Error: template instance std.metastrings.FormatString!("*/\x0a @property Node right%s(Node newNode)\x0a {\x0a _right%s = newNode;\x0a if(newNode !is null)\x0a newNode._parent%s = &this;\x0a return newNode;\x0a }\x0a",N,N,N) recursive expansion

Created attachment 1016 dmd doesn't like Format

Now that std.format is CTFE-able, std.metastrings has been deprecated. Maybe this bug can be closed?

Depends on whether you consider it a dmd issue or a phobos issue. I just took a second look at this and.. you get a recursion depth of exactly 500 for templates. I don't recall this restriction being documented anywhere, so if it isn't, it should be, and after that maybe take up the question of whether it is an acceptable restriction (with ctfe maybe it is now). import std.string; enum s = "abcdefghijklmorpshtn"; static assert(s.length == 20); enum t = s ~ s; static assert(t.length == 40); enum u = t ~ t; static assert(u.length == 80); enum v = u ~ u; static assert(v.length == 160); enum w = v ~ v; static assert(w.length == 320); enum x = w ~ v ~ s; static assert(x.length == 500); /* // ok! enum x = w ~ v ~ s[1..$]; static assert(x.length == 499); */ template iter(string s, T) { static if(s.length == 0) { enum iter = ""; }else{ enum iter = toUpper(s[0..1]) ~ iter!(s[1..$], T); } } pragma(msg, iter!(x, int)); void main() {}

std.metastrings has been deprecated and removed, so this bug will never be "fixed". Use std.format instead (it is now supported in CTFE).