[Issue 7339] New: [2.058] std.string.indexOf doesn't instantiate with inlining

[d-bugmail at puremagic.com]

http://d.puremagic.com/issues/show_bug.cgi?id=7339

           Summary: [2.058] std.string.indexOf doesn't instantiate with
                    inlining
           Product: D
           Version: D2
          Platform: All
        OS/Version: All
            Status: NEW
          Keywords: rejects-valid
          Severity: regression
          Priority: P2
         Component: DMD
        AssignedTo: nobody at puremagic.com
        ReportedBy: dsimcha at yahoo.com
        Depends on: 4724


--- Comment #0 from David Simcha <dsimcha at yahoo.com> 2012-01-21 12:08:17 PST ---
This is related to Bug 4724, but I'm filing it as a separate issue and labeling
it a regression because something (probably the more aggressive inlining that
was added for 2.058) expands the effects of that bug in ways that break
previously working code.  If we can't fix the compiler bug then we need to work
around it to avoid an awful regression.

import std.string;

void main() {
    "".indexOf("");
}

dmd -O -inline -release -c test.d
/home/dsimcha/dmd2/linux/bin64/../../src/phobos/std/algorithm.d(2970): Error:
function
std.string.indexOf!(char,char).indexOf.simpleMindedFind!(__lambda6,const(char)[],const(char)[]).simpleMindedFind
is a nested function and cannot be accessed from main

Here's a reduced version that doesn't import any Phobos modules:

void main() {
    "".indexOf("") ;
}

void indexOf(const(char)[] s, const(char)[] sub) {
    find!((dchar a, dchar b){return a == b;})
        (s, sub);
}

void find(alias pred)(const(char)[] haystack, const(char)[] needle) {
     pred(haystack[0], needle[0]);
}

$ dmd -c -O -inline -release test.d
test.d(11): Error: function test.indexOf.find!(delegate pure nothrow @safe
bool(dchar a, dchar b)
{
return cast(uint)a == cast(uint)b;
}
).find is a nested function and cannot be accessed from main

-- 
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------