Bug 10891 – regression(2.063.2 => v2.064-devel-a3e447c): Error: function std.uni.TrieBuilder

dmd_from_gitmaster -c -inline -release -noboundscheck -Ipath/to/Dscanner/ main.d phobos/std/uni.d(3547): Error: function std.uni.TrieBuilder!(ubyte, dchar, 1114112, sliceBits!(13LU, 21LU), sliceBits!(6LU, 13LU), sliceBits!(0LU, 6LU)).TrieBuilder.putAt errors compiling the function if you use dmd (2.063.2) instead of dmd_from_gitmaster, it works. if you remove any of the flags -inline -release -noboundscheck, it works. ---- import stdx.d.lexer; void main() { LexerConfig config; auto a=(ubyte[]).init.byToken(config); } ----

Confirmed, with: dscanner - 42cf96c6d24bb21ad076ac81d37211f2c389057a dmd - 1399b1f7b94ef1c1b24f2ce842b6f247e2e1fc52 phobos - 153034ccf843cc00052ede6febc79a3d26bd1dff Also, DMD crashes after the error. Stack trace: 0018f9c0 004e7dbd image00400000!FuncDeclaration::toObjFile+0x4a [C:\dmd-git\dmd2\src\dmd\src\glue.c @ 553] 0018f9ec 004e86ea image00400000!StructDeclaration::toObjFile+0x14e [C:\dmd-git\dmd2\src\dmd\src\toobj.c @ 988] 0018fa14 004e099c image00400000!TemplateInstance::toObjFile+0x9b [C:\dmd-git\dmd2\src\dmd\src\toobj.c @ 1329] 0018fac4 00405823 image00400000!Module::genobjfile+0x28a [C:\dmd-git\dmd2\src\dmd\src\glue.c @ 360] 0018ff0c 00405a53 image00400000!tryMain+0x32c5 [C:\dmd-git\dmd2\src\dmd\src\mars.c @ 1662] 0018ff48 0057aced image00400000!main+0x43 [C:\dmd-git\dmd2\src\dmd\src\mars.c @ 1724] windbg highlights this line (the if line): // If errors occurred compiling it, such as bugzilla 6118 if (type && type->ty == Tfunction && ((TypeFunction *)type)->next->ty == Terror) return; I'm assuming 'next' is probably null at this point (can't verify this yet, maybe later).

(In reply to comment #0) > phobos/std/uni.d(3547): Error: function std.uni.TrieBuilder!(ubyte, dchar, > 1114112, sliceBits!(13LU, 21LU), sliceBits!(6LU, 13LU), sliceBits!(0LU, > 6LU)).TrieBuilder.putAt errors compiling the function (In reply to comment #1) > windbg highlights this line (the if line): > > // If errors occurred compiling it, such as bugzilla 6118 > if (type && type->ty == Tfunction && ((TypeFunction *)type)->next->ty == > Terror) > return; > > I'm assuming 'next' is probably null at this point (can't verify this yet, > maybe later). I think this would be a dup of bug 10836, and it was recently fixed. Could you confirm this bug again with git-head?

(In reply to comment #2) > I think this would be a dup of bug 10836, and it was recently fixed. > Could you confirm this bug again with git-head? Confirmed, commit 01d22e4f3b2a6c07dabc0307eb4ee6b066ae27dd fixes it, previous commit had the issue. *** This issue has been marked as a duplicate of issue 10836 ***