Bug 7485 – Incorrect BigInt Error Message

When I try to compile this code: public struct Test1 { BigInt maxCoefficient = (BigInt(10)^^9 )- 1; } The error message is: src\phobos\std\internal\math\biguintx86.d(306): Error: asm statements cannot be interpreted at compile time. I don't know if the code should actually compile, but the error message is not helpful

public struct Test1 { BigInt maxCoefficient = (BigInt(10)^^9 )- 1; // line 1 } If I change line 1 to: BigInt maxCoefficient = BigInt(10)^^9 - 1; // remove parens It crashes the compiler.

I was mistaken about the compiler crashing. It does crash on that line in my code but not in the stripped-down test case.

That's still not the error message I get at all (2.057, Win32): ----- C:\Users\Stewart\Documents\Programming\D\Tests\bugs>dmd bz7485.d bz7485.d(2): Error: undefined identifier BigInt ----- Please, always post a complete, self-contained testcase, and the full output from that testcase. Anyway, adding the necessary import gives for me: ----- bz7485.d ----- import std.bigint; public struct Test1 { BigInt maxCoefficient = (BigInt(10)^^9 )- 1; } ---------- C:\Users\Stewart\Documents\Programming\D\Tests\bugs>dmd bz7485.d d:\dmd2\windows\bin\..\..\src\phobos\std\internal\math\biguintx86.d(306): Error: asm statements cannot be interpreted at compile time d:\dmd2\windows\bin\..\..\src\phobos\std\internal\math\biguintcore.d(434): called from here: multibyteShl(result[words..words + this.data.length],this.data,bits) d:\dmd2\windows\bin\..\..\src\phobos\std\internal\math\biguintcore.d(692): called from here: result.opShl(cast(ulong)(evenbits + firstnonzero * 8u * 4u) * y) d:\dmd2\windows\bin\..\..\src\phobos\std\bigint.d(188): called from here: pow(this.data,u) d:\dmd2\windows\bin\..\..\src\phobos\std\bigint.d(188): called from here: this.data.opAssign(pow(this.data,u)) d:\dmd2\windows\bin\..\..\src\phobos\std\bigint.d(250): called from here: r.opOpAssign(y) bz7485.d(4): called from here: BigInt(BigUint([10u]),false).opBinary(9) bz7485.d(4): called from here: BigInt(BigUint([10u]),false).opBinary(9) ---------- It seems to me it's a combination of two things: - in certain circumstances, it fails to give this CTFE call backtrace (what compiler version are you using?) - the implementation of power on BigInt relies on inline assembly, thereby preventing it working with CTFE. A testcase for the complier crashing might still be useful. http://pr.stewartsplace.org.uk/d/bugreport.html might help you with creating one.

I can't see any bug here. The error message is correct.