go/test/fixedbugs/issue11326.go
Robert Griesemer 732e2cd746 cmd/compile: don't truncate tiny float constants to 0 in error messages
Fixes #13559.

Change-Id: I6fe8b5083192e8eb6c1b3ca1919fde81a00ccb7e
Reviewed-on: https://go-review.googlesource.com/17695
Run-TryBot: Robert Griesemer <gri@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Alan Donovan <adonovan@google.com>
2015-12-10 16:39:46 +00:00

31 lines
1.2 KiB
Go

// errorcheck
// Copyright 2015 The Go Authors. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
// Tests for golang.org/issue/11326.
package main
func main() {
// The gc compiler implementation uses the minimally required 32bit
// binary exponent, so these constants cannot be represented anymore
// internally. However, the language spec does not preclude other
// implementations from handling these. Don't check the error.
// var _ = 1e2147483647 // "constant too large"
// var _ = 1e646456993 // "constant too large"
// Any implementation must be able to handle these constants at
// compile time (even though they cannot be assigned to a float64).
var _ = 1e646456992 // ERROR "1e\+646456992 overflows float64"
var _ = 1e64645699 // ERROR "1e\+64645699 overflows float64"
var _ = 1e6464569 // ERROR "1e\+6464569 overflows float64"
var _ = 1e646456 // ERROR "1e\+646456 overflows float64"
var _ = 1e64645 // ERROR "1e\+64645 overflows float64"
var _ = 1e6464 // ERROR "1e\+6464 overflows float64"
var _ = 1e646 // ERROR "1e\+646 overflows float64"
var _ = 1e309 // ERROR "1e\+309 overflows float64"
var _ = 1e308
}