Bug 746 – incorrect code generation for "new Type[][](1, 1)"

ID	Filename	Summary	Content-Type	Size
88	array_initialization_26_E.o	Attached array_initialization_26_E.o generated by GDC-0.20 from http://dstress.kuehne.cn/run/a/array_initialization_26_E.d	application/octet-stream	1612

Comments

Comment #0 by thomas-dloop — 2006-12-25T11:01:26Z

incorrect code generation if compiled with "-O": # # void main(){ # int[][] a = new int [][](1, 1); # } # always incorrect code generation: # # void main(){ # float[][] a = new float [][](1, 1); # } # test cases: http://dstress.kuehne.cn/run/a/array_initialization_26_B.d http://dstress.kuehne.cn/run/a/array_initialization_26_C.d http://dstress.kuehne.cn/run/a/array_initialization_26_E.d

Comment #1 by bugzilla — 2007-01-02T00:24:51Z

It works when I try it, with and without -O. Also, the float array is (and should be) initialized to float.nan, not 0.

Comment #2 by thomas-dloop — 2007-01-02T09:17:55Z

Created attachment 88 Attached array_initialization_26_E.o generated by GDC-0.20 from http://dstress.kuehne.cn/run/a/array_initialization_26_E.d

Comment #3 by dvdfrdmn — 2007-03-10T10:27:06Z

Fixed in 0.23. The array_initialization_26_E case is invalid: (0.0 !<>= float.nan) should test as true.

Comment #4 by thomas-dloop — 2007-03-12T00:29:22Z

Thanks, Fixed