[ARM] Fix phobos unit tests #1861
Conversation
| @@ -5793,7 +5793,7 @@ template Largest(T...) if(T.length >= 1) | |||
|
|
|||
| unittest | |||
| { | |||
| static assert(is(Largest!(uint, ubyte, ulong, real) == real)); | |||
| static assert(is(Largest!(uint, ubyte, real) == real)); | |||
There was a problem hiding this comment.
Maybe (ubyte, ushort, uint, real) == real ?
There was a problem hiding this comment.
I added the ushort type, but I think the values were intentionally not sorted by size?
|
Other than nits, LTGM. |
| @@ -2605,7 +2605,7 @@ unittest | |||
| assert(to!Float("123e+2") == Literal!Float(123e+2)); | |||
| assert(to!Float("123e-2") == Literal!Float(123e-2)); | |||
| assert(to!Float("123.") == Literal!Float(123.0)); | |||
| assert(to!Float(".375") == Literal!Float(.375)); | |||
| assert(to!Float(".125") == Literal!Float(.125)); | |||
There was a problem hiding this comment.
This fails on ARM? I'd expect an exact conversion for all floating point sizes, no?
There was a problem hiding this comment.
No, that was a merge fail on my part. The unit test originally used 0.456 and that failed on ARM.
I then opened a pull request to change this to 0.375:
d44104e
But I somehow used 0.125 in the ARM port and it got merged into this change. Anyway, I'll remove this change.
|
ping... |
|
thanks! |
Changes to make the unit tests pass on ARM. Mostly reducing required precision for math tests,
adding replacements for tests which required 80 bit reals and so on.
Note: This is part of a set of changes which are required to get test suite & unit tests passing on ARM GDC. All necessary changes for GDC are here for reference:
https://github.com/jpf91/GDC/commits/arm-old
Once these pull requests are merged upstream I'll merge them into GDC as well.