Remove obsolete nan-propagation.wast.

[gahaas]

NaN propagation has been removed from the design repo, which made the nan-propagation tests obsolete. this PR removes the nan-propagation tests.

[rossberg]

Does this require changes to the implementation as well? If so, we should perhaps do them together.

In any case, can you please retarget the PR on the binary-0xc branch?

[titzer]

The change landed on master, IIUC. If so, then it's probably best to go
ahead and remove these tests on master, too.

On Tue, Aug 23, 2016 at 11:48 AM, rossberg-chromium <
notifications@github.com> wrote:

Does this require changes to the implementation as well? If so, we should
perhaps do them together.

In any case, can you please retarget the PR on the binary-0xc branch?

—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub
#321 (comment), or mute
the thread
https://github.com/notifications/unsubscribe-auth/ALnq1PZ7x3MxjyzHhjAmXC12UgagS3_aks5qisIGgaJpZM4JquyL
.

[rossberg]

I see. In that case sgtm, but we should merge to binary-0xc after.

@sunfishcode, you probably know best: IIUC the change is purely a relaxation, but is there anything that can/should be simplified in the impl of arithmetics now to reflect it?

[sunfishcode]

This change lgtm.

determine_binary_nan and determine_unary_nan could indeed be simplified, though as you observe, it's not necessary.

I've updated #286 with notes about the remaining missing NaN test coverage.

[rossberg]

@gahaas, feel free to land this. Bonus points if you also make the simplifications that @sunfishcode mentioned (see float.ml). :)