Update AstSemantics.md

[titzer]

It looks like loads and stores of i64 with sizes < 64bits got dropped on the floor. This PR adds them back and revises the text to no longer refer to memory types.

[sunfishcode]

They accept as input an index and an offset which are described in the Addressing section immediately following this section.

[sunfishcode]

lgtm, other than a comment above.

[titzer]

I just removed the references to indexing, since they are redundant with
the text in the next section.

On Wed, Sep 23, 2015 at 3:19 PM, Dan Gohman notifications@github.com
wrote:

lgtm, other than a comment above.

—
Reply to this email directly or view it on GitHub
#366 (comment).

[lukewagner]

This PR undoes the explicit change in #326/#341/spec/#55 that @rossberg-chromium proposed and everyone (including you) lgtm'd which was to avoid including every possible coercion-on-load except those that were necessary to avoid introducing unwanted intermediate types (int8/int16). I don't have a strong opinion either way, but the change was intentional and we'll need to revise spec/ (again) with this design change.

[titzer]

It was an oversight on my part to not notice that the coercions for i64
were removed in rossberg's proposal.

I think it's important to keep them since otherwise load/stores working
with i64 and smaller sizes would require another explicit coercion.

On Thu, Sep 24, 2015 at 5:05 PM, Luke Wagner notifications@github.com
wrote:

This PR undoes the explicit change in #326
#326
#341
WebAssembly/spec#55 that @rossberg-chromium
https://github.com/rossberg-chromium proposed and everyone (including
you
#326 (comment))
lgtm'd which was to avoid including every possible coercion-on-load except
those that were necessary to avoid introducing unwanted intermediate types
(int8/int16). I don't have a strong opinion either way, but the change was
intentional and we'll need to revise spec/ (again) with this design change.

—
Reply to this email directly or view it on GitHub
#366 (comment).

[lukewagner]

Ok; this is just the usual pattern-match vs. EITBI argument. I generally lean to the latter, so I guess I'm actually happy to see this PR; I mostly just wanted to point out that the change wasn't accidental (at least for some). I can update spec/ to match this.

[lukewagner]

Btw: if you search AstSemantics.md there is only one more dangling reference to memory type (in Addressing) and then we could remove "Memory type".