| Commit message (Collapse) | Author | Age |
|
|
|
| |
This seemes to be making it slower though....
|
| |
|
|
|
|
| |
Maybe it'll result in better output code with fewer zeros?
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
| |
Also reimplement it with a shift and a mask
|
| |
|
| |
|
| |
|
| |
|
|
|
|
| |
Need to fix the display of mulsplit and addwithcarry
|
| |
|
| |
|
| |
|
|
|
|
| |
We still need to idtac, because tc resolution eats messages from fail at any level
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
| |
It seems that something gets unfolded which should not get unfolded.
But we no longer block on lists.
|
| |
|
|
|
|
|
|
|
| |
The new parameterized definitions and proofs are in
WordByWord/Abstract/Dependent/*; the old ones are untouched (and unused)
in WordByWord/Abstract/*. I replaced definitions I didn't know how to
write in the Saturated API with the use of an axiom.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
As per
https://github.com/mit-plv/fiat-crypto/pull/203#issuecomment-309060964
and
https://github.com/mit-plv/fiat-crypto/pull/203#issuecomment-309101747
Revert "update ocq2C sed script"
This reverts commit 4a39f39e195b9b7273810a83de78dfd1d150783e.
Revert "make display"
This reverts commit cbf6d013c533d5165d749d0f9405a15d1ff0b43e.
Revert "Make use of CArrayNotations"
This reverts commit cae0e80ae76b76091e7fb86fcd794358a4fe55bb.
Revert "Fix CArrayNotations"
This reverts commit d0d0fbd4499296a2164e209466227892671556f0.
Revert "Revert "Revert "Add CArrayNotations"""
This reverts commit b2b8403ca76f6fd461d9a71ac2e9add4359bba8c.
|
| |
|
|
|
|
|
| |
Note that the sed rules for addcarryx and sbb need to change, to use
arrays.
|
| |
|
|
|
|
|
|
| |
Work around [bug #5608](https://coq.inria.fr/bugs/show_bug.cgi?id=5608),
Anomaly: No printing rule found for _ _ [1] = { _ } ; return ( _ , _ ,
.. , _ ). Please report at http://coq.inria.fr/bugs/.
|
|
|
|
| |
This reverts commit 29ad742d76dca90ec9c8d03ab6f4359ccf053a90.
|
| |
|
| |
|
|
|
|
| |
This change is necessary because we've added -compat 8.6.
|
|
|
|
|
|
|
| |
This reverts commit 44359b29d99ab52154dcfdf2b2b16bca7dbaf339.
It triggers [bug #5469](https://coq.inria.fr/bugs/show_bug.cgi?id=5469),
which is present in 8.6, but not v8.6 (nor 8.6.1, once it comes out).
|
|
|
| |
This closes #182.
|
| |
|
| |
|
|
|
|
|
|
| |
I didn't want to bother redoing all of the proofs that I'd already done,
so instead I prove the cps'ified version equal to the non-cps version,
and transfer over the proofs that way.
|
| |
|
|
|
|
|
|
| |
This allows fiat-crypto to continue working with trunk, after the merge
of https://github.com/coq/coq/pull/220. We will remove this when we
migrate to requiring 8.6.1 or 8.7 (neither of which is released yet).
|
| |
|
|
|
|
| |
This is the lighter-weight solution to #197.
|
|
|
|
| |
This closes #195
|