mit-plv / coqutil Goto Github PK

I have the feeling that using tuples in that file was an unfortunate choice and benefits almost nothing; getting rid of it would be a clear improvement. @samuelgruetter do you agree? If so, plan:

1. define LittleEndian.combine and LittleEndian.split in terms of LittleEndianList.combine and LittleEndianList.split
2. prove the lemmas we currently have about them? this might actually be more annoying than the next point
3. change the callers to use LittleEndianList directly
4. atomic rename LittleEndianList -> LittleEndian across all repositories (but we get most of the benefit even if we don't do this)

Thoughts?

supersedes #34

Quoting Enrico in coq/opam#914 (comment):

For a .dev package I believe it is fine, but if you want to put it under released at some point please craft a more precise description of it.

This is in no way urgent, but it's possible we'll at some point want to release fiat-crypto, and if the released fiat-crypto depends on coqutil via bedrock2, this will then be a blocker for releasing fiat-crypto on opam.

If we want to use coqutil in fiat-crypto, we need some subset (and target) of it that builds with Coq 8.9.

In Coq 8.12, build failures are documented in #38
In Coq 8.11, build failures are documented in #40
In Coq 8.10, there's at least one issue with lia in List.v, Ltac2 does not exist, and neither does Coq.micromega.ZifyClasses
In Coq 8.9, skipn_cons and Z.div_mod_to_equations don't exist

cc @andres-erbsen @samuelgruetter

I see

coq_makefile -f _CoqProject INSTALLDEFAULTROOT = coqutil  /home/jgross/Documents/repos/fiat-crypto/bedrock2/deps/coqutil/src/coqutil/Macros/symmetry.v /home/jgross/Docu
ments/repos/fiat-crypto/bedrock2/deps/coqutil/src/coqutil/Macros/unique.v /home/jgross/Documents/repos/fiat-crypto/bedrock2/deps/coqutil/src/coqutil/Macros/subst.v /hom
e/jgross/Documents/repos/fiat-crypto/bedrock2/deps/coqutil/src/coqutil/Word/Interface.v /home/jgross/Documents/repos/fiat-crypto/bedrock2/deps/coqutil/src/coqutil/Word/
LittleEndian.v /home/jgross/Documents/repos/fiat-crypto/bedrock2/deps/coqutil/src/coqutil/Word/Properties.v /home/jgross/Documents/repos/fiat-crypto/bedrock2/deps/coqut
il/src/coqutil/Word/Naive.v /home/jgross/Documents/repos/fiat-crypto/bedrock2/deps/coqutil/src/coqutil/Map/Interface.v /home/jgross/Documents/repos/fiat-crypto/bedrock2
/deps/coqutil/src/coqutil/Map/Solver.v /home/jgross/Documents/repos/fiat-crypto/bedrock2/deps/coqutil/src/coqutil/Map/SlowGoals.v /home/jgross/Documents/repos/fiat-cryp
to/bedrock2/deps/coqutil/src/coqutil/Map/Z_keyed_SortedListMap.v /home/jgross/Documents/repos/fiat-crypto/bedrock2/deps/coqutil/src/coqutil/Map/SortedList.v /home/jgros
s/Documents/repos/fiat-crypto/bedrock2/deps/coqutil/src/coqutil/Map/Funext.v /home/jgross/Documents/repos/fiat-crypto/bedrock2/deps/coqutil/src/coqutil/Map/SortedListWo
rd.v /home/jgross/Documents/repos/fiat-crypto/bedrock2/deps/coqutil/src/coqutil/Map/SortedListString.v /home/jgross/Documents/repos/fiat-crypto/bedrock2/deps/coqutil/sr
c/coqutil/Map/TestLemmas.v /home/jgross/Documents/repos/fiat-crypto/bedrock2/deps/coqutil/src/coqutil/Map/Properties.v /home/jgross/Documents/repos/fiat-crypto/bedrock2
/deps/coqutil/src/coqutil/Map/Empty_set_keyed_map.v /home/jgross/Documents/repos/fiat-crypto/bedrock2/deps/coqutil/src/coqutil/Map/SortedListString_test.v /home/jgross/
Documents/repos/fiat-crypto/bedrock2/deps/coqutil/src/coqutil/Map/TestGoals.v /home/jgross/Documents/repos/fiat-crypto/bedrock2/deps/coqutil/src/coqutil/Z/PushPullMod.v
 /home/jgross/Documents/repos/fiat-crypto/bedrock2/deps/coqutil/src/coqutil/Z/ZLib.v /home/jgross/Documents/repos/fiat-crypto/bedrock2/deps/coqutil/src/coqutil/Z/bitbla
st.v /home/jgross/Documents/repos/fiat-crypto/bedrock2/deps/coqutil/src/coqutil/Z/div_to_equations.v /home/jgross/Documents/repos/fiat-crypto/bedrock2/deps/coqutil/src/
coqutil/Z/HexNotation.v /home/jgross/Documents/repos/fiat-crypto/bedrock2/deps/coqutil/src/coqutil/Z/Lia.v /home/jgross/Documents/repos/fiat-crypto/bedrock2/deps/coquti
l/src/coqutil/Z/div_mod_to_equations.v /home/jgross/Documents/repos/fiat-crypto/bedrock2/deps/coqutil/src/coqutil/Z/BitOps.v /home/jgross/Documents/repos/fiat-crypto/be
drock2/deps/coqutil/src/coqutil/Tactics/rdelta.v /home/jgross/Documents/repos/fiat-crypto/bedrock2/deps/coqutil/src/coqutil/Tactics/syntactic_unify.v /home/jgross/Docum
ents/repos/fiat-crypto/bedrock2/deps/coqutil/src/coqutil/Tactics/letexists.v /home/jgross/Documents/repos/fiat-crypto/bedrock2/deps/coqutil/src/coqutil/Tactics/autoforw
ard.v /home/jgross/Documents/repos/fiat-crypto/bedrock2/deps/coqutil/src/coqutil/Tactics/Tactics.v /home/jgross/Documents/repos/fiat-crypto/bedrock2/deps/coqutil/src/co
qutil/Tactics/destr.v /home/jgross/Documents/repos/fiat-crypto/bedrock2/deps/coqutil/src/coqutil/Tactics/eabstract.v /home/jgross/Documents/repos/fiat-crypto/bedrock2/d
eps/coqutil/src/coqutil/Tactics/simpl_rewrite.v /home/jgross/Documents/repos/fiat-crypto/bedrock2/deps/coqutil/src/coqutil/sanity.v /home/jgross/Documents/repos/fiat-cr
ypto/bedrock2/deps/coqutil/src/coqutil/dlet.v /home/jgross/Documents/repos/fiat-crypto/bedrock2/deps/coqutil/src/coqutil/Decidable.v /home/jgross/Documents/repos/fiat-c
rypto/bedrock2/deps/coqutil/src/coqutil/Datatypes/PropSet.v /home/jgross/Documents/repos/fiat-crypto/bedrock2/deps/coqutil/src/coqutil/Datatypes/PrimitivePair.v /home/j
gross/Documents/repos/fiat-crypto/bedrock2/deps/coqutil/src/coqutil/Datatypes/HList.v /home/jgross/Documents/repos/fiat-crypto/bedrock2/deps/coqutil/src/coqutil/Datatyp
es/String.v /home/jgross/Documents/repos/fiat-crypto/bedrock2/deps/coqutil/src/coqutil/Datatypes/Prod.v /home/jgross/Documents/repos/fiat-crypto/bedrock2/deps/coqutil/s
rc/coqutil/Datatypes/List.v /home/jgross/Documents/repos/fiat-crypto/bedrock2/deps/coqutil/src/coqutil/Datatypes/Option.v -o Makefile.coq.all

every time I run make, and it's a bit unpleasant. Please suppress the output, at least whenever VERBOSE is empty.
(You can copy this snipped to allow conditional showing/hiding:

VERBOSE?=
SHOW := $(if $(VERBOSE),@true "",@echo "")
HIDE := $(if $(VERBOSE),,@)

and then do

foo:
	$(SHOW)short description
	$(HIDE)verbose actual command

Fiat Crypto no longer builds with the newest version of coqutil. What happened?

Hi folks, I see your makefile has:

# absolute paths so that emacs compile mode knows where to find error
# use cygpath -m because Coq on Windows cannot handle cygwin paths
SRCDIR := $(shell cygpath -m "$$(pwd)" 2>/dev/null || pwd)/src

then you do

	@$(COQ_MAKEFILE) $(ALL_VS) -o Makefile.coq.all

with the above files, being absolute paths.

Using absolute paths is really not supported by coq_makefile (paths should be relative to the flags given in _CoqProject) , it works because coqdep implements some (fragile magic) but we'd like to get rid of it.

What is this stuff about emacs needing absolute paths? cc @samuelgruetter

It's breaking the bedrock2 dev package (because coqutil.dev gets source from coqutil tested branch)
alternatively you can change the dev package to use master

LittleEndian.combine should not inline tuple.to_list. It could just take a list as input and have a simpler type.

We can fix the first one with

diff --git a/Makefile b/Makefile
index d87b01a..a84f2b5 100644
--- a/Makefile
+++ b/Makefile
@@ -1,6 +1,6 @@
 default_target: all

-.PHONY: clean force all install uninstall
+.PHONY: clean force all install uninstall validate

 # absolute paths so that emacs compile mode knows where to find error
 # use cygpath -m because Coq on Windows cannot handle cygwin paths
@@ -31,3 +31,6 @@ install::

 uninstall::
	$(MAKE) -f Makefile.coq.all uninstall
+
+validate:
+	$(MAKE) -f Makefile.coq.all validate

but the second is a bug in Coq coq/coq#12066 that coqutil hits because it's using absolute paths.

Please add an AUTHORS file and a LICENSE file; I cannot create an opam package if I don't know which authors to list and what license to use.

Should any/all/most of the fiat-crypto src/Util directory be migrated to coqutil? (Currently this directory is more-or-less replicated across fiat, rewriter, and fiat-crypto, with fiat-crypto containing the most complete and up-to-date version.)

An incomplete list of tactics, roughly in order from most-used to least-used:

• destructing discriminees of match subject to various conditions such as "only this type of variable", "only terms which contain no matches inside themselves", "only in the hyps", "only in the conclusion", "only variables" (with intelligent insertion of equations)
• destructing or inverting all hypotheses of a given type family, with performance-optimized variants for common heads
• [a tactic for specializing all hypotheses with non-dependent premises which can be proven by a given tactic (such as assumption, auto, lia, etc))[https://github.com/mit-plv/fiat-crypto/blob/master/src/Util/Tactics/SpecializeBy.v]
• variants of pose and assert which are suitable for use in repeat blocks by ensuring that they don't pose duplicates
• split hypotheses using iff, and, produnderneath binders in a way that preserves the binders, e.g., turningforall x y, P x y /\ Q x yintoforall x y, P x yandforall x y, Q x y`
• finding the head of a term
• attempting to rewrite with any hypothesis that can be rewritten with
• call set or subst on all evars
• a variant of subst that works modulo relations
• printing the context and the goal
• testing if an identifier has a body / can be unfolded
• calling subst on all local definitions
• transparent assert
• a tactic to evar-normalize terms
• a variant of etransitivity that accepts arguments
• run a tactic on any hypothesis that it succeeds on
• a destruct that attempts to work with convoy-pattern terms appearing in the goal
• especialize
• a tactic for getting the goal in fewer characters
• general normalization of hyps and goal with respect to a commutativity lemma
• get all instances of a typeclass (family)
• tac_on_subterms, to be used as e.g. tac_on_subterms ring_simplify
• revert non-dependent hypotheses, roughly the inverse of intros *
• tactics for running tactics while returning constrs and a tryif for returning constrs
• replace_with_vm_compute for getting around vm_compute not inserting casts
• saturate the context with all ways of specializing hypotheses with other hyps
• simplifications of nested if statements

An incomplete list of other utilities:

• A large but not-especially-coherently-organized library of utility lemmas about Z, including automation for:
  • named hint databases that can be used to have (e)auto invoke lia, nia, or lra at the leaves
  • rewrite databases for distributing and factoring various operations, the most useful of which is Z.modulo, which is factored into a tactic
  • general simplification rewrite databases, divided into categories of (a) general simplification that's reversible, (b) ones that further don't require any side conditions, and (c) ones that strip out operations on constants
  • a version of subst which will solve linear equations in variables over Z to substitute
  • tactics for making progress on or simplifying fractions, inequalities, bounds, min/max, Z.testbit
• smaller libraries for lemmas about nat, list, N, Q, MSets
• a general error monad
• a library of lemmas about reflect culminating in a general tactic for turning boolean equality hypotheses into propositions
• a collection of lemmas about string, including
  • a small parser combinator library used in fiat-crypto to parse command-line options and assembly files
  • typeclasses for Show, ShowLevel (for printing with precedence), and ShowLines (printing to lists of strings rather than single strings, for better efficiency) and theory around them
  • sorting
• telescopes and wf and more facts
• a port and adaptation of OCaml's Arg module (maybe this should be split off to a separate project and added on opam? (perhaps to be combined with the method for extracting main functions and executables in OCaml and Haskell?)
• notations for CPS
• function composition and (un)currying
• theory of decidable propositions and relations
• a class for types with default elements
  • a variant for "pointed" types and props
• heterogeneous lists indexed over a list of types (and a primitive projections variant)
• heterogeneous lists not indexed over a list of types
• length-indexed lists/tuples
• some theory about eq
• factorization of Ns into primes
• lemmas and tactics about subsingleton types, including a tactic to replace proofs of x = x with eq_refl when equality on x is decidable
• the Let_In constant for not inlining let ... in ... and a monad around this
• theory of finitely listable types
• fueled loops
• some hints about PERs
• basic theory, including specialized inversion lemmas and tactics (a la inversion_sigma) for option, nat, list, prod, prod with primitive projections, sum, primitive projection versions of sig, sigT, etc, sumbool
• reflective simpliciation of Props
• theory around option (list _) and list (option _)

Should https://github.com/mit-plv/bedrock2/blob/master/bedrock2/src/bedrock2/ZnWords.v be adapted to hook into the zify mechanism a la https://github.com/coq/coq/blob/master/theories/micromega/ZifyUint63.v and be integrated into coqutil?

@samuelgruetter

As sign extension goes from one word size to another word size, shouldn't it be a derived function from a word of one size to a word of another size?

The Coq team released Coq 8.16+rc1 on June 01, 2022.
The corresponding Coq Platform release 2022.09 should be released before September 15, 2022.
It can be delayed in case of difficulties until October 15, 2022, but this should be an exception.

This issue is to inform you that the opam package we are currently testing in Coq Platform CI works fine with Coq 8.16+rc1.

Coq Platform CI is currently testing opam package coq-coqutil.0.0.1
from official repository https://coq.inria.fr/opam/released/packages/coq-coqutil/coq-coqutil.0.0.1/opam.

In case this is the version you want to see in Coq Platform, there is nothing to do for you - please just close this issue.

In case you would prefer to see an updated or an older version in the upcoming Coq Platform 2022.09, please inform us as soon as possible and before August 31, 2022!

The working branch of Coq Platform, can be found here main.
It contains package pick ~8.16+rc1~2022.09~preview1 which already supports Coq version 8.16+rc1 and contains already working (possibly patched / commit pinned) Coq Platform packages.

In case you want to select a different version, please don't close this issue, even after creating the new tag and/or opam package.
We will close the issue after updating Coq Platform.
This is a measure of 'double book keeping' in order to avoid that we ship the wrong version.

In any case, Coq Platform won't be released before this issue is closed!

Thanks!

P.S.: this issue has been created automatically based on CI status.

CC: coq/platform#274

coqchk gives:

CONTEXT SUMMARY
===============

* Theory: Set is predicative

* Axioms:
    Coq.ssr.ssrunder.Under_rel.Over_rel
    Coq.ssr.ssrunder.Under_rel.Under_rel_from_rel
    Coq.ssr.ssrunder.Under_rel.over_rel
    Coq.Logic.FunctionalExtensionality.functional_extensionality_dep
    coqutil.Map.SortedList.TODO_andres
    Coq.ssr.ssrunder.Under_rel.over_rel_done
    Coq.ssr.ssrunder.Under_rel.Under_rel
    Coq.Logic.PropExtensionality.propositional_extensionality
    Coq.ssr.ssrunder.Under_rel.Under_relE
    coqutil.Map.SortedListString.TODO_andres
    Coq.ssr.ssrunder.Under_rel.under_rel_done

* Constants/Inductives relying on type-in-type: <none>

* Constants/Inductives relying on unsafe (co)fixpoints: <none>

* Inductives whose positivity is assumed: <none>

You can ignore the ssr.ssrunder ones (coq/coq#5030), but it'd be nice to remove the other ones.

The Coq team released Coq 8.18.0 on September 7th, 2023.
The corresponding Coq Platform release 2023.10 should be released before November 30th, 2023.
It can be delayed in case of difficulties until January 15th, 2023, but this should be an exception.

This issue is to inform you that the opam package we are currently testing in Coq Platform CI works fine with Coq 8.18.0.

Coq Platform CI is currently testing opam package coq-coqutil.0.0.3
from official repository https://coq.inria.fr/opam/released/packages/coq-coqutil/coq-coqutil.0.0.3/opam.

In case this is the version you want to see in Coq Platform, there is nothing to do for you - please just close this issue.

In case you would prefer to see an updated or an older version in the upcoming Coq Platform 2023.10, please inform us as soon as possible and before October 31st, 2023!

The working branch of Coq Platform, can be found here main.
It contains package pick ~8.18+beta1 which already supports Coq version 8.18.0 and contains already working (possibly patched / commit pinned) Coq Platform packages.

In case you want to select a different version, please don't close this issue, even after creating the new tag and/or opam package.
We will close the issue after updating Coq Platform.
This is a measure of 'double book keeping' in order to avoid that we ship the wrong version.

In any case, Coq Platform won't be released before this issue is closed!

Thanks!

P.S.: this issue has been created automatically based on CI status.

CC: coq/platform#372

We have

This breaks testing fiat-crypto's make install target on docker: https://github.com/mit-plv/fiat-crypto/actions/runs/6909719309/job/18801575975?pr=1733#step:18:3269

This is currently blocking mit-plv/fiat-crypto#686 (cc @jadephilipoom )

If I do

Require Import Coq.extraction.Extraction.
Require Import coqutil.Map.SortedListString.
Extraction Language OCaml.
Recursive Extraction coqutil.Map.SortedListString.map.

I get:

type __ = Obj.t

type bool =
| True
| False

(** val negb : bool -> bool **)

let negb = function
| True -> False
| False -> True

type 'a option =
| Some of 'a
| None

type ('a, 'b) prod =
| Pair of 'a * 'b

(** val fst : ('a1, 'a2) prod -> 'a1 **)

let fst = function
| Pair (x, _) -> x

type 'a list =
| Nil
| Cons of 'a * 'a list

type comparison =
| Eq
| Lt
| Gt

(** val eqb : bool -> bool -> bool **)

let eqb b1 b2 =
  match b1 with
  | True -> b2
  | False -> (match b2 with
              | True -> False
              | False -> True)

(** val fold_right : ('a2 -> 'a1 -> 'a1) -> 'a1 -> 'a2 list -> 'a1 **)

let rec fold_right f a0 = function
| Nil -> a0
| Cons (b, t) -> f b (fold_right f a0 t)

(** val find : ('a1 -> bool) -> 'a1 list -> 'a1 option **)

let rec find f = function
| Nil -> None
| Cons (x, tl) -> (match f x with
                   | True -> Some x
                   | False -> find f tl)

type positive =
| XI of positive
| XO of positive
| XH

type n =
| N0
| Npos of positive

module Pos =
 struct
  (** val succ : positive -> positive **)

  let rec succ = function
  | XI p -> XO (succ p)
  | XO p -> XI p
  | XH -> XO XH

  (** val add : positive -> positive -> positive **)

  let rec add x y =
    match x with
    | XI p ->
      (match y with
       | XI q -> XO (add_carry p q)
       | XO q -> XI (add p q)
       | XH -> XO (succ p))
    | XO p ->
      (match y with
       | XI q -> XI (add p q)
       | XO q -> XO (add p q)
       | XH -> XI p)
    | XH -> (match y with
             | XI q -> XO (succ q)
             | XO q -> XI q
             | XH -> XO XH)

  (** val add_carry : positive -> positive -> positive **)

  and add_carry x y =
    match x with
    | XI p ->
      (match y with
       | XI q -> XI (add_carry p q)
       | XO q -> XO (add_carry p q)
       | XH -> XI (succ p))
    | XO p ->
      (match y with
       | XI q -> XO (add_carry p q)
       | XO q -> XI (add p q)
       | XH -> XO (succ p))
    | XH -> (match y with
             | XI q -> XI (succ q)
             | XO q -> XO (succ q)
             | XH -> XI XH)

  (** val mul : positive -> positive -> positive **)

  let rec mul x y =
    match x with
    | XI p -> add y (XO (mul p y))
    | XO p -> XO (mul p y)
    | XH -> y

  (** val compare_cont : comparison -> positive -> positive -> comparison **)

  let rec compare_cont r x y =
    match x with
    | XI p ->
      (match y with
       | XI q -> compare_cont r p q
       | XO q -> compare_cont Gt p q
       | XH -> Gt)
    | XO p ->
      (match y with
       | XI q -> compare_cont Lt p q
       | XO q -> compare_cont r p q
       | XH -> Gt)
    | XH -> (match y with
             | XH -> r
             | _ -> Lt)

  (** val compare : positive -> positive -> comparison **)

  let compare =
    compare_cont Eq
 end

module N =
 struct
  (** val compare : n -> n -> comparison **)

  let compare n0 m =
    match n0 with
    | N0 -> (match m with
             | N0 -> Eq
             | Npos _ -> Lt)
    | Npos n' -> (match m with
                  | N0 -> Gt
                  | Npos m' -> Pos.compare n' m')

  (** val ltb : n -> n -> bool **)

  let ltb x y =
    match compare x y with
    | Lt -> True
    | _ -> False
 end

module Coq_N =
 struct
  (** val add : n -> n -> n **)

  let add n0 m =
    match n0 with
    | N0 -> m
    | Npos p -> (match m with
                 | N0 -> n0
                 | Npos q -> Npos (Pos.add p q))

  (** val mul : n -> n -> n **)

  let mul n0 m =
    match n0 with
    | N0 -> N0
    | Npos p -> (match m with
                 | N0 -> N0
                 | Npos q -> Npos (Pos.mul p q))
 end

type ascii =
| Ascii of bool * bool * bool * bool * bool * bool * bool * bool

(** val eqb0 : ascii -> ascii -> bool **)

let eqb0 a b =
  let Ascii (a0, a1, a2, a3, a4, a5, a6, a7) = a in
  let Ascii (b0, b1, b2, b3, b4, b5, b6, b7) = b in
  (match match match match match match match eqb a0 b0 with
                                       | True -> eqb a1 b1
                                       | False -> False with
                                 | True -> eqb a2 b2
                                 | False -> False with
                           | True -> eqb a3 b3
                           | False -> False with
                     | True -> eqb a4 b4
                     | False -> False with
               | True -> eqb a5 b5
               | False -> False with
         | True -> eqb a6 b6
         | False -> False with
   | True -> eqb a7 b7
   | False -> False)

(** val n_of_digits : bool list -> n **)

let rec n_of_digits = function
| Nil -> N0
| Cons (b, l') ->
  Coq_N.add (match b with
             | True -> Npos XH
             | False -> N0) (Coq_N.mul (Npos (XO XH)) (n_of_digits l'))

(** val n_of_ascii : ascii -> n **)

let n_of_ascii = function
| Ascii (a0, a1, a2, a3, a4, a5, a6, a7) ->
  n_of_digits (Cons (a0, (Cons (a1, (Cons (a2, (Cons (a3, (Cons (a4, (Cons (a5,
    (Cons (a6, (Cons (a7, Nil))))))))))))))))

type string =
| EmptyString
| String of ascii * string

module Coq_map =
 struct
  type ('key, 'value) map = { get : (__ -> 'key -> 'value option); empty : 
                              __; put : (__ -> 'key -> 'value -> __);
                              remove : (__ -> 'key -> __);
                              fold : (__ -> (__ -> 'key -> 'value -> __) -> __ ->
                                     __ -> __) }
 end

module Coq_parameters =
 struct
  type parameters =
    __ -> __ -> bool
    (* singleton inductive, whose constructor was Build_parameters *)

  type key = __

  type value = __

  (** val ltb : parameters -> key -> key -> bool **)

  let ltb parameters0 =
    parameters0
 end

(** val eqb1 :
    Coq_parameters.parameters -> Coq_parameters.key -> Coq_parameters.key -> bool **)

let eqb1 p k1 k2 =
  match negb (Coq_parameters.ltb p k1 k2) with
  | True -> negb (Coq_parameters.ltb p k2 k1)
  | False -> False

(** val put0 :
    Coq_parameters.parameters -> (Coq_parameters.key, Coq_parameters.value) prod
    list -> Coq_parameters.key -> Coq_parameters.value -> (Coq_parameters.key,
    Coq_parameters.value) prod list **)

let rec put0 p m k v =
  match m with
  | Nil -> Cons ((Pair (k, v)), Nil)
  | Cons (y, m') ->
    let Pair (k', v') = y in
    (match Coq_parameters.ltb p k k' with
     | True -> Cons ((Pair (k, v)), m)
     | False ->
       (match Coq_parameters.ltb p k' k with
        | True -> Cons ((Pair (k', v')), (put0 p m' k v))
        | False -> Cons ((Pair (k, v)), m')))

(** val remove0 :
    Coq_parameters.parameters -> (Coq_parameters.key, Coq_parameters.value) prod
    list -> Coq_parameters.key -> (Coq_parameters.key, Coq_parameters.value) prod
    list **)

let rec remove0 p m k =
  match m with
  | Nil -> Nil
  | Cons (y, m') ->
    let Pair (k', v') = y in
    (match Coq_parameters.ltb p k k' with
     | True -> m
     | False ->
       (match Coq_parameters.ltb p k' k with
        | True -> Cons ((Pair (k', v')), (remove0 p m' k))
        | False -> m'))

type rep =
  (Coq_parameters.key, Coq_parameters.value) prod list
  (* singleton inductive, whose constructor was Build_rep *)

(** val value0 :
    Coq_parameters.parameters -> rep -> (Coq_parameters.key, Coq_parameters.value)
    prod list **)

let value0 _ r =
  r

(** val map0 :
    Coq_parameters.parameters -> (Coq_parameters.key, Coq_parameters.value)
    Coq_map.map **)

let map0 p =
  let wrapped_put = fun m k v -> put0 p (value0 p m) k v in
  let wrapped_remove = fun m k -> remove0 p (value0 p m) k in
  { Coq_map.get = (fun m k ->
  match find (fun p0 -> eqb1 p k (fst p0)) (value0 p (Obj.magic m)) with
  | Some p0 -> let Pair (_, v) = p0 in Some v
  | None -> None); Coq_map.empty = (Obj.magic Nil); Coq_map.put =
  (Obj.magic wrapped_put); Coq_map.remove = (Obj.magic wrapped_remove);
  Coq_map.fold = (fun _ f r0 m ->
  fold_right (fun pat r -> let Pair (k, v) = pat in f r k v) r0
    (value0 p (Obj.magic m))) }

module Ascii =
 struct
  (** val ltb : ascii -> ascii -> bool **)

  let ltb c d =
    N.ltb (n_of_ascii c) (n_of_ascii d)
 end

(** val ltb0 : string -> string -> bool **)

let rec ltb0 a b =
  match a with
  | EmptyString -> (match b with
                    | EmptyString -> False
                    | String (_, _) -> True)
  | String (x, a') ->
    (match b with
     | EmptyString -> False
     | String (y, b') ->
       (match eqb0 x y with
        | True -> ltb0 a' b'
        | False -> Ascii.ltb x y))

(** val build_parameters : Coq_parameters.parameters **)

let build_parameters =
  Obj.magic ltb0

(** val map1 : (Coq_parameters.key, Coq_parameters.value) Coq_map.map **)

let map1 =
  map0 build_parameters

Running ocamlc foo.ml gives:

File "foo.ml", line 356, characters 4-8:
Error: The type of this expression, ('_weak1, '_weak2) Coq_map.map,
       contains type variables that cannot be generalized

The Coq team released Coq 8.17+rc1 on December 29th, 2022 and plans to release Coq 8.17.0 around March 7th, 2023.
The corresponding Coq Platform release 2023.03 should be released before April 14th, 2023.
It can be delayed in case of difficulties until June 30, 2023, but this should be an exception.

This issue is to inform you that the opam package we are currently testing in Coq Platform CI works fine with Coq 8.17+rc1.

Coq Platform CI is currently testing opam package coq-coqutil.0.0.2
from official repository https://coq.inria.fr/opam/released/packages/coq-coqutil/coq-coqutil.0.0.2/opam.

In case this is the version you want to see in Coq Platform, there is nothing to do for you - please just close this issue.

In case you would prefer to see an updated or an older version in the upcoming Coq Platform 2023.03, please inform us as soon as possible and before March 21st, 2023!

The working branch of Coq Platform, can be found here main.
It contains package pick ~8.17~2023.03+preview1 which already supports Coq version 8.17+rc1 and contains already working (possibly patched / commit pinned) Coq Platform packages.

In case you want to select a different version, please don't close this issue, even after creating the new tag and/or opam package.
We will close the issue after updating Coq Platform.
This is a measure of 'double book keeping' in order to avoid that we ship the wrong version.

In any case, Coq Platform won't be released before this issue is closed!

Thanks!

P.S.: this issue has been created automatically based on CI status.

CC: coq/platform#335

I get an email from GitHub actions every time the tested branch fails to update, probably because I'm the person who last edited the .yml file. @samuelgruetter could you edit it, so I stop getting these emails and they go to someone who can make better use of them instead?

COQC /home/runner/work/coqutil/coqutil/src/coqutil/Datatypes/ListSet.v
Error: Tactic failure: Proof search failed.

Please fix, or revert 4986e26. This breaks fiat-crypto's opam package CI

map solver tests take a while compared to rest of the repo