| Age | Commit message (Collapse) | Author |
|---|
|
|
|
We typecheck arguments like previously, using bidirectionality hints,
but ultimately replace them with user-provided arguments on which we
replay coercion traces.
This is a fix which should be easy to backport, but there are two
directions of future work:
- Coercion traces for `Program` coercions (in these cases, we currently
use the inferred arguments)
- Separate the Coercion API in two phases: inference and application of
coercions. It will make the approach taken here cleaner, and probably
make it easier to interleave typing steps with coercion inference.
Co-Authored-By: Gaëtan Gilbert <gaetan.gilbert@skyskimmer.net>
|
|
Reviewed-by: ejgallego
Reviewed-by: herbelin
|
|
clauses in pattern matching decompilation algorithm
Ack-by: Zimmi48
Reviewed-by: ppedrot
|
|
The `Print Canonical Projections` command now can take constants and prints only
the unification rules that involves or are synthesized from given constants.
|
|
Namely, Evd.evar_env and Evd.evar_filtered_env now take an additional
environment instead of querying the imperative global one. We percolate
this change as higher up as possible.
|
|
We restrict to those that are actually related to typeclasses, and
perform the following renamings:
Classops --> Coercionops
Class --> ComCoercion
|
|
This change improves the relaxed ambiguous path condition of coercions (#9743)
to check that any circular inheritance path of `C >-> C` is definitionally equal
to the identity function of the class `C`. Moreover, for a new inheritance path
`p : C >-> D` and existing (valid) one `q : C >-> D`, the new mechanism does not
report the ambiguity of `p` and `q` if they have a common element, that is to
say:
`p = p1 @ [c] @ p2` and `q = q1 @ [c] @ q2`
for some coercion `c` and inheritance paths `p1`, `p2`, `q1`, and `q2`.
In that case, convertibility of `p1` and `q1`, also, `p2` and `q2` should be
checked; thus, checking the ambiguity of `p` and `q` is redundant with them.
If the new mechanism does not report any ambiguous path, the inheritance graph
must be coherent [Barthe 1995, Sect. 3.2] [Saïbi 1997, Sect. 7]:
1. for any circular path `p : C >-> C`, `p` is definitionally equal to the
identity function, and
2. for any two paths `p, q : C >-> D`, `p` and `q` are convertible.
[Barthe 1995] Gilles Barthe, Implicit coercions in type systems, In: TYPES '95,
LNCS, vol 1158, Springer, 1996, pp 1-15.
[Saïbi 1997] Amokrane Saïbi, Typing algorithm in type theory with inheritance,
In: POPL '97, ACM, 1997, pp 292-301.
|
|
Ack-by: SkySkimmer
Reviewed-by: ejgallego
Reviewed-by: gares
|
|
|
|
|
|
For now, the API is unchanged and this commit only splits pretyping code for
each glob_constr constructor into a record field.
|
|
There are no users in Coq but maybe some plugin wants it so don't
completely remove it
|
|
We renounce to the ad hoc rule preferring a notation w/o delimiter
for a term with coercions stripped over a notation for the
fully-applied terms with coercions not removed.
Instead, we interleave removal of coercions and search for notations:
we prefer a notation for the fully applied term, and, if not, try to
remove one coercion, and try again a notation for the remaining term,
and if not, try to remove the next coercion, etc.
Note: the flatten_application could be removed if prim_token were able
to apply on a prefix of an application node.
|
|
The missing dependency impacted the algorithm for detecting default
clauses.
|
|
(inh_conv_coerces_to is unused so we remove it)
This makes the code simpler, removing dead match branches and Option.maps/gets
|
|
Rels that exist inside the environment at the time of the closure creation
are fetched in the global environment, while we only use the local list of
relevance for FRels. All the infrastructure was implicitly relying on this
kind of behaviour before the introduction of SProp.
Fixes #11150: pattern is 10x slower in Coq 8.10.0
|
|
We also remove trailing whitespace.
Script used:
```bash
for i in `find . -name '*.ml' -or -name '*.mli' -or -name '*.mlg'`; do expand -i "$i" | sponge "$i"; sed -e's/[[:space:]]*$//' -i.bak "$i"; done
```
|
|
* Float added to is_value/get_value to avoid stack overflows
(cf. #7646)
* beware of the use of Array.map with floats (cf. comment in the
makeblock function)
NB: From here one, the configure option "-native-compiler no"
is no longer needed.
|
|
|
|
Replace `option comparison` with `float_comparison` (:= `FEq | FLt |
FGt | FNotComparable`) as suggested by Guillaume Melquiond to avoid
boxing and an extra match when using primitive float comparison.
|
|
* This commit add float instructions to the VM, their encoding in bytecode
and the interpretation of primitive float values after the reduction.
* The flag '-std=c99' could be added to the C compiler flags to ensure
that float computation strictly follows the norm (ie. i387 80-bits
format is not used as an optimization).
Actually, we use '-fexcess-precision=standard' instead of '-std=c99'
because the latter would disable GNU asm used in the VM.
|
|
Beware of 0. = -0. issue for primitive floats
The IEEE 754 declares that 0. and -0. are treated equal but we cannot
say that this is true with Leibniz equality.
Therefore we must patch the equality and the total comparison inside the
kernel to prevent inconsistency.
|
|
If you have access to a kernel name you also should have the
environment in which it is defined, barring hacks. In order to
disfavor hacks we make the standard lookups raise anomalies so that
people are forced to admit they rely on the internals of the
environment.
We find that hackers operated on the code for side effects, for
finding inductive schemes, for simpl and for Print Assumptions. They
attempted to operate on funind but the error handling code they wrote
would have raised another Not_found instead of being useful.
All these uses are indeed hacky so I am satisfied that we are not
forcing new hacks on callers.
|
|
More precisely: The equivalent in #7288 (4dab4fc) to the former call
to ltac_interp_name_env was not done anymore when interpreting
uconstr's.
|
|
|
|
This is not quite pretty, but it is needed by the section mechanism to rebuild
an inductive when closing a section. Hopefully this can be moved back at some
point.
|
|
|
|
Reviewed-by: SkySkimmer
|
|
Reviewed-by: mattam82
|
|
We move `binder_kind` to the pretyping AST, removing the last data
type in the now orphaned file `Decl_kinds`.
This seems a better fit, as this data is not relevant to the lower
layers but only used in `Impargs`.
We also move state keeping to `Impargs`, so now implicit declaration
must include the type. We also remove a duplicated function.
|
|
This file is unrelated to library, but to pretyping/unification.
This commit, along with others already submitted go towards the goal
of `library` containing only the handling of library objects.
|
|
This source of slowness has been observed in VST, but it is probably
pervasive. Most of the unification problems are not mentioning evars,
it is thus useless to compute the set of frozen evars upfront.
We also seize the opportunity to reverse the flag, because it is always
used negatively.
|
|
We split the function used to retrieve the local context from the one used to
provide the implicit status of each binder. Most of the users only rely on the
former indeed.
|
|
rewrite
Reviewed-by: gares
Reviewed-by: ppedrot
|
|
Fixes #10300 and #10285.
|
|
|
|
|
|
Reviewed-by: SkySkimmer
Ack-by: ppedrot
|
|
Not pretty, but it had to be done some day, as `Globnames` seems to be
on the way out.
I have taken the opportunity to reduce the number of `open` in the
codebase.
The qualified style would indeed allow us to use a bit nicer names
`GlobRef.Inductive` instead of `IndRef`, etc... once we have the
tooling to do large-scale refactoring that could be tried.
|
|
The changes are large due to `Pervasives` deprecation:
- the `Pervasives` module has been deprecated in favor of `Stdlib`, we
have opted for introducing a few wrapping functions in `Util` and
just unqualified the rest of occurrences. We avoid the shims as in
the previous attempt.
- a bug regarding partial application have been fixed.
- some formatting functions have been deprecated, but previous
versions don't include a replacement, thus the warning has been
disabled.
We may want to clean up things a bit more, in particular
w.r.t. modules once we can move to OCaml 4.07 as the minimum required
version.
Note that there is a clash between 4.08.0 modules `Option` and `Int`
and Coq's ones. It is not clear if we should resolve that clash or
not, see PR #10469 for more discussion.
On the good side, OCaml 4.08.0 does provide a few interesting
functionalities, including nice new warnings useful for devs.
|
|
Reviewed-by: maximedenes
Reviewed-by: ppedrot
|
|
Reviewed-by: SkySkimmer
Ack-by: herbelin
|
|
new one
|
|
See discussion in #10417
|
|
Ack-by: SkySkimmer
Reviewed-by: ejgallego
Reviewed-by: ppedrot
|
|
|
|
Instead we get the symbols from a Environ.env.
We make them accessible to the produced code through a reference
managed by the kernel, similar to the return values except inverting
when it's written and when it's read.
|
|
This replaces the mismatched context error, which occurred when
Instance := {} was used with strictly more fields than declared.
Since we later check that field names match those declared for the
instance, now that we reject duplicates we know that there are no
extra fields.
|
|
This lets us avoid having to cache the SearchBlacklist.elements call
in search as we can just use the set module's for_all function.
|
