| Age | Commit message (Collapse) | Author |
|---|
|
Reviewed-by: SkySkimmer
Reviewed-by: gares
|
|
This was virtually dead code. The only place really accessing this data was the
user pretty-printer, but actually the tables were not installed for vanilla vo
files.
In practice, that meant that the only case where an access to this table could
have been triggered would have been to print a term coming from a vio file,
or a vo file generated via vio2vo. In all other cases, the printer would not
have displayed the internal universes. While the latter might be considered
a bug, I am instead convinced that this notion of user-facing internal universes
needs to be handled by another mechanism, the current one making little sense.
The fact it was broken all along without anybody noticing proves my point.
|
|
|
|
Before, we would store futures, but it was actually ensured by the
upper layers that they were either evaluated or stored by the STM
somewhere else. We simply replace this type with an option, thus
removing the Future.computation type from vo/vio files.
This also enhances debug printing, as the latter is unable to properly
print futures.
|
|
Kernel should be mostly correct, higher levels do random stuff at
times.
|
|
Note currently it's impossible to define inductives in SProp because
indtypes.ml and the pretyper aren't fully plugged.
|
|
It used to simply remember the normal form of the type of the constructor.
This is somewhat problematic as this is ambiguous in presence of
let-bindings. Rather, we store this data in a fully expanded way, relying
on rel_contexts.
Probably fixes a crapload of bugs with inductive types containing
let-bindings, but it seems that not many were reported in the bugtracker.
|
|
I think the usage looks cleaner this way.
|
|
This work makes it possible to take advantage of a compact
representation for integers in the entire system, as opposed to only
in some reduction machines. It is useful for heavily computational
applications, where even constructing terms is not possible without such
a representation.
Concretely, it replaces part of the retroknowledge machinery with
a primitive construction for integers in terms, and introduces a kind of
FFI which maps constants to operators (on integers). Properties of these
operators are expressed as explicit axioms, whereas they were hidden in
the retroknowledge-based approach.
This has been presented at the Coq workshop and some Coq Working Groups,
and has been used by various groups for STM trace checking,
computational analysis, etc.
Contributions by Guillaume Bertholon and Pierre Roux <Pierre.Roux@onera.fr>
Co-authored-by: Benjamin Grégoire <Benjamin.Gregoire@inria.fr>
Co-authored-by: Vincent Laporte <Vincent.Laporte@fondation-inria.fr>
|
|
|
|
Rename Univ.Level.{Qualid -> UGlobal}, remove Univ.Level.Id.
Remove the ability to split the argument of `Univ.Level.Level` into a
dirpath*int pair (except by going through string hacks like
detyping/pretyping(/funind) does).
Id.of_string_soft to turn unnamed universes into qualid is pushed up
to detyping. (TODO some followup PR clean up more)
This makes it pointless to have an opaque type for ints in
Univ.Level: it would only be used as argument to
Univ.Level.UGlobal.make, ie
~~~
open Univ.Level
let x = UGlobal.make dp (Id.make n)
(* vs *)
let x = UGlobal.make dp n
~~~
Remaining places which create levels from ints are various hacks (eg
the dummy in inductive.ml, the Type.n universes in ugraph
sort_universes) and univgen.
UnivGen does have an opaque type for ints used as univ ids since they
get manipulated by the stm.
NB: build breaks due to ocamldep issue if UGlobal is named Global instead.
|
|
Remote counters were trying to build universe levels (as opposed to
simple integers), but did not have access to the right dirpath at
construction time. We fix it by constructing the level only at use time,
and we introduce some abstractions for qualified and unqualified level
names.
|
|
|
|
|
|
|
|
|
|
Now that the checker shares code with the kernel using md5 cic.mli
doesn't work. We could md5 declarations.ml but this would miss changes
to constr (and cic.mli already missed changes to names, univs).
Instead we just print a bit of information (the last seen type
name/annotation) when validate fails. This should help debugging when
forgetting to update values.ml without the full verbosity of -debug.
As such there is no need to -debug in the makefile validate
target (NB: CI has an independent implementation of the validate rule
since the vos are installed there).
|
|
There is little point in having a list, as there is virtually no sharing
nor expansion of bound universe names. This representation is thus more
compact.
|
|
|
|
The kernel no longer has to read the configure flag, its value can now
be overriden by a coqtop/coqc argument, and more generally is easier to
set from a toplevel (such as the checker).
We also add a `-bytecode-compiler` flag.
Fixes #4607
|
|
We remove sections paths from kernel names. This is a cleanup as most of the times this information was unused. This implies a change in the Kernel API and small user visible changes with regards to tactic qualification. In particular, the removal of "global discharge" implies a large cleanup of code.
Additionally, the change implies that some machinery in `library` and `safe_typing` must now take an `~in_section` parameter, as to provide the information whether a section is open or not.
|
|
|
|
For some reason the code was implementing substitutions as pairs of maps,
but the invariant ensured actually no observable difference between fetching
a module ident from one or the other. The split seems to be mostly due to
historical reasons. We make this invariant static by representing substitutions
as a single map.
|
|
We move the global declaration of that argument to the environment, and reuse
the Global module to handle this flag.
Note that the checker was not using this flag before this patch, and still
doesn't use it. This should probably be fixed in a later patch.
|
|
The upper layers still need a mapping constant -> projection, which is
provided by Recordops.
|
|
We use a dedicated mutable constructor to perform a Landin knot.
|
|
|
|
This brings more compatibility with handling of mutual primitive records
in the kernel.
|
|
This is a first step towards the acceptance of mutual record types in the
kernel.
|
|
|
|
This eliminates 3 uses of Obj from TCB.
|
|
This was completely wrong, such a term could not even be type-checked by
the kernel as it was internally using a match construct over a negative
record. They were luckily only used in upper layers, namley printing
and extraction.
Recomputing the projection body might be costly in detyping, but this only
happens when the compatibility flag is turned on, which is not the default.
Such flag is probably bound to disappear anyways.
Extraction should be fixed though so as to define directly primitive
projections, similarly to what has been done in native compute.
|
|
This field was not used inside the kernel and not used in
performance-critical code where caching is essential, so we extrude
the code that computes it out of the kernel.
|
|
This field used to signal that a constant was the compatibility
eta-expansion of a primitive projections, but since a previous cleanup in
the kernel it had become useless.
|
|
Unluckily, this is completely wrong as we trust the inlined term to be
well-typed in some unavailable environment. To start with, the checker
should not even rely on substitutions as it does not trust functors,
but it does anyways. I have thus commented this code as a useful backdoor
for when Coq is used to implement the next blockchain Ponzi scheme.
We really need to sort this out though.
|
|
Instead of having the projection data in the constant data we have it
independently in the environment.
|
|
|
|
|
|
We defer the computation of the universe quantification to the upper layer,
outside of the kernel.
|
|
Since cumulativity of an inductive type is the universe constraints
which make a term convertible with its universe-renamed copy, the only
constraints we can get are between a universe and its copy.
As such we do not need to be able to represent arbitrary constraints
between universes and copied universes in a double-sized ucontext,
instead we can just keep around an array describing whether a bound
universe is covariant, invariant or irrelevant (CIC has no
contravariant conversion rule).
Printing is fairly obtuse and should be improved: when we print the
CumulativityInfo we add marks to the universes of the instance: = for
invariant, + for covariant and * for irrelevant. ie
Record Foo@{i j k} := { foo : Type@{i} -> Type@{j} }.
Print Foo.
gives
Cumulative Record Foo : Type@{max(i+1, j+1)} := Build_Foo
{ foo : Type@{i} -> Type@{j} }
(* =i +j *k |= *)
|
|
We also have to update the checker to deserialize this additional data,
but it is not using it in type-checking yet.
|
|
|
|
Also use constant_universes_entry instead of a bool flag to indicate
polymorphism in ParameterEntry.
There are a few places where we convert back to ContextSet because
check_univ_decl returns a UContext, this could be improved.
|
|
This will allow to merge back `Names` with `API.Names`
|
|
|
|
This should save a lot of useless reallocations and hashset crawling, which
end up costing a lot.
|
|
|
|
As explained in edf85b9, the original commit that merged the module_body
and module_type_body representations, this was delayed to a later time
assumedly due to OCaml lack of GADTs. Actually, the only thing that was
needed was polymorphic recursion, which has been around already for a
relatively long time (since 3.12).
|
|
The use of template polymorphism in constants was quite limited, as it
only applied to definitions that were exactly inductive types without any
parameter whatsoever. Furthermore, it seems that following the introduction
of polymorphic definitions, the code path enforced regular polymorphism as
soon as the type of a definition was given, which was in practice almost
always.
Removing this feature had no observable effect neither on the test-suite,
nor on any development that we monitor on Travis. I believe it is safe to
assume it was nowadays useless.
|
|
|
