| Age | Commit message (Collapse) | Author |
|---|
|
This ensures that the API is self-contained and is, well, an API.
Before this patch, the contents of `API.mli` bore little relation with
what was used by the plugins [example: `Metasyntax` in tacentries.ml].
Many missing types had to be added.
A sanity check of the `API.mli` file can be done with:
`ocamlfind ocamlc -rectypes -package camlp5 -I lib API/API.mli`
|
|
4844 and 4824)
The commit 8e257d4 (which consider the dummy type Tdummy to be polymorphic and hence
immune of the need for unsafeCoerce) is quite wrong, even if in pratice it worked ok
most of the time. The confusion comes from the fact that the dummy value (__ aka MLdummy
internally) is implemented in Haskell as Prelude.error, hence it has indeed a polymorphic
unrestricted type. But the dummy type Tdummy used when extracting types must be monomorphic
(otherwise type parameters would have to be propagated out of any type definition involving
Tdummy). We implement Tdummy by Haskell's (), which for instance isn't convertible to Any,
as shown by the examples in bug reports 4844 and 4824.
This fix will bring back some more unsafeCoerce in Haskell extraction, including possibly
a few spurious ones. And these extra unsafeCoerce might also hinder further code optimisations.
We tried to mitigate that by directly removing [MLmagic] constructors in front of [MLdummy _].
NB: even if the original bug report mentions universe polymorphism, this issue is
almost unrelated to it. It just happens that when universe polymorphism is off,
an inductive instance is fully placed in Prop (cf. template polymorphism) in the example,
avoiding triggering the issue.
Warning: the test-suite file is there for archiving the repro case, but currently it doesn't
test much (we should run ghc on the extracted code).
|
|
Avoid Anomaly (or Assert False) when a module signature contains
an applied functor followed by a "with Definition" or "with Module"
Also fix the dependency computation in presence of a "with Definition"
Concerning 4720, the code extracted out of this bug report was suboptimal
in Coq 8.4 (it was compilable, but could have probably been tweaked into a
real issue producing faulty code).
But the example of 4720 (and some variants of it) was broken since 8.5,
for the same reasons as 5177 and 5240. And the good news is that after
these repairs, the example of bug 4720 is now extracted to correct code
(the "with" is preserved).
|
|
|
|
|
|
Extraction TestCompile foo
is equivalent to:
Extraction "/tmp/testextraction1234.ml" foo
ocamlfind ocamlc -I /tmp -c /tmp/testextraction1234.mli /tmp/testextraction1234.ml
This command isn't meant for the end user, but rather as an helper
for test-suite scripts. It only works with extraction to OCaml,
and the generated code should be standalone.
|
|
|
|
The user now has to manually load them, respectively via:
Require Extraction
Require Import FunInd
The "Import" in the case of FunInd is to ensure that the
tactics functional induction and functional inversion are indeed
in scope.
Note that the Recdef.v file is still there as well (it contains
complements used when doing Function with measures), and it also
triggers a load of FunInd.v.
This change is correctly documented in the refman, and the test-suite
has been adapted.
|
|
|
|
|
|
|
|
|
|
As per https://github.com/coq/coq/pull/716#issuecomment-305140839
Partially using
```bash
git grep --name-only 'anomaly\s*\(~label:"[^"]*"\s*\)\?\(Pp.\)\?(\(\(Pp.\)\?str\)\?\s*".*[^\.!]")' | xargs sed s'/\(anomaly\s*\(~label:"[^"]*"\s*\)\?\(Pp.\)\?(\(\(Pp.\)\?str\)\?\s*".*\s*[^\.! ]\)\s*")/\1.")/g' -i
```
and
```bash
git grep --name-only ' !"' | xargs sed s'/ !"/!"/g' -i
```
The rest were manually edited by looking at the results of
```bash
git grep anomaly | grep '\.ml' | grep -v 'anomaly\s*\(~label:"[^"]*"\s*\)\?\(Pp\.\)\?(\(\(Pp.\)\?str\)\?\s*".*\(\.\|!\)")' | grep 'anomaly\($\|[^_]\)' | less
```
|
|
|
|
Deprecations which can't be fixed in 4.02.3 are locally wrapped with
[@@@ocaml.warning "-3"]. The only ones encountered are
- capitalize to capitalize_ascii and variants. Changing to ascii would
break coqdoc -latin1 and maybe other things though.
- external "noalloc" to external [@@noalloc]
|
|
This is the continuation of #244, we now deprecate `CErrors.error`,
the single entry point in Coq is `user_err`.
The rationale is to allow for easier grepping, and to ease a future
cleanup of error messages. In particular, we would like to
systematically classify all error messages raised by Coq and be sure
they are properly documented.
We restore the two functions removed in #244 to improve compatibility,
but mark them deprecated.
|
|
|
|
|
|
Inspired by https://coq.inria.fr/bugs/show_bug.cgi?id=5229 , which
this PR solves, I propose to remove support for non-synchronous
options.
It seems the few uses of `optsync = false` we legacy and shouldn't
have any impact.
Moreover, non synchronous options may create particularly tricky
situations as for instance, they won't be propagated to workers.
|
|
|
|
This completes the Loc.ghost removal, the idea is to gear the API
towards optional, but uniform, location handling.
We don't print <unknown> anymore in the case there is no location.
This is what the test suite expects.
The old printing logic for located items was a bit inconsistent as
it sometimes printed <unknown> and other times it printed nothing as
the caller checked for `is_ghost` upstream.
|
|
|
|
This reverts commit 6d2802075606dcddb02dd13cbaf38ff76f8bf242, which is
an 8.6 only commit.
|
|
|
|
This way, after we merge PR#220, scripts can be fixed in a way that is
compatible with the 8.6 and trunk branches.
|
|
Previous implementations of `Pp` flushed on newline, however,
depending on the formatter this may not be always the case.
We now alwayas flush the formatters before closing the file as this is
the intended behavior.
|
|
Previously to this patch, Coq featured to distinct logging paths: the
console legacy one, based on `Pp.std_ppcmds` and Ocaml's `Format`
module, and the `Feedback` one, intended to encapsulate message inside a
more general, GUI-based feedback protocol.
This patch removes the legacy logging path and makes feedback
canonical. Thus, the core of Coq has no dependency on console code
anymore.
Additionally, this patch resolves the duplication of "document" formats
present in the same situation. The original console-based printing code
relied on an opaque datatype `std_ppcmds`, (mostly a reification of
`Format`'s format strings) that could be then rendered to the console.
However, the feedback path couldn't reuse this type due to its opaque
nature. The first versions just embedded rending of `std_ppcmds` to a
string, however in 8.5 a new "rich printing" type, `Richpp.richpp` was
introduced.
The idea for this type was to be serializable, however it brought
several problems: it didn't have proper document manipulation
operations, its format was overly verbose and didn't preserve the full
layout, and it still relied on `Format` for generation, making
client-side rendering difficult.
We thus follow the plan outlined in CEP#9, that is to say, we take a
public and refactored version of `std_ppcmds` as the canonical "document
type", and move feedback to be over there. The toplevel now is
implemented as a feedback listener and has ownership of the console.
`richpp` is now IDE-specific, and only used for legacy rendering. It
could go away in future versions. `std_ppcmds` carries strictly more
information and is friendlier to client-side rendering and display
control.
Thus, the new panorama is:
- `Feedback` has become a very module for event dispatching.
- `Pp` contains a target-independent box-based document format.
It also contains the `Format`-based renderer.
- All console access lives in `toplevel`, with console handlers private
to coqtop.
_NOTE_: After this patch, many printing parameters such as printing
width or depth should be set client-side. This works better IMO,
clients don't need to notify Coq about resizing anywmore. Indeed, for
box-based capable backends such as HTML or LaTeX, the UI can directly
render and let the engine perform the word breaking work.
_NOTE_: Many messages could benefit from new features of the output
format, however we have chosen not to alter them to preserve output.
A Future commits will move console tag handling in `Pp_style` to
`toplevel/`, where it logically belongs.
The only change with regards to printing is that the "Error:" header was
added to console output in several different positions, we have removed
some of this duplication, now error messages should be a bit more
consistent.
|
|
We also remove flushing operations `msg_with`, now the flushing
responsibility belong to the owner of the formatter.
|
|
Mostly unused, we ought to limit spacing in the boxes themselves.
|
|
No functional change.
|
|
This is cumbersome, because now code may fail at link time if it's not
referring to the correct module name. Therefore, one has to add corresponding
open statements a the top of every file depending on a Ltac module. This
includes seemingly unrelated files that use EXTEND statements.
|
|
|
|
This removes quite a few unsafe casts. Unluckily, I had to reintroduce
the old non-module based names for these data structures, because I could
not reproduce easily the same hierarchy in EConstr.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
This reverts commit 69c4e7cfa0271f024b2178082e4be2e3ca3be263.
String.capitalize_ascii are only available for ocaml >= 4.03, sorry...
|
|
|
|
|
|
- A few tweaks of string are now done via the Bytes module
- lots of String.capitalize_ascii and co
|
|
No more pp_alias_spec et pp_alias_decl.
Instead, we use "include" and "module type of".
The extracted code might hence need OCaml 3.12 (quite rarely)
|
|
Not_found)
This clarifies the execution flow
|
|
A double call to pp_module_type inside Ocaml.pp_specif was
causing an complexity blowup when pretty-printing heavily modular
extracted code.
I wasn't able to figure out why this double call is there. It could
be the leftover of some intermediate work in 2007 before commit 350398eae
(which introduced global printing phases Pre/Impl/Intf).
Anyway I'm reasonably sure that today these two pp_module_type calls produce
the exact same pretty-printed signature (even if there's a large bunch of
imperative states around). Moreover, this duplicated signature is actually
slightly wrong: when we alias a module M with a unambiguous name like Coq__123,
the type of Coq__123 should not be an exact copy of the type of M, but rather
a "strengthened" version of it (with equality between inductive types).
So the best solution is now to use this funny feature of OCaml introduced in 3.12 :
module Coq__123 : module type of struct include M end
This "module type of struct include" is slightly awkward, but short, correct,
and trivial to produce :-). And I doubt anybody will object to the (rare) use
of some 3.12 features in extracted code of 2017...
|
