This PR makes it harder to create "fake" theorems about definitions that
are stubbed-out with `sorry` by ensuring that each `sorry` is not
definitionally equal to any other. For example, this now fails:
```lean
example : (sorry : Nat) = sorry := rfl -- fails
```
However, this still succeeds, since the `sorry` is a single
indeterminate `Nat`:
```lean
def f (n : Nat) : Nat := sorry
example : f 0 = f 1 := rfl -- succeeds
```
One can be more careful by putting parameters to the right of the colon:
```lean
def f : (n : Nat) → Nat := sorry
example : f 0 = f 1 := rfl -- fails
```
Most sources of synthetic sorries (recall: a sorry that originates from
the elaborator) are now unique, except for elaboration errors, since
making these unique tends to cause a confusing cascade of errors. In
general, however, such sorries are labeled. This enables "go to
definition" on `sorry` in the Infoview, which brings you to its origin.
The option `set_option pp.sorrySource true` causes the pretty printer to
show source position information on sorries.
**Details:**
* Adds `Lean.Meta.mkLabeledSorry`, which creates a sorry that is labeled
with its source position. For example, `(sorry : Nat)` might elaborate
to
```
sorryAx (Lean.Name → Nat) false
`lean.foo.12.8.12.13.8.13._sorry._@.lean.foo._hyg.153
```
It can either be made unique (like the above) or merely labeled. Labeled
sorries use an encoding that does not impact defeq:
```
sorryAx (Unit → Nat) false (Function.const Lean.Name ()
`lean.foo.14.7.13.7.13.69._sorry._@.lean.foo._hyg.174)
```
* Makes the `sorry` term, the `sorry` tactic, and every elaboration
failure create labeled sorries. Most are unique sorries, but some
elaboration errors are labeled sorries.
* Renames `OmissionInfo` to `DelabTermInfo` and adds configuration
options to control LSP interactions. One field is a source position to
use for "go to definition". This is used to implement "go to definition"
on labeled sorries.
* Makes hovering over a labeled `sorry` show something friendlier than
that full `sorryAx` expression. Instead, the first hover shows the
simplified ``sorry `«lean.foo:48:11»``. Hovering over that hover shows
the full `sorryAx`. Setting `set_option pp.sorrySource true` makes
`sorry` always start with printing with this source position
information.
* Removes `Lean.Meta.mkSyntheticSorry` in favor of `Lean.Meta.mkSorry`
and `Lean.Meta.mkLabeledSorry`.
* Changes `sorryAx` so that the `synthetic` argument is no longer
optional.
* Gives `addPPExplicitToExposeDiff` awareness of labeled sorries. It can
set `pp.sorrySource` when source positions differ.
* Modifies the delaborator framework so that delaborators can set Info
themselves without it being overwritten.
Incidentally closes#4972.
Inspired by [this Zulip
thread](https://leanprover.zulipchat.com/#narrow/channel/287929-mathlib4/topic/Is.20a.20.60definition_wanted.60.20keyword.20possible.3F/near/477260277).
@Kha This commit addresses an issue reported by Kevin. Holes and tactic
blocks represent a discontinuity in the elaboration process.
By introducing inaccessible variables (or "things" as Kevin calls
them), we create error message that are harder to understand (see
affected test), and goals where we didn't allow the user to select the
variable name and/or eagerly unfolded a definition.
BTW, I first considered using "reducible" setting when deciding
whether to insert implicit lambdas or not. This is a bad idea.
See `monotone.lean` test. The decision should not depend on
reducibility status, but whether there is "discontinuity" on the
elaboration process or not. As Kevin pointed out,
"introducing implicits work great if you finish the job".
@Kha we do that in Lean 3. It helps when the error is due to incorrect universe levels.
BTW, I had to update `tests/lean/server/content_diag.json` since the
error message is different, but a few other stuff changed too.
Could you please take a look whether the test is still correct?
@Kha I am using `_shadowed.<idx>` suffix for marking variables that
have been shadowed. It is a bit verbose, but at least it is easy to
understand understand error messages such as
```
shadow.lean:4:0: error: type mismatch
h
has type
x._shadowed.1 = x._shadowed.1
but it is expected to have type
x = x
```
It is better than the old cryptic version
```
shadow.lean:4:0: error: type mismatch
h
has type
x = x
but it is expected to have type
x = x
```
@Kha, we now support variable/constant shadowing in patterns.
A constant may occur in a pattern if it is a constructor or tagged with
the new [pattern] attribute. In the standard library, I have tagged
'add', 'zero', 'one', 'bit0', 'bit1' and 'rfl' with this new attribute.
BTW, arbitrary constants and variables may occur nested in type ascriptions and
inaccessible terms.
Here is an example:
meta_definition tactic_result_to_string {A : Type} : tactic_result A → string
| (success a s) := to_string a
| (exception ⌞A⌟ e s) := "Exception: " ++ to_string (e ())
I had to use the inaccessible ⌞A⌟ in the example above, otherwise, we would be shadowing the parameter
{A : Type}, and we would get a type error.
The new validation is performed at to_pattern_fn (parser.cpp).
