Fix unstable comment just before or-patterns

[Julow]

Fixes #1164

[facebook-github-bot]

Hi Julow! Thank you for your pull request and welcome to our community. We require contributors to sign our Contributor License Agreement, and we don't seem to have you on file.In order for us to review and merge your code, please sign at https://code.facebook.com/cla. If you are contributing on behalf of someone else (eg your employer), the individual CLA may not be sufficient and your employer may need to sign the corporate CLA.

If you have received this in error or have any questions, please contact us at [email protected]. Thanks!

[Julow]

I didn't find a case where break 1000 0 was necessary.

[Julow]

pro must be noop when adj is not to avoid extra empty lines.

[facebook-github-bot]

Thank you for signing our Contributor License Agreement. We can now accept your code for this (and any) Facebook open source project. Thanks!

[gpetiot]

Looks good to me, thanks.
For completeness, here is the diff for the conventional profile (no diff with default options, and with janestreet profile):

diff --git a/infer/src/clang/cTrans.ml b/infer/src/clang/cTrans.ml
index 09fca70c1..1da1f0cd0 100644
--- a/infer/src/clang/cTrans.ml
+++ b/infer/src/clang/cTrans.ml
@@ -3374,8 +3374,7 @@ struct
       let is_by_ref =
         (* see http://en.cppreference.com/w/cpp/language/lambda *)
         match lci_capture_kind with
-        | `LCK_ByRef (* explicit with [&x] or implicit with [&] *)
-        | `LCK_This
+        | `LCK_ByRef (* explicit with [&x] or implicit with [&] *) | `LCK_This
         (* explicit with [this] or implicit with [&] *)
         | `LCK_VLAType
         (* capture a variable-length array by reference. we probably don't handle
diff --git a/infer/src/nullsafe/typeOrigin.ml b/infer/src/nullsafe/typeOrigin.ml
index 7099f0713..e2a9a5a45 100644
--- a/infer/src/nullsafe/typeOrigin.ml
+++ b/infer/src/nullsafe/typeOrigin.ml
@@ -56,8 +56,7 @@ let equal = [%compare.equal: t]
 let get_nullability = function
   | NullConst _ -> Nullability.Null
   | NonnullConst _ | This (* `this` can not be null according to Java rules *)
-  | New (* In Java `new` always create a non-null object  *)
-  | ArrayLengthResult
+  | New (* In Java `new` always create a non-null object  *) | ArrayLengthResult
   (* integer hence non-nullable *)
   | InferredNonnull _
   (* WARNING: we trade soundness for usability.
@@ -67,8 +66,7 @@ let get_nullability = function
      We currently don't have a nice way to detect initialization, neither automatical nor manual.
      Hence we make potentially dangerous choice in favor of pragmatism.
   *)
-  | ArrayAccess
-  | OptimisticFallback
+  | ArrayAccess | OptimisticFallback
   (* non-null is the most optimistic type *)
   | Undef
   (* This is a very special case, assigning non-null is a technical trick *) ->
diff --git a/lambda/lambda.ml b/lambda/lambda.ml
index db23401fe..8513e4474 100644
--- a/lambda/lambda.ml
+++ b/lambda/lambda.ml
@@ -440,8 +440,7 @@ let make_key e =
     | Lsend (m, e1, e2, es, _loc) ->
         Lsend (m, tr_rec env e1, tr_rec env e2, tr_recs env es, Location.none)
     | Lifused (id, e) -> Lifused (id, tr_rec env e)
-    | Lletrec _ | Lfunction _ | Lfor _
-    | Lwhile _
+    | Lletrec _ | Lfunction _ | Lfor _ | Lwhile _
     (* Beware: (PR#6412) the event argument to Levent
        may include cyclic structure of type Type.typexpr *)
     | Levent _ ->

which is a fix considering break-cases=fit is used for conventional.

[Julow]

The last two indeed look like a fix. I'm not 100% sure for the first two, I added a test for that.

[gpetiot]

I doubted as well at first because it kinda looks weird since the comment is long, no doubt for a shorter comment. But this is what is expected for break-cases=fit.

[Julow]

I see. I tested too and it's a fix.
Should I put that in the changelog ? If yes, how ?

[gpetiot]

If you don't find this too long I would modify your entry: Fix an unstable comment bug in or-patterns; and fix wrapping of or-patterns in presence of comments when break-cases=fit

[Julow]

That's a bit long for a single line, I tried something.