ASSERT failures in testsuite

Joachim writes: our secondary CI infrastructure is in bad shape since a week.

The first push that broke it was the one ending with commit fc4ef667 by SPJ:

• fc4ef667 - Comments only (vor 7 Tagen)<Simon Peyton Jones>
• b612da66 - Fix impredicativity (again) (vor 7 Tagen)<Simon Peyton Jones>
• 3012c431 - Add Outputable Report in TcErrors (vor 7 Tagen)<Simon Peyton Jones>
• 66a8c194 - Fix a bug in occurs checking (vor 7 Tagen)<Simon Peyton Jones>
• 2fbfbca2 - Fix desugaring of pattern bindings (again) (vor 7 Tagen)<Simon Peyton Jones>
• 0b533a25 - A bit of tracing about flattening (vor 7 Tagen)<Simon Peyton Jones>

--- /dev/null	2015-01-28 16:31:58.000000000 +0000
+++ /tmp/ghctest-gSX4vv/test   spaces/./boxy/Base1.run/Base1.comp.stderr.normalised	2016-09-30 12:44:57.182659875 +0000
@@ -0,0 +1,18 @@
+
+Base1.hs:20:13:
+     Couldn't match type ‘a0 -> a0’ with ‘forall a. a -> a’
+      Expected type: MEither Sid b
+        Actual type: MEither (a0 -> a0) b
+     In the expression: MLeft fid
+      In an equation for ‘test1’: test1 fid = MLeft fid
+
+Base1.hs:25:39:
+     Couldn't match type ‘a1 -> a1’ with ‘forall a. a -> a’
+      Expected type: Maybe (Sid, Sid)
+        Actual type: Maybe (a1 -> a1, a2 -> a2)
+     In the expression: Just (x, y)
+      In a case alternative: MRight y -> Just (x, y)
+      In the expression:
+        case m of {
+          MRight y -> Just (x, y)
+          _ -> Nothing }
*** unexpected failure for Base1(normal)

Compile failed (exit code 1) errors were:
ghc-stage2: panic! (the 'impossible' happened)
  (GHC version 8.1.20160930 for x86_64-unknown-linux):
	ASSERT failed!
  m_aAI
  Call stack:
      CallStack (from HasCallStack):
        prettyCurrentCallStack, called at compiler/utils/Outputable.hs:1076:58 in ghc:Outputable
        callStackDoc, called at compiler/utils/Outputable.hs:1125:22 in ghc:Outputable
        assertPprPanic, called at compiler/typecheck/TcType.hs:979:47 in ghc:TcType
  Call stack:
      CallStack (from HasCallStack):
        prettyCurrentCallStack, called at compiler/utils/Outputable.hs:1076:58 in ghc:Outputable
        callStackDoc, called at compiler/utils/Outputable.hs:1080:37 in ghc:Outputable
        pprPanic, called at compiler/utils/Outputable.hs:1123:5 in ghc:Outputable
        assertPprPanic, called at compiler/typecheck/TcType.hs:979:47 in ghc:TcType

Please report this as a GHC bug:  https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fwww.haskell.org%2Fghc%2Freportabug&data=01%7C01%7Csimonpj%40microsoft.com%7Cb15d11554ecf4f68173d08d3ee64ef42%7C72f988bf86f141af91ab2d7cd011db47%7C1&sdata=eG%2F4%2FH6lhGUFjL1my3GUE5Ho1hP3C4j0tkEihKvddyQ%3D&reserved=0


*** unexpected failure for RaeBlogPost(normal)

Compile failed (exit code 1) errors were:
ghc-stage2: panic! (the 'impossible' happened)
  (GHC version 8.1.20160930 for x86_64-unknown-linux):
	ASSERT failed!
  m_awK
  Call stack:
      CallStack (from HasCallStack):
        prettyCurrentCallStack, called at compiler/utils/Outputable.hs:1076:58 in ghc:Outputable
        callStackDoc, called at compiler/utils/Outputable.hs:1125:22 in ghc:Outputable
        assertPprPanic, called at compiler/typecheck/TcType.hs:979:47 in ghc:TcType
  Call stack:
      CallStack (from HasCallStack):
        prettyCurrentCallStack, called at compiler/utils/Outputable.hs:1076:58 in ghc:Outputable
        callStackDoc, called at compiler/utils/Outputable.hs:1080:37 in ghc:Outputable
        pprPanic, called at compiler/utils/Outputable.hs:1123:5 in ghc:Outputable
        assertPprPanic, called at compiler/typecheck/TcType.hs:979:47 in ghc:TcType

Please report this as a GHC bug:  https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fwww.haskell.org%2Fghc%2Freportabug&data=01%7C01%7Csimonpj%40microsoft.com%7Cb15d11554ecf4f68173d08d3ee64ef42%7C72f988bf86f141af91ab2d7cd011db47%7C1&sdata=eG%2F4%2FH6lhGUFjL1my3GUE5Ho1hP3C4j0tkEihKvddyQ%3D&reserved=0


*** unexpected failure for TypeLevelVec(normal)

A subsequent commit by SPJ fixed the first one. It now compield without -DDEBUG again, but the other two failures (observable only with -DDEBUG) are still there.

Since then, Travis has been reporting failures for the master branch. I only noticed now as I pushed something to master, and I got an email. Did you not get notifications about the breakage?

If you did, was it unclear how to get to the log file?

In any case: Simon, could you have a look and see if the ASSERT is pointing out a real bug introduced with your commits, or whether the ASSERT is wrong, so that we can build master with -DDEBUG again?

Trac metadata

Trac field	Value
Version	8.0.1
Type	Bug
TypeOfFailure	OtherFailure
Priority	normal
Resolution	Unresolved
Component	Compiler
Test case
Differential revisions
BlockedBy
Related
Blocking
CC
Operating system
Architecture