[arch-commits] Commit in ghc/trunk (PKGBUILD)

[Thomas Dziedzic]

    Date: Monday, June 11, 2012 @ 17:22:56
  Author: tdziedzic
Revision: 161493

note all test failures and comments explaining tests

Modified:
  ghc/trunk/PKGBUILD

----------+
 PKGBUILD |   17 ++++++++++++++---
 1 file changed, 14 insertions(+), 3 deletions(-)

Modified: PKGBUILD
===================================================================
--- PKGBUILD	2012-06-11 21:12:03 UTC (rev 161492)
+++ PKGBUILD	2012-06-11 21:22:56 UTC (rev 161493)
@@ -102,10 +102,21 @@
   sed -e 's/PYTHON = python/&2/' -i testsuite/mk/boilerplate.mk
   sed -e 's_#!/usr/bin/env python_&2_' -i testsuite/timeout/calibrate testsuite/timeout/timeout.py
 
-  # note: there is one test failure on i686 only: http://hackage.haskell.org/trac/ghc/ticket/5856
+  # upstream known failures on both i686 and x86_64:
+  #    ghci.debugger/scripts     break011 [bad stderr] (ghci)
+  #    perf/compiler             parsing001 [stat not good enough] (normal)
+  #    simplCore/should_compile  spec-inline [stderr mismatch] (optasm)
+  # upstream known failures on i686 only:
+  #    numrun012 test fails on i686 works on x86_64 - http://hackage.haskell.org/trac/ghc/ticket/5856
+  make test
+
+  # enabling threads to speed up
+  # causes more failures than make test, need to investigate
   #make THREADS=9 test
-  make test
-  # might be some failures
+
+  # will most likely cause lots of failures
+  # enable this when upstream has 0 test failures from a simple 'make test'
+  # zero unexpected failures on all tier 1 platforms - http://hackage.haskell.org/trac/ghc/ticket/5757
   #make fulltest
 }