Unexpected failures in testsuite due to .ghci

[daniel.is.fischer]

Running the testsuite, I get e.g.

=====> 3171(normal) 375 of 2596 [0, 1, 0]
cd ./ghci/should_run && $MAKE -s --no-print-directory 3171    </dev/null >3171.run.stdout 2>3171.run.stderr
=====> bug1465(normal) 376 of 2596 [0, 1, 0]
cd ./typecheck/bug1465 && $MAKE -s --no-print-directory bug1465    </dev/null >bug1465.run.stdout 2>bug1465.run.stderr
Actual stdout output differs from expected:
--- ./ghci/should_run/3171.stdout.normalised	2010-10-07 13:20:16.000000000 +0200
+++ ./ghci/should_run/3171.run.stdout.normalised	2010-10-07 13:20:16.000000000 +0200
@@ -1 +1,3 @@
+ghci: warnings (except type defaulting and unused do-binds) on; MR off.
+(0.05 secs, 2799172 bytes)
 Interrupted.
*** unexpected failure for 3171(normal)

since it reads ~/.ghci.

I think -ignore-dot-ghci should be passed in the testsuite to avoid such gratuitous unexpected failures.

[daniel.is.fischer]

Hmm, actually, in testsuite/driver/testlib.py, it seems to be passed in

def ghci_script( name, way, script ):
...
   cmd = "HC='" + config.compiler + "' " + \
          "HC_OPTS='" + join(flags,' ') + "' " + \
          "'" + config.compiler + "'" + \
          ' --interactive -v0 -ignore-dot-ghci ' + \
          join(flags,' ')

Nevertheless, apparently .ghci is not ignored.

[igloo]

Thanks for the report. Fixed by

Wed Oct 13 18:52:15 BST 2010  Ian Lynagh <igloo@earth.li>
  * Add -ignore-dot-ghci to a ghci test

in hpc and

Wed Oct 13 18:53:48 BST 2010  Ian Lynagh <igloo@earth.li>
  * Add -ignore-dot-ghci to a ghci test

in testsuite.