On 10/11/2009 00:13, Manuel M T Chakravarty wrote:
Simon Marlow:On Leopard, the lists is still longer:Unexpected failures: 2302(normal) 2816(ghci) T1969(normal) T3294(normal) conc070(ghci) ffi005(normal) ghci028(ghci) print021(ghci)How are the others failing, can you post the output?I attach the error messages of all failures of the HEAD from today on *Leopard*. (IIRC think the unicode failures don't happen on Snow Leopard.) Manuel -=-=====> 2302(normal) cd ./parser/unicode && '/Users/chak/Code/ghc-test/bindisttest/install dir/bin/ghc' -fforce-recomp -dcore-lint -dcmm-lint -no-user-package-conf -dno-debug-output -c 2302.hs >2302.comp.stderr 2>&1 Actual stderr output differs from expected: --- ./parser/unicode/2302.stderr.normalised 2009-11-10 10:49:46.000000000 +1100 +++ ./parser/unicode/2302.comp.stderr.normalised 2009-11-10 10:49:46.000000000 +1100 @@ -1,2 +1,2 @@ -2302.hs:1:4: Not in scope: data constructor `À' +2302.hs:1:4: Not in scope: data constructor `<stderr>: hPutChar: invalid argument (Illegal byte sequence) *** unexpected failure for 2302(normal)
This is caused by not having a UTF-8 locale. The testsuite driver makes some attempt to set a UTF-8 locale, but perhaps it isn't working on Leopard. If you have a chance, could you investigate? (the top of testsuite/driver/runtests.py).
=====> 2816(ghci) cd ./ghci/scripts && HC='/Users/chak/Code/ghc-test/bindisttest/install dir/bin/ghc' HC_OPTS='-dcore-lint -dcmm-lint -no-user-package-conf -dno-debug-output ' '/Users/chak/Code/ghc-test/bindisttest/install dir/bin/ghc' --interactive -v0 -ignore-dot-ghci -dcore-lint -dcmm-lint -no-user-package-conf -dno-debug-output <2816.script >2816.run.stdout 2>2816.run.stderr Actual stderr output differs from expected: --- ./ghci/scripts/2816.stderr.normalised 2009-11-10 10:46:27.000000000 +1100 +++ ./ghci/scripts/2816.run.stderr.normalised 2009-11-10 10:46:27.000000000 +1100 @@ -1,2 +0,0 @@ - -<interactive>:1:0: Not in scope: `α' *** unexpected failure for 2816(ghci)
locale again.
=====> T1969(normal) cd ./space_leaks && '/Users/chak/Code/ghc-test/bindisttest/install dir/bin/ghc' -fforce-recomp -dcore-lint -dcmm-lint -no-user-package-conf -dno-debug-output -c T1969.hs +RTS -V0 -tT1969.comp.stats --machine-readable -RTS >T1969.comp.stderr 2>&1 peak_megabytes_allocated 20 is more than maximum allowed 19 max_bytes_used 7015308 is more than maximum allowed 7000000 bytes allocated 257219432 is more than maximum allowed 250000000 *** unexpected failure for T1969(normal)=====> T3294(normal) cd ./space_leaks && '/Users/chak/Code/ghc-test/bindisttest/install dir/bin/ghc' -fforce-recomp -dcore-lint -dcmm-lint -no-user-package-conf -dno-debug-output -c T3294.hs +RTS -V0 -tT3294.comp.stats --machine-readable -RTS >T3294.comp.stderr 2>&1 bytes allocated 837155440 is more than maximum allowed 830000000 *** unexpected failure for T3294(normal)
These two are indicating a recent performance regression in GHC that we still need to look into.
conc070 works now!=====> ffi005(normal) cd ./ffi/should_run && '/Users/chak/Code/ghc-test/bindisttest/install dir/bin/ghc' -fforce-recomp -dcore-lint -dcmm-lint -no-user-package-conf -dno-debug-output -o ffi005 ffi005.hs -fglasgow-exts >ffi005.comp.stderr 2>&1 cd ./ffi/should_run && ./ffi005 </dev/null >ffi005.run.stdout 2>ffi005.run.stderr Actual stdout output differs from expected: --- ./ffi/should_run/ffi005.stdout.normalised 2009-11-10 10:45:37.000000000 +1100 +++ ./ffi/should_run/ffi005.run.stdout.normalised 2009-11-10 10:45:37.000000000 +1100 @@ -1,13 +1,13 @@ Testing sin==mysin (should return lots of Trues) -[True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True] +[True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,False,True,True,True,True,False,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,False,True,False,True,True,True,True,True,True,True,True,True,False,True,True,True,True,True,True,True,True,True] Testing puts (and withString) Testing peekArray0 Test successful Testing sin==dynamic_sin (should return lots of Trues) -[True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True] +[True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,False,True,True,True,True,False,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,False,True,False,True,True,True,True,True,True,True,True,True,False,True,True,True,True,True,True,True,True,True] Testing sin==IO wrapped_sin (should return lots of Trues) [True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True] *** unexpected failure for ffi005(normal)
Perhaps some floating-point weirdness. The test is testSin f g = [ (f x == g x) | x <- [0,0.01 .. 1] ] and it looks like it is returning a list of different length on Leopard.However, it seems we also have some different failures in this test on x86/Linux when running the full testsuite (optasm, threaded2, dyn). So perhaps there's an x86 codegen issue, I'll look into it.
=====> ghci028(ghci) cd ./ghci/scripts && HC='/Users/chak/Code/ghc-test/bindisttest/install dir/bin/ghc' HC_OPTS='-dcore-lint -dcmm-lint -no-user-package-conf -dno-debug-output ' '/Users/chak/Code/ghc-test/bindisttest/install dir/bin/ghc' --interactive -v0 -ignore-dot-ghci -dcore-lint -dcmm-lint -no-user-package-conf -dno-debug-output <ghci028.script >ghci028.run.stdout 2>ghci028.run.stderr Actual stdout output differs from expected: --- ./ghci/scripts/ghci028.stdout.normalised 2009-11-10 10:46:25.000000000 +1100 +++ ./ghci/scripts/ghci028.run.stdout.normalised 2009-11-10 10:46:25.000000000 +1100 @@ -1 +0,0 @@ -"HSIF" *** unexpected failure for ghci028(ghci)
locale again.
=====> print021(ghci) cd ./ghci.debugger/scripts && HC='/Users/chak/Code/ghc-test/bindisttest/install dir/bin/ghc' HC_OPTS='-dcore-lint -dcmm-lint -no-user-package-conf -dno-debug-output ' '/Users/chak/Code/ghc-test/bindisttest/install dir/bin/ghc' --interactive -v0 -ignore-dot-ghci -dcore-lint -dcmm-lint -no-user-package-conf -dno-debug-output -ignore-dot-ghci <print021.script >print021.run.stdout 2>print021.run.stderr Actual stdout output differs from expected: --- ./ghci.debugger/scripts/print021.stdout.normalised 2009-11-10 10:46:37.000000000 +1100 +++ ./ghci.debugger/scripts/print021.run.stdout.normalised 2009-11-10 10:46:37.000000000 +1100 @@ -1,2 +1,2 @@ test = (_t1::[T]) -test = [À,Α,Ⴀ,Ϣ,А,Ա] +test = [*** Exception: <stdout>: hPutChar: invalid argument (Illegal byte sequence) *** unexpected failure for print021(ghci)
and again.
Cheers,
Simon
_______________________________________________
Cvs-ghc mailing list
Cvs-ghc@haskell.org
http://www.haskell.org/mailman/listinfo/cvs-ghc
