okay I only realized one thing: this test of running unboxed tuples of
course won't currently work with the `ghci` mode. Does the test need to
be specially notated in any way because of that?
Also my other test, parser/should_compile/read063, (and probably others)
seem to be getting a failure for optimization mode just because of a
NOTE in stderr? :
=====> read063(normal)
cd . &&
'/Users/me/modified/ghc-unboxed-prefix/compiler/stage1/ghc-inplace'
-no-recomp -dcore-lint -dcmm-lint -Di386_unknown_linux -c read063.hs
>read063.comp.stderr 2>&1
=====> read063(optc)
cd . &&
'/Users/me/modified/ghc-unboxed-prefix/compiler/stage1/ghc-inplace'
-no-recomp -dcore-lint -dcmm-lint -Di386_unknown_linux -c read063.hs -O
-fvia-C >read063.comp.stderr 2>&1
Actual stderr output differs from expected:
--- /dev/null 2007-12-30 14:48:00.000000000 -0500
+++ ./read063.comp.stderr.normalised 2008-01-04 08:55:51.000000000 -0500
@@ -0,0 +1 @@
+NOTE: Simplifier still going after 4 iterations; bailing out. Size = 26
*** unexpected failure for read063(optc)
Fri Jan 4 08:48:43 EST 2008 Isaac Dupree <[EMAIL PROTECTED]>
* add test for curried unboxed-tuple constructors
New patches:
[add test for curried unboxed-tuple constructors
Isaac Dupree <[EMAIL PROTECTED]>**20080104134843] {
hunk ./tests/ghc-regress/parser/should_run/all.T 4
+test('read004', normal, compile_and_run, [''])
addfile ./tests/ghc-regress/parser/should_run/read004.hs
hunk ./tests/ghc-regress/parser/should_run/read004.hs 1
+{-# LANGUAGE UnboxedTuples, MagicHash #-}
+
+-- should_run to make sure linking succeeds
+-- (curried unboxed tuples with both boxed
+-- and unboxed components).
+-- See Trac #1509; also Note [Primop wrappers] in Id.lhs
+
+import GHC.Exts
+
+main = do
+ case curried 9.0## 't'# of
+ (# i#, u@(), d1#, c1#, f#, d2#, c2# #)
+ -> print ( I# i#, u, D# d1#, C# c1#, F# f#, D# d2#, C# c2# )
+ print $ map_ ((#,#) True) ['a','b','c']
+
+-- try NOINLINE to make sure the currying isn't eliminated
+-- too soon, but also test the other one without NOINLINE
+-- for variety of testing
+{-# NOINLINE curried #-}
+curried :: Double# -> Char# ->
+ (# Int#, (), Double#, Char#, Float#, Double#, Char# #)
+curried = (#,,,,,,#) 3# () 4.0## 'f'# 5.0#
+
+map_ :: (a -> (# b, c #)) -> [a] -> [(b,c)]
+map_ f [] = []
+map_ f (a:as) = case f a of
+ (# b, c #) -> (b, c) : map_ f as
+
addfile ./tests/ghc-regress/parser/should_run/read004.stdout
hunk ./tests/ghc-regress/parser/should_run/read004.stdout 1
+(3,(),4.0,'f',5.0,9.0,'t')
+[(True,'a'),(True,'b'),(True,'c')]
}
Context:
[add test for prefix unboxed tuples
Isaac Dupree <[EMAIL PROTECTED]>**20080102132824]
[Accept output for tcfail172, too
Tim Chevalier <[EMAIL PROTECTED]>**20071225190352]
[Accept output for gadt13, gadt7
Tim Chevalier <[EMAIL PROTECTED]>**20071225185124]
[Enable an old test for Trac 323
[EMAIL PROTECTED]
[Test for Trac 1988
[EMAIL PROTECTED]
[Test for Trac 1495
[EMAIL PROTECTED]
[Add test for Trac 1981
[EMAIL PROTECTED]
[Replace mentions of way 'opt' by the new 'optc' (often adding optasm too)
[EMAIL PROTECTED]
Its really a testsuite-framework bug that these mentions of 'opt' have
lingered so long. They should have been rejected as a non-existent way.
]
[Tests for generalised list comprehensions
[EMAIL PROTECTED]
[Give a ticket # to expect_broken
Pepe Iborra <[EMAIL PROTECTED]>**20071220093444]
[Unsoundness in the RTTI scheme
Pepe Iborra <[EMAIL PROTECTED]>**20071219191932]
[add test for #1980
Simon Marlow <[EMAIL PROTECTED]>**20071218092731]
[Fix outofmem for other 32-bit Linux platforms
Ian Lynagh <[EMAIL PROTECTED]>**20071217231845]
[6.8 updates
Ian Lynagh <[EMAIL PROTECTED]>**20071215205108]
[Simplify the testsuite driver
Ian Lynagh <[EMAIL PROTECTED]>**20071215195350
Now instead of saying, e.g.
namebase_if_compiler_lt('ghc','6.9', 'foo-6.8')
you say
if_compiler_lt('ghc','6.9', namebase('foo-6.8'))
]
[add 6.8 output
Simon Marlow <[EMAIL PROTECTED]>**20071212142523]
[Update tests for Trac #1972
[EMAIL PROTECTED]
[More 6.8 fixes
Manuel M T Chakravarty <[EMAIL PROTECTED]>**20071213033710]
[Make GADT tests work with 6.8 again
Manuel M T Chakravarty <[EMAIL PROTECTED]>**20071213031332]
[Test for #1723
Manuel M T Chakravarty <[EMAIL PROTECTED]>**20071211071414]
[Test for #1722
Manuel M T Chakravarty <[EMAIL PROTECTED]>**20071211062431]
[Test from #1815
Manuel M T Chakravarty <[EMAIL PROTECTED]>**20071211055023]
[accept output
Simon Marlow <[EMAIL PROTECTED]>**20071212141945]
[add 6.8 output
Simon Marlow <[EMAIL PROTECTED]>**20071212141109]
[countReaders001 doesn't work on Windows
Simon Marlow <[EMAIL PROTECTED]>**20071212134000]
[fix this test on windows (avoid non-portable use of :!)
Simon Marlow <[EMAIL PROTECTED]>**20071212133745]
[Adapt to changes in :print (contents of references)
Pepe Iborra <[EMAIL PROTECTED]>**20071209182141]
[Remaining regression of change from GADT refinement to equalities
Manuel M T Chakravarty <[EMAIL PROTECTED]>**20071208111646
* gadt/lazypatok fails
* tcfail167's error message gets worse; ie, no more
"Inaccessible case alternative: Can't match types `Char' and `Float'"
Both of these are minor regressions that will be addressed in due time.
]
[gadt: slighly changed error msg with equalities
Manuel M T Chakravarty <[EMAIL PROTECTED]>**20071208110824]
[type families: Temporary fail
Manuel M T Chakravarty <[EMAIL PROTECTED]>**20071208110317
- GADT3 fails with equalities instead of GADT refinement due to a known bug
]
[Fixed by removing GADT refinements in favour of equalities
Manuel M T Chakravarty <[EMAIL PROTECTED]>**20071208110253]
[Fixed by correcting normalisation of dicts
Manuel M T Chakravarty <[EMAIL PROTECTED]>**20071207071954]
[Test for Trac #1370
[EMAIL PROTECTED]
[test now passes
Simon Marlow <[EMAIL PROTECTED]>**20071206092511]
[add test for #1959
Simon Marlow <[EMAIL PROTECTED]>**20071205152025]
[add test for #1048
Simon Marlow <[EMAIL PROTECTED]>**20071205102619]
[Added test for :print of a reference
Pepe Iborra <[EMAIL PROTECTED]>**20071205114017]
[GADT9 passes now that we have fixed #1919
[EMAIL PROTECTED]
[Skip arrowrun004 in the 6.8 branch
Ian Lynagh <[EMAIL PROTECTED]>**20071201214110]
[Skip arrowrun002 for the 6.8 branch
Ian Lynagh <[EMAIL PROTECTED]>**20071201213910]
[countReaders001 now passes in the 6.8 branch
Ian Lynagh <[EMAIL PROTECTED]>**20071201174337]
[Add 6.8 branch output for break021
Ian Lynagh <[EMAIL PROTECTED]>**20071201174241]
[Add 6.8 branch output for break003
Ian Lynagh <[EMAIL PROTECTED]>**20071201172005]
[6.8 branch no longer needs special output for ghci026
Ian Lynagh <[EMAIL PROTECTED]>**20071201171445]
[Update ghci024 output for the 6.8 branch
Ian Lynagh <[EMAIL PROTECTED]>**20071201171241]
[Add tcfail187 output for the 6.8 branch
Ian Lynagh <[EMAIL PROTECTED]>**20071201170905]
[Add 6.8-branch output for tcfail178
Ian Lynagh <[EMAIL PROTECTED]>**20071201170626]
[tcfail175 output is now the same in 6.8 and HEAD
Ian Lynagh <[EMAIL PROTECTED]>**20071201170041]
[Give GHC 6.8 output for tcfail168
Ian Lynagh <[EMAIL PROTECTED]>**20071201165537]
[Add 6.8-branch output for rw
Ian Lynagh <[EMAIL PROTECTED]>**20071201164015]
[rebindable8 is still broken in GHC 6.8
Ian Lynagh <[EMAIL PROTECTED]>**20071201163425]
[FD1 now behaves the same in 6.8 and HEAD
Ian Lynagh <[EMAIL PROTECTED]>**20071201162525]
[fromdos test FD1
Ian Lynagh <[EMAIL PROTECTED]>**20071201162347]
[Add expected output for 1744
Ian Lynagh <[EMAIL PROTECTED]>**20071201162018]
[add test for #1914
Simon Marlow <[EMAIL PROTECTED]>**20071130132214]
[add test for #1744
Simon Marlow <[EMAIL PROTECTED]>**20071130100927]
[Update output for change to deriving (Trac #1935)
[EMAIL PROTECTED]
[Test for Trac #1935
[EMAIL PROTECTED]
[add test for #1916
Simon Marlow <[EMAIL PROTECTED]>**20071128115654]
[fix this test when CLEANUP is not done
Simon Marlow <[EMAIL PROTECTED]>**20071127123338]
[Be more consistent in printing about framework failures
Ian Lynagh <[EMAIL PROTECTED]>**20071123181712]
[Make Church2 fail as it (currently) should
[EMAIL PROTECTED]
[Simple{20,22,23} now pass in the 6.8 branch
Ian Lynagh <[EMAIL PROTECTED]>**20071122134526]
[openFile008: do "ulimit -n 1024" first
Simon Marlow <[EMAIL PROTECTED]>**20071122102223
Hopefully fixes this on MacOS where the limit is 256 by default
]
[remove the _with_prefix forms of compile_and_run, add cmd_prefix() config
instead
Simon Marlow <[EMAIL PROTECTED]>**20071122102129]
[add test for #1753
Simon Marlow <[EMAIL PROTECTED]>**20071122094048]
[Update output (slight improvements on the whole)
[EMAIL PROTECTED]
[Update output
[EMAIL PROTECTED]
[Update test output
[EMAIL PROTECTED]
This is actually *slightly* worse than before, because of the
error-message refactoring in TcUnify. The previous message
reported a mis-match between
(a->Bool) and (Bool -> Bool)
whereas the new one only complains of mismatch between
a and Bool
Trouble is, the old one was vulnerable to the problem
described in the comments on tcSubExp.
So I'm accepting this regression for now; let's see if
anyone notices!
]
[Update for rebindable changes (Trac #1537)
[EMAIL PROTECTED]
[Test for Trac #1913
[EMAIL PROTECTED]
[countReaders001 now passes
Simon Marlow <[EMAIL PROTECTED]>**20071120143539]
[test repeated open/close of 1000 files
Simon Marlow <[EMAIL PROTECTED]>**20071120114757]
[Tests for Trac #1825
[EMAIL PROTECTED]
[augment ghci025 to cover #1847
[EMAIL PROTECTED]
- need to test :browse! in different contexts
- -s has disappeared, see #1799 for sorting
- 'imported from' clarified to 'imported via'
]
[Test for Trac #959
[EMAIL PROTECTED]
[Test for Trac #1806
[EMAIL PROTECTED]
[update to test #1873, #1360
Simon Marlow <[EMAIL PROTECTED]>**20071116155948]
[Add -XExistentialQuantification flag (should have been there all along)
[EMAIL PROTECTED]
[small tidyups to thie Makefile
Simon Marlow <[EMAIL PROTECTED]>**20071116131914]
[fix this test for case sensitive filenames
Simon Marlow <[EMAIL PROTECTED]>**20071116131837]
[1603 depends on the gcc version, which we can't easily account for, so skip
Simon Marlow <[EMAIL PROTECTED]>**20071116093428]
[outofmem2 passes on the stable branch
Simon Marlow <[EMAIL PROTECTED]>**20071115162357]
[add test for #1679
Simon Marlow <[EMAIL PROTECTED]>**20071115131517]
[FIX #1840: outofmem and outofmem2 need ulimit settings to avoid filling up swap
Simon Marlow <[EMAIL PROTECTED]>**20071114164525]
[Accept output
Pepe Iborra <[EMAIL PROTECTED]>**20071114232553
The test now actually passes
]
[Add test for #1892 - :info of a bkpt binding panics
Pepe Iborra <[EMAIL PROTECTED]>**20071114233829]
[Adapt output to improvements in the TTerm pretty printer
Pepe Iborra <[EMAIL PROTECTED]>**20071114232445]
[simplify ghci024.py, fix -fprint-bind-result, -fno-ignore-breakpoints,
-fprint-bind-contents
[EMAIL PROTECTED]
- as ghc 6.9 flags are beginning to diverge, we need
a simpler way of specifying version-specific flags
or settings: have two entirely separate sections,
instead of one section with too many patches.
even if that doubles the script size, it should
be easier to edit. also add a warning that
ghci024.stdout is a generated file.
- default for -fprint-bind-result has changed
- -fno-ignore-breakpoints is gone
- -fprint-bind-contents is new
]
[Add test for #782
Simon Marlow <[EMAIL PROTECTED]>**20071114141111]
[Test for Trac #1662
[EMAIL PROTECTED]
[Test for Trac #1888
[EMAIL PROTECTED]
[Add a test for #1827 (:print panicswith overloaded values))
Pepe Iborra <[EMAIL PROTECTED]>**20071113172048]
[Remove some tabs in break020.hs
Pepe Iborra <[EMAIL PROTECTED]>**20071113163912]
[Accept output
Pepe Iborra <[EMAIL PROTECTED]>**20071113163451]
[FIX ghci024 for unregisterised, powerpc_apple_darwin, and ghc-6.8
[EMAIL PROTECTED]
- for unregisterised platforms, default is '-fno-asm-mangling'
- powerpc_apple_darwin fails on ':set -package ghc' (#1845)
- for ghc 6.8, -fno-run-cps and -fno-convert-to-zipper-and-back
do not exist
]
[add test for #1603
Simon Marlow <[EMAIL PROTECTED]>**20071108164056]
[add test for #1852
Simon Marlow <[EMAIL PROTECTED]>**20071108143533]
[Update output
Ian Lynagh <[EMAIL PROTECTED]>**20071111223835]
[TAG 2007-11-11
Ian Lynagh <[EMAIL PROTECTED]>**20071111170042]
Patch bundle hash:
4b19936fd0f9511b0840cf07f1cd5f75647e25e9
_______________________________________________
Cvs-ghc mailing list
Cvs-ghc@haskell.org
http://www.haskell.org/mailman/listinfo/cvs-ghc
