patch applied (packages/base): new Control.Category, ghc ticket #1773

Simon Marlow simonmarhaskell at gmail.com
Mon Oct 29 09:48:29 EDT 2007


Ashley Yakeley wrote:
> Sun Oct 28 19:25:26 PDT 2007  Ashley Yakeley <ashley at semantic.org>
>   * new Control.Category, ghc ticket #1773
> 
>      ./Control/Compositor.hs -> ./Control/Category.hs
>     M ./Control/Applicative.hs -3 +1 r1
>     M ./Control/Arrow.hs -7 +7 r1
>     M ./Control/Category.hs -9 +21 r2
>     M ./base.cabal r1

This broke GHC's build.  Haddock fails to parse Control.Category due to a 
bug in Haddock, I've committed a workaround for this.  Second issue: all 
the arrow tests are failing:

=====> arrowfail001(normal)
cd ./arrows/should_fail && 
'/64playpen/simonmar/testing/compiler/stage2/ghc-inpl
ace' -no-recomp -dcore-lint -dcmm-lint -Dx86_64_unknown_linux  -c 
arrowfail001.h
s    >arrowfail001.comp.stderr 2>&1
Actual stderr output differs from expected:
--- ./arrows/should_fail/arrowfail001.stderr.normalised 2007-10-29 
13:30:42.0000
00000 +0000
+++ ./arrows/should_fail/arrowfail001.comp.stderr.normalised    2007-10-29 
13:30
:42.000000000 +0000
@@ -1,5 +1,8 @@

-arrowfail001.hs:15:35:
-    Ambiguous type variable `a' in the constraint:
-      `Foo a' arising from a use of `foo' at arrowfail001.hs:15:35-37
-    Probable fix: add a type signature that fixes these type variable(s)
+arrowfail001.hs:15:6:
+    Can't find interface-file declaration for variable Control.Arrow.>>>
+      Probable cause: bug in .hi-boot file, or inconsistent .hi file
+      Use -ddump-if-trace to get an idea of which file caused the error
+    In the expression: proc x -> case x of Bar a -> foo -< a
+    In the definition of `get':
+       get = proc x -> case x of Bar a -> foo -< a


It appears that GHC is looking for Control.Arrow.>>>, which has moved to 
Control.Category.  Third issue:

=====> tc192(normal)
cd ./typecheck/should_compile && 
'/64playpen/simonmar/testing/compiler/stage2/gh
c-inplace' -no-recomp -dcore-lint -dcmm-lint -Dx86_64_unknown_linux  -c 
tc192.hs
    -fno-warn-incomplete-patterns >tc192.comp.stderr 2>&1
Compile failed (status 256) errors were:

tc192.hs:25:21: Not in scope: `>>>'

tc192.hs:40:8: Not in scope: `>>>'

tc192.hs:48:27: Not in scope: `>>>'

tc192.hs:50:22: Not in scope: `>>>'

tc192.hs:54:36: Not in scope: `>>>'

tc192.hs:141:21: Not in scope: `>>>'

tc192.hs:141:48: Not in scope: `>>>'


Could someone fix these up please?

Cheers,
	Simon



More information about the Cvs-libraries mailing list