! (the 'impossible' happened)
(GHC version 7.7.20130109 for arm-unknown-linux):
expectJust initTcInteractive
This is a known bug <http://hackage.haskell.org/trac/ghc/ticket/7490>. If
you just try the build again after getting the error, it usually works.
--
St
Somebody claiming to be Stephen Paul Weber wrote:
I've decided to move on to the essential ARM port of my GHC
cross-compiler. I got pretty far.
I also needed to install LLVM, since there is no native code generator for
ARM. However, it seems like something in the build system i
Somebody signing messages as Stephen Paul Weber wrote:
I'm not 100% sure that the attached patch is the right solution, but it
seems to work both on my local system as well as with my cross-compiler,
and is not so dramatic as to remove the integralness check entirely.
Feedback welcome.
Somebody signing messages as Stephen Paul Weber wrote:
Somebody claiming to be Daniel Fischer wrote:
On Monday 07 January 2013, 12:17:41, Stephen Paul Weber wrote:
I think my vote would be to restructure the checks such that non-integral
types are specially handled.
The purpose of the test
Somebody claiming to be Daniel Fischer wrote:
On Monday 07 January 2013, 12:17:41, Stephen Paul Weber wrote:
I think my vote would be to restructure the checks such that non-integral
types are specially handled.
The purpose of the test was to check whether two types have the same size, if
I
Somebody signing messages as Stephen Paul Weber wrote:
Somebody signing messages as Stephen Paul Weber wrote:
Somebody claiming to be Ian Lynagh wrote:
On Thu, Dec 20, 2012 at 01:20:46PM -0500, Stephen Paul Weber wrote:
Somebody claiming to be Stephen Paul Weber wrote:
libraries/base/Foreign
Somebody signing messages as Stephen Paul Weber wrote:
Somebody claiming to be Ian Lynagh wrote:
On Thu, Dec 20, 2012 at 01:20:46PM -0500, Stephen Paul Weber wrote:
Somebody claiming to be Stephen Paul Weber wrote:
libraries/base/Foreign/C/Types.hs:162:25:
Do you know why the test is failing
Somebody claiming to be Ian Lynagh wrote:
On Thu, Dec 20, 2012 at 01:20:46PM -0500, Stephen Paul Weber wrote:
Somebody claiming to be Stephen Paul Weber wrote:
>libraries/base/Foreign/C/Types.hs:162:25:
>
>Not in scope: type constructor or class `HTYPE_FLOAT'
After much more di
Somebody claiming to be Ian Lynagh wrote:
On Sat, Dec 08, 2012 at 06:32:42PM +, Stephen Paul Weber wrote:
checking sys/eventfd.h usability... no
checking sys/eventfd.h presence... yes
What does this mean? Is it a problem?
There should be more information about what failed in config.log
Somebody claiming to be Nathan Hüsken wrote:
On 01/05/2013 05:20 PM, Stephen Paul Weber wrote:
Somebody claiming to be Nathan Hüsken wrote:
Time.hsc: In function '_hsc2hs_test45':
Time.hsc:266:20: error: storage size of 'test_array' isn't constant
Which Time.hsc?
Hmm. Perhaps the libraries did *not* in fact build properly. What happens
when you try to build by itself?
Indeed, /home/ls/src/ghc-7.7.20121213/libraries/base/dist-install/build/Prelude.hi
does not exist.
--
Stephen Paul Weber, @singpolyma
See <http://sin
Somebody claiming to be Nathan Hüsken wrote:
OK, I managed to set the sysroot. Now the buil fails with this:
Try building ghc-prim with your cross-compiler by hand. Sometimes the GHC
build system eats useful error messages.
--
Stephen Paul Weber, @singpolyma
See <http://singpolyma.net&g
, though I have no NDK experience.
--
Stephen Paul Weber, @singpolyma
See <http://singpolyma.net> for how I prefer to be contacted
edition right joseph
signature.asc
Description: Digital signature
___
Cvs-ghc mailing list
Cvs-ghc@haskell.org
http:/
in the first place.
Hard to say. I've used it a couple times, mostly when there are C enums
involved. Usually in my FFI work I just write the imports by hand (though
this also allows for dangerously writing `Int` when you mean `CInt`, etc).
--
Stephen Paul Weber, @singpolyma
See <http
Somebody claiming to be Stephen Paul Weber wrote:
libraries/base/Foreign/C/Types.hs:162:25:
Not in scope: type constructor or class `HTYPE_FLOAT'
After much more digging into the source, it seems that the `Storable` class
for the CFloat and CDouble types assumes that HTYPE_FLOA
est solution would be to preserve that target
triple from `./configure` and always pass it through to LLVM.
Thoughts?
--
Stephen Paul Weber, @singpolyma
See <http://singpolyma.net> for how I prefer to be contacted
edition right joseph
signature.asc
Description: Digital signature
Somebody claiming to be Stephen Paul Weber wrote:
Somebody claiming to be Stephen Paul Weber wrote:
When I try to build base for QNXNTO (BlackBerry 10) using my
cross-compiler I get an error about "test_array is not of static
size" or similar.
Looking up CLK_TCK in the QNX
Somebody claiming to be Stephen Paul Weber wrote:
The attached patches are needed to cross-compile with GHC for QNXNTO
And I forgot to attach... Attached now.
--
Stephen Paul Weber, @singpolyma
See <http://singpolyma.net> for how I prefer to be contacted
edition right joseph
t with:
. $BBNDK_INSTALL_PATH/bbndk-env.sh
# Create a build.mk that uses INTEGER_LIBRARY=integer-simple
perl boot
./configure --target=i486-pc-nto-qnx8.0.0 --build=i686-linux-gnu
--host=i486-pc-nto-qnx8.0.0
make
replace i486-pc-nto-qnx8.0.0 with arm-unknown-nto-qnx8.0.0eabi to target ARM.
--
Stephe
Somebody claiming to be Stephen Paul Weber wrote:
When building base with my QNXNTO cross-compiling GHC, I get the following:
libraries/base/Foreign/C/Types.hs:162:25:
Not in scope: type constructor or class `HTYPE_FLOAT'
Another question: if `base` cannet build without these constants,
Somebody claiming to be Stephen Paul Weber wrote:
When I try to build base for QNXNTO (BlackBerry 10) using my
cross-compiler I get an error about "test_array is not of static size" or
similar.
-#if defined(CLK_TCK)
+#if defined(CLK_TCK) && !defined(__QNXNTO__)
(#co
Somebody claiming to be Stephen Paul Weber wrote:
Somebody claiming to be Duncan Coutts wrote:
On 8 December 2012 20:36, Stephen Paul Weber wrote:
The ./Setup program is passing the selected gcc to hsc2hs, so if you
want to use a different gcc, then use --with-gcc=whatever, and then
./Setup
s there another build setting I need to change?
--
Stephen Paul Weber, @singpolyma
See <http://singpolyma.net> for how I prefer to be contacted
edition right joseph
___
Cvs-ghc mailing list
Cvs-ghc@haskell.org
http://www.haskell.org/mailman/listinfo/cvs-ghc
Somebody claiming to be Duncan Coutts wrote:
On 8 December 2012 20:36, Stephen Paul Weber wrote:
The ./Setup program is passing the selected gcc to hsc2hs, so if you
want to use a different gcc, then use --with-gcc=whatever, and then
./Setup will also pass that on to hsc2hs.
This seems likely
e working for sure. What are these values primarily for, and
how should I improve the configure system to work in my case?
--
Stephen Paul Weber, @singpolyma
See <http://singpolyma.net> for how I prefer to be contacted
edition right joseph
___
Cvs-ghc mailing list
Cvs-ghc@haskell.org
http://www.haskell.org/mailman/listinfo/cvs-ghc
ven having hsc2hs accept multiple
--cc and using the last one would work in this case.
--
Stephen Paul Weber, @singpolyma
See <http://singpolyma.net> for how I prefer to be contacted
edition right joseph
___
Cvs-ghc mailing list
Cvs-ghc@
hecking for sys/eventfd.h... no
checking for eventfd... no
What does this mean? Is it a problem?
--
Stephen Paul Weber, @singpolyma
See <http://singpolyma.net> for how I prefer to be contacted
edition right joseph
___
Cvs-ghc mailing list
Cvs
unsafePerformIO (sysconf (#const _SC_CLK_TCK) >>= return . fromIntegral)
But I don't actually understand what this code does or what the error I get
means. I'm not sure my patch is safe. I'd like to get more clarity on
this.
--
Stephen Paul Weber, @singpolyma
See <
28 matches
Mail list logo
