problem getting cabal-install to work on CentOS 5.2

Hi, I am trying to install cabal-install 0.8.2 with GHC 6.12.1 on a CentOS 5.2 system and am running up against a 'ExitFailure 127' error. I can download the package and run 'cabal install' inside the package directory and all will be fine, but if I let cabal download and install the package cabal runs into problems as soon as it tries to configure. For example, if I try 'cabal install permutation --verbose=3' it finishes like this: link: done *** Deleting temp files: Deleting: /tmp/ghc1324_0/ghc1324_0.lpp *** Deleting temp dirs: Deleting: /tmp/ghc1324_0 /tmp/permutation-0.4.132758/permutation-0.4.1/dist/setup/setup configure --verbose=3 --ghc --prefix=/home/chris/.cabal --user --constraint=base ==3.0.3.2 --constraint=ghc-prim ==0.2.0.0 cabal: Error: some packages failed to install: permutation-0.4.1 failed during the configure step. The exception was: ExitFailure 127 I have tried the same exercise with GHC 6.10.4 with cabal-install 0.8.0 the same results. I have downloaded the Cabal sources to try and add some instrumentation to find out what is going wrong I have yet to succeed. Does anybody have any suggestions? Is there an easier way of getting cabal-install working on CentOS? Any help much appreciated! Chris

I suspect CentOS might disable execute permissions on anything under
/tmp. To verify this just try creating a simple sh script under /tmp
and executing it directly with ./test.sh.
-Corey O'Connor
On Mon, Apr 12, 2010 at 5:05 AM, Chris Dornan
Hi,
I am trying to install cabal-install 0.8.2 with GHC 6.12.1 on a CentOS 5.2 system and am running up against a ‘ExitFailure 127’ error.
I can download the package and run ‘cabal install’ inside the package directory and all will be fine, but if I let cabal download and install the package cabal runs into problems as soon as it tries to configure.
For example, if I try ‘cabal install permutation --verbose=3’ it finishes like this:
link: done
*** Deleting temp files:
Deleting: /tmp/ghc1324_0/ghc1324_0.lpp
*** Deleting temp dirs:
Deleting: /tmp/ghc1324_0
/tmp/permutation-0.4.132758/permutation-0.4.1/dist/setup/setup configure
--verbose=3 --ghc --prefix=/home/chris/.cabal --user --constraint=base
==3.0.3.2 --constraint=ghc-prim ==0.2.0.0
cabal: Error: some packages failed to install:
permutation-0.4.1 failed during the configure step. The exception was:
ExitFailure 127
I have tried the same exercise with GHC 6.10.4 with cabal-install 0.8.0 the same results.
I have downloaded the Cabal sources to try and add some instrumentation to find out what is going wrong I have yet to succeed.
Does anybody have any suggestions? Is there an easier way of getting cabal-install working on CentOS?
Any help much appreciated!
Chris
_______________________________________________ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe

Thanks very much Corey -- your suspicion is very well founded, and of course it neatly explains the peculiar behaviour. It looks as if exec has obviously been modified on RHEL and friends to refuse to execute anything in /tmp.
With a bit of luck I should be able to get cabal-install working now...
Chris
-----Original Message-----
From: Corey O'Connor [mailto:coreyoconnor@gmail.com]
Sent: 13 April 2010 2:37 AM
To: Chris Dornan
Cc: Haskell-Cafe@haskell.org
Subject: Re: [Haskell-cafe] problem getting cabal-install to work on CentOS 5.2
I suspect CentOS might disable execute permissions on anything under
/tmp. To verify this just try creating a simple sh script under /tmp
and executing it directly with ./test.sh.
-Corey O'Connor
On Mon, Apr 12, 2010 at 5:05 AM, Chris Dornan
Hi,
I am trying to install cabal-install 0.8.2 with GHC 6.12.1 on a CentOS 5.2 system and am running up against a ‘ExitFailure 127’ error.
I can download the package and run ‘cabal install’ inside the package directory and all will be fine, but if I let cabal download and install the package cabal runs into problems as soon as it tries to configure.
For example, if I try ‘cabal install permutation --verbose=3’ it finishes like this:
link: done
*** Deleting temp files:
Deleting: /tmp/ghc1324_0/ghc1324_0.lpp
*** Deleting temp dirs:
Deleting: /tmp/ghc1324_0
/tmp/permutation-0.4.132758/permutation-0.4.1/dist/setup/setup configure
--verbose=3 --ghc --prefix=/home/chris/.cabal --user --constraint=base
==3.0.3.2 --constraint=ghc-prim ==0.2.0.0
cabal: Error: some packages failed to install:
permutation-0.4.1 failed during the configure step. The exception was:
ExitFailure 127
I have tried the same exercise with GHC 6.10.4 with cabal-install 0.8.0 the same results.
I have downloaded the Cabal sources to try and add some instrumentation to find out what is going wrong I have yet to succeed.
Does anybody have any suggestions? Is there an easier way of getting cabal-install working on CentOS?
Any help much appreciated!
Chris
_______________________________________________ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe

On Wed, Apr 14, 2010 at 06:27:53AM -0500, Chris Dornan wrote:
Thanks very much Corey -- your suspicion is very well founded, and of course it neatly explains the peculiar behaviour. It looks as if exec has obviously been modified on RHEL and friends to refuse to execute anything in /tmp.
I would suspect that you have SELinux to thank for this. It has a lovely tendency to have lots of fun policies about what can be done where, by who, and on files with particular labels. -- Lars Viklund | zao@acc.umu.se

On Apr 14, 2010, at 07:27 , Chris Dornan wrote:
Thanks very much Corey -- your suspicion is very well founded, and of course it neatly explains the peculiar behaviour. It looks as if exec has obviously been modified on RHEL and friends to refuse to execute anything in /tmp.
Unlikely that exec has been modified. SELinux and "noexec" in the fstab are the most likely causes. -- brandon s. allbery [solaris,freebsd,perl,pugs,haskell] allbery@kf8nh.com system administrator [openafs,heimdal,too many hats] allbery@ece.cmu.edu electrical and computer engineering, carnegie mellon university KF8NH

Right: in this case /tmp is being mounted noexec. Thanks again. Chris -----Original Message----- From: Brandon S. Allbery KF8NH [mailto:allbery@ece.cmu.edu] Sent: 14 April 2010 10:41 AM To: Chris Dornan Cc: Brandon S. Allbery KF8NH; 'Corey O'Connor'; Haskell-Cafe@haskell.org Subject: Re: [Haskell-cafe] problem getting cabal-install to work on CentOS 5.2 On Apr 14, 2010, at 07:27 , Chris Dornan wrote:
Thanks very much Corey -- your suspicion is very well founded, and of course it neatly explains the peculiar behaviour. It looks as if exec has obviously been modified on RHEL and friends to refuse to execute anything in /tmp.
Unlikely that exec has been modified. SELinux and "noexec" in the fstab are the most likely causes. -- brandon s. allbery [solaris,freebsd,perl,pugs,haskell] allbery@kf8nh.com system administrator [openafs,heimdal,too many hats] allbery@ece.cmu.edu electrical and computer engineering, carnegie mellon university KF8NH
participants (4)
-
Brandon S. Allbery KF8NH
-
Chris Dornan
-
Corey O'Connor
-
Lars Viklund