
Dear Bjorn, Attached is a patch that fixes a context reduction stack overflow in your dimensional package. I noticed something weird though (that's why I'm CCing the ghc list). When I cabal build dimensional-0.8.2 I first get the context reduction stack overflow when I then build it again I get the error again (as expected). However when I try to build it for the third time it succeeds! Why is this? I attached the build log. Regards, Bas

Am 29.03.2011 13:46, schrieb Bas van Dijk:
Dear Bjorn,
Attached is a patch that fixes a context reduction stack overflow in your dimensional package.
I noticed something weird though (that's why I'm CCing the ghc list). When I cabal build dimensional-0.8.2 I first get the context reduction stack overflow when I then build it again I get the error again (as expected). However when I try to build it for the third time it succeeds! Why is this?
Your log shows that it re-compiles fine in the second attempt (don't ask me why) saying "7 of 7". The same happens when compiling for profiling (see .p_o files). C.
I attached the build log.
Regards,
Bas
_______________________________________________ Glasgow-haskell-users mailing list Glasgow-haskell-users@haskell.org http://www.haskell.org/mailman/listinfo/glasgow-haskell-users

On Tue, Mar 29, 2011 at 15:46, Bas van Dijk
Attached is a patch that fixes a context reduction stack overflow in your dimensional package.
Thanks Bas, was the build failure on GHC 7.0.3? I don't believe this problem occurred on 7.0.1 and 7.0.2 although I'm not sure about the latter. I'm a little behind on GHC versions myself, waiting for 7.0.3 to get in the HP.
I noticed something weird though (that's why I'm CCing the ghc list). When I cabal build dimensional-0.8.2 I first get the context reduction stack overflow when I then build it again I get the error again (as expected). However when I try to build it for the third time it succeeds! Why is this?
I attached the build log.
It looks like the first two times it build all seven modules, failing on CGS, while the last time it decides to build only CGS. Strange indeed. Anyway, I'll upload a patched dimensional to Hackage right now. But I'll need to figure out how to get my ssh key on code.haskell.org before I can push the patch to the repo... Thanks, Bjorn

On 29 March 2011 14:12, Bjorn Buckwalter
On Tue, Mar 29, 2011 at 15:46, Bas van Dijk
wrote: Attached is a patch that fixes a context reduction stack overflow in your dimensional package.
Thanks Bas, was the build failure on GHC 7.0.3? I don't believe this problem occurred on 7.0.1 and 7.0.2 although I'm not sure about the latter. I'm a little behind on GHC versions myself, waiting for 7.0.3 to get in the HP.
Yes, I'm using GHC-7.0.3
I noticed something weird though (that's why I'm CCing the ghc list). When I cabal build dimensional-0.8.2 I first get the context reduction stack overflow when I then build it again I get the error again (as expected). However when I try to build it for the third time it succeeds! Why is this?
I attached the build log.
It looks like the first two times it build all seven modules, failing on CGS, while the last time it decides to build only CGS. Strange indeed.
Anyway, I'll upload a patched dimensional to Hackage right now. But I'll need to figure out how to get my ssh key on code.haskell.org before I can push the patch to the repo...
Just mail your public key to support@community.haskell.org. Thanks, Bas
participants (3)
-
Bas van Dijk
-
Bjorn Buckwalter
-
Christian Maeder