
So, let's talk about plans for xmonad and ghc 6.12. I am currently building xmonad and xmonad-contrib on ghc 6.12.1rc2 and it works fine, except I get a lot of warnings, which fall into three main categories: * Warnings about using base-3 instead of base-4. Currently xmonad doesn't build with base-4 because of the Exception vs. OldException thing. What's our plan for dealing with this? We should switch to using the extensible exception stuff, but how do we continue to support base-3? (*Should* we continue to support it?) * Warnings about redundant imports; apparently 6.12 is more aggressive about noticing these. These can be cleaned up, which is probably a good idea anyway. Or, if people don't like this for some reason, we can disable this warning. * Lots of warnings about actions which return something other than () but whose results are neither bound nor explicitly discarded. As discussed on IRC the right thing to do here is probably to just pass -fno-warn-unused-do-bind to ghc (as opposed to "fixing" all of these). Thoughts? -Brent

byorgey:
So, let's talk about plans for xmonad and ghc 6.12. I am currently building xmonad and xmonad-contrib on ghc 6.12.1rc2 and it works fine, except I get a lot of warnings, which fall into three main categories:
* Warnings about using base-3 instead of base-4. Currently xmonad doesn't build with base-4 because of the Exception vs. OldException thing. What's our plan for dealing with this? We should switch to using the extensible exception stuff, but how do we continue to support base-3? (*Should* we continue to support it?)
base-3 is around for at least one more release. I would propose we make the change to base 4 in the next stable release.
* Warnings about redundant imports; apparently 6.12 is more aggressive about noticing these. These can be cleaned up, which is probably a good idea anyway. Or, if people don't like this for some reason, we can disable this warning.
If it is accurate, there's no harm being tidy.
* Lots of warnings about actions which return something other than () but whose results are neither bound nor explicitly discarded. As discussed on IRC the right thing to do here is probably to just pass -fno-warn-unused-do-bind to ghc (as opposed to "fixing" all of these).
Warning are annoying :) -- Don
participants (2)
-
Brent Yorgey
-
Don Stewart