Enabling -fwarn-tabs for commit test (for contrib)
Hi, I want to add '-fwarn-tabs' to the warnings run when testing patches? (and clean up the couple uses of tabs). The warning is documented in ghc-6.8, missing in ghc-6.6 (but it could still be there?), ghc-6.6 does not support the current darcs contrib, due to various pragmas in use, such as -XDeriveDataTypeable, among others. Changing the -fglasgow-exts flags to more specific pragmas should carry the same risks too. Cheers, Adam
On Tue, Jun 23, 2009 at 11:08:32PM -0400, Adam Vogt wrote:
Hi,
I want to add '-fwarn-tabs' to the warnings run when testing patches? (and clean up the couple uses of tabs).
The warning is documented in ghc-6.8, missing in ghc-6.6 (but it could still be there?), ghc-6.6 does not support the current darcs contrib, due to various pragmas in use, such as -XDeriveDataTypeable, among others.
Changing the -fglasgow-exts flags to more specific pragmas should carry the same risks too.
Cheers, Adam
I think we were waiting on -fwarn-tabs until we stopped supporting 6.6. Now that we've done that we should absolutely activate -fwarn-tabs. In the case of -fglasgow-exts, do be careful. I think there were some bugs between 6.8 and 6.10 and extension flags. Cheers, Spencer Janssen
* On Tuesday, June 23 2009, Spencer Janssen wrote:
I think we were waiting on -fwarn-tabs until we stopped supporting 6.6. Now that we've done that we should absolutely activate -fwarn-tabs.
In the case of -fglasgow-exts, do be careful. I think there were some bugs between 6.8 and 6.10 and extension flags.
I've updated most uses of -fglasgow-exts, except in the XMonad.Config namespace, and XMonad.Hooks.ManageDocks (since it could still be moved into core, where 6.6 compatiblity is desirable). vav tested it on ghc-6.8 too. Thanks, Adam
vogt.adam:
Hi,
I want to add '-fwarn-tabs' to the warnings run when testing patches? (and clean up the couple uses of tabs).
Seems reasonable.
The warning is documented in ghc-6.8, missing in ghc-6.6 (but it could still be there?), ghc-6.6 does not support the current darcs contrib, due to various pragmas in use, such as -XDeriveDataTypeable, among others.
Changing the -fglasgow-exts flags to more specific pragmas should carry the same risks too.
--- Don
participants (3)
-
Adam Vogt -
Don Stewart -
Spencer Janssen