magnus:

On Sun, Aug 16, 2009 at 6:00 AM, Don Stewart wrote:

...

Hey all,

Anyone have thoughts on whether we treat haddock failing at build time as a non-fatal error now?

A couple of projects aren't building, e.g.

   haskell-xosd 0.1.1-1

due to a haddock failure. I'm wondering if we should handle that gracefully.

Maybe the way to deal with is to have cabal2arch fail to create a PKGBUILD if the hackage page flags build failure?

Hmm. Hackage is overly restrictive though (i.e. it doesn't have C libraries installed). We build more things successfully than Hackage does. I'd prefer it not to be fatal for a package's docs to fail to build. -- Don