Hi,
we have several issues on Haddock's ghc-7.6 branch.  I fixed some of
those, but some still persist.

Isolated, none of those issues are a big deal.  But combined they are.
Is there a way we can prevent situations like this in the future?  Would
it work to establish the convention that if you change Haddock you have
to build a vanilla Haddock (say without in-ghc-tree) and run the test
suite?

I fixed the following:

 (a) deepseq was missing from the 'library' dependencies

 (b) All test were broken due to the version bump to 2.12.0

 (c) Some tests were broken due to the fact that () is not hyperlinked
     anymore.

 (d) The CrossPackageDocs test was broken


@David My changes are on [1].  Can you look at those, and let me know
what you think.  Notes about the individual points are further bellow.


The following issues still persist:

 (e) The library component does not build anymore, due to a refactoring
     of Main.hs, which was done in:

      commit 315338287ea84b525da7d8fa8252cc9ec99597bb
      Author: Ian Lynagh <ig...@earth.li>
      Date:   Tue Jun 12 14:38:01 2012 +0100

          Follow changes in GHC

 (f) The Unicode test fails for me.  Looks like characters outside the
     ASCII range are just striped from the documentation.

 (g) The TypeOperators test is brocken.


I guess (c) was an intended change, right?  What commit introduced that?

What about (d)?  I made two commits to fix that, [2] and [3].  I'm
confident about [2], but is [3] an intended change?

I looked into (e) some time ago, and that time it was not obvious to me
how to fix it.  IIRC, it would be easy to fix it by changing the type of
`createInterfaces`, but this would require a user to initialize a GHC
session, etc..  I guess from a users point of view, the current type is
more useful, right?

I'm entirely puzzled about (f).  Was this an intended change?

(g) is a compile error.  I guess this is due to a GHC change.  @David
can you help with that?

Cheers,
Simon

[1] https://github.com/sol/haddock/commits/dev/
[2] 
https://github.com/sol/haddock/commit/a427422fae865a06f5fa23d904e7e12cc5b1af6e
[3] 
https://github.com/sol/haddock/commit/66cc7cc1db9fdc0b09651dbac6e5454380c11340

_______________________________________________
Cvs-ghc mailing list
Cvs-ghc@haskell.org
http://www.haskell.org/mailman/listinfo/cvs-ghc

Reply via email to