[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

ghc6 6.12.3-1 going to experimental



Just FYI, I'm about to upload GHC 6.12.3 to experimental.  Here's the
changelog entry:

ghc6 (6.12.3-1) experimental; urgency=low

  * New upstream release
    * ghci :m now works with module names containing ' (Closes: #580568)
  * Put haddock and the shared libraries into their own packages.
    (Closes: #571537)
  * Only insert packages in gen_contents_index which are both registered
    and for which the html directory exists.
  * Use ghc-pkg --global in ghc6's trigger to avoid accessing HOME.
    (Closes: #578679)
  * Fix directories in haddock manpage.  (Closes: #578321)
  * Apply some patches from Gentoo (thanks, Sergei Trofimovich) hopefully
    fixing ia64 build.
  * Downgrade ghc6's dependency on perl to Suggests and have ghc6-doc
    depend on perl.  (Closes: #553332)
  * Standards-Version 3.9.0 (no changes necessary)

 -- Kari Pahula <kaol@debian.org>  Sun, 18 Jul 2010 14:57:32 +0300


This one'll have to go through NEW.  I put haddock outside of the ghc6
package but I didn't want to name the package as haddock, if only
since it would have ghc's version number and not haddock's.  I used
ghc6-haddock instead.  I don't really expect there to be other haddock
builds around anytime soon, since it's pretty tightly tied up with
ghc's API right now.

Another new introduction was ghc6-dynamic.  Many people can avoid
installing that, saving a good chunk of disk space.

I'm not really happy about the state of index and contents list
generation in ghc6-doc.  Currently, it uses haddock for that and it
needs the .haddock files for the task and those reside in -dev
packages and we can't just keep copies of them in -doc packages, due
to ABI breakage between GHC releases.

I was thinking of solving this by not using haddock in ghc6's trigger.
The individual packages keep their own index and contents files and it
wouldn't be too difficult to write a, say, perl script for combining
those.  The only immediate loss would be the annotations about
multiple entries with the same name being either functions or whatever
in the contents index.  If you didn't understand the last sentence
then you didn't need it.

Attachment: signature.asc
Description: Digital signature


Reply to: