Hi,
On Thu, Jun 19, 2014 at 03:23:27PM +0100, Zé Loff wrote:
> On Thu, Jun 19, 2014 at 10:22:17AM -0400, Ian Darwin wrote:
> > On 2014-06-19, 9:42 AM, Zé Loff wrote:
> > >Has anyone tried / started to port pandoc, or knows of any good reason
> > >not to? I'd like to give it a shot, but I don't want do duplicate
> > >efforts...
> > >
> > >
> > Yes, I have it mostly done (a dozen or so new dependencies, of course), but
> > it's on hold waiting for the annual churnover of Haskell GHC port.
>
> Nice, thanks. Let me know if I can help.
Short story: try older versions of pandoc which can be built with
aeson-0.6, *not* requiring any library which meta/haskell-platform
depends on to be changed. pandoc-1.12.3.3 may be worth a try.
Long story:
The biggest problem is that the latest version of pandoc needs
aeson-0.7, which needs attoparsec->=0.11.3.4, but the haskell-platform
dictates attoparsec-0.10.4.0 (and even the current drafts of
haskell-platform-2014 still depends on that same version of
attoparsec).
That means that even after the yearly churn of haskell ports (which
I'll hopefully start in two or three weeks), our port of attoparsec
has to stay at 0.10.4.0. For pandoc (or aeson-0.7), we would need
a separate port/package hs-attoparsec-0.11 (or maybe -0.12?).
But with two different ports of hs-attoparsec, for every existing port
depending on hs-attoparsec (0.10), you'll have to make sure that either
- it doesn't pick up hs-attoparsec-0.11 by accident when it's
installed (by looking at its cabal file, and maybe patching it to
depend on attoparsec-<0.11), or
- it builds and works with attoparsec-0.11, depends on it (in the port's
Makefile) and *never* picks up an installed hs-attoparsec (0.10)
because of other dependency constraints.
I'd prefer to not doing such a stunt, because it's unmaintainable.
Better go back a few releases of pandoc (see above) and then try
to find out whether the newer releases really need aeson-0.7 (and
probably other stuff causing similar problems) and wether they could
be patched to still work with aeson-0.6. (Or try to get aeson-0.7
working with attoparsec-0.10)
Ciao,
Kili
