HPDF needs upper bounds

[nomeata]

Some users of my tttool software, who I previously instructed to use cabal to build the software (I am now starting to provide static binaries) fell over HPDF not building with newer versions of GHC (because of MonadFail and use of old functions from containers). Upstream is dormant, and those who maintain forks so far did not agree to take over the package.

I would appreciate if a trustee could put in upper bounds so that https://matrix.hackage.haskell.org/#/package/HPDF is no longer red.

[hvr]

Before: https://matrix.hackage.haskell.org/package/HPDF@1553983961

After: https://matrix.hackage.haskell.org/package/HPDF@1553991520

---

I've mostly fixed it up and it should be all green how; however, there's some corner cases I can't fix up via revisions yet ...

It's due to improper CPP conditionals such as

#if __GLASGOW_HASKELL__ >= 710
import qualified Control.Monad.Except as EXC
#else
import qualified Control.Monad.Error as EXC
#endif

(which should have been #if MIN_VERSION_mtl(2,2,1) instead) and this is e.g. observable via

$ cabal v2-repl -zb HPDF==1.4.10  -w ghc-7.10.3 --constraint 'mtl < 2.2.1'
...
Configuring library for HPDF-1.4.10..
Preprocessing library for HPDF-1.4.10..
Building library for HPDF-1.4.10..

Graphics/PDF/Image.hs:49:18:
    Could not find module ‘Control.Monad.Except’

so in order to make the metadata accurate, I'd need to be able to revise in a conditional tighter lower bound on mtl if GHC >= 7.10, i.e.

if impl(ghc >= 7.10)
   build-depends: mtl >= 2.2.1

[nomeata]

Thanks!

[hvr]

@nomeata btw, you should take a look at https://matrix.hackage.haskell.org/package/tttool (you may need to look at a non-latest report in case the current one doesn't have the ghc-8.0/8.2 columns yet) :-)

[nomeata]

Ah, thanks!