The IPython specification for the metadata does not
define that all fields in metadata must be textual.
Latest Jupyters lab uses that feature so the parsing
mechanism became incomplatible with the protocol.
We change Metadata to be an Aeson.Object and introduce
a 'Metadata' newtype in order to keep that explicit
and open for the later changes if needed.
Fixes#912.
Hoogle v5 uses a different json representation for the data, requires a
new URL and also contains HTML markup inside the json fields.
This commit removes all markup to be able to handle the hoogle responses
using the same structure as with the previous response structure (from
hoogle v4).
As per #716, GHC-7 support has at the moment no priority. Use the GHC7 tag
for the last version that retains the original GHC-7.10 support.
https://github.com/gibiansky/IHaskell/releases/tag/GHC7
(However, it is still quite feasible to support GHC-7.10 in master, see
#747.)
In `haskell-src-exts-1.17`, `Module` was not parameterised yet, so trying
to compile with that version gives
src/IHaskell/Eval/Lint.hs:36:19:
‘SrcExts.Module’ is applied to too many type arguments
In the type ‘SrcExts.Module SrcSpanInfo’
In the type declaration for ‘ExtsModule’
Previously, `cabal install ihaskell` would happily use ipython-kernel-0.8.3, resulting in error
‘protocolVersion’ is not a (visible) field of constructor ‘KernelInfoReply’
Preprocessing test suite 'hspec' for ihaskell-0.8.4.0...
Hspec.hs:20:18:
Could not find module ‘Data.String.Here’
It is a member of the hidden package ‘here-1.2.7@here_1j09jYtlmBB7SYT9WO9k1d’.
Perhaps you need to add ‘here’ to the build-depends in your .cabal file.
Use -v to see a list of the files searched for.
Signed-off-by: Sergei Trofimovich <siarheit@google.com>
