Update to compile under recent Haskell Platform about lslforge HOT 10 CLOSED

This seems like a worthy quest, and I've made a start.

Currently I have a fork that works with ghc-6.12.3, and I'll be hooking up the tests into a proper Cabal test-suite before investigating a Monad Transformer issue. Once that is resolved, I'll be working on ghc-7.0.* changes.

from lslforge.

OK, my ghc-7.0.4 branch is now merged to master.

Travis CI is building and testing LSLForge successfully for all but a few CompilationServer unit tests, and those don't actually look to be in sync with the source.

A quick test with Eclipse shows the LSLForge binary to be functional, but there are a bunch of fail calls throughout the codebase, and these cause failures with transformers >= 0.2.2.0.

It can be built with the latest cabal-install binary download, on Linux and macOS at least, so the build story is much nicer now, but there's a long way to go yet.

from lslforge.

Now at ghc-8.0.2

from lslforge.

Awesome news, @simon-nicholls !

from lslforge.

Over to you for review! It would be good if you could trial my pull request against some production LSL code before merging.

from lslforge.

Agreed. It might be a good idea to compare compiled LSL-files for starters.

from lslforge.

Simon, we have some issues with the Java generated code from Haskell. Hope you see the message

from lslforge.

@simon-nicholls, I'm about to close this, but would appreciate if you look at the recent changes at the https://github.com/raysilent/lslforge/tree/0.1.9.4 tree. I had to introduce a few workarounds that resulted from the code that you upgraded to the latest Haskell libraries. Please have a look. Thank you

from lslforge.

from lslforge.

No problem @simon-nicholls. You did such a fantastic job, that wasn't even possible to imagine that would ever be accomplished.

After testing for quite a while, with the silly patch I provided, it seems working, although for some reason I noticed that that "TupleX" initialization has to happen earlier in the generated/InitAll.java, so I'm manually moving 2 lines up at the moment. The symptom is that the whole Eclipse hangs during initialization, the window freezes and doesn't want to continue. LSLForge executable, even if stopped, restarts itself and nothing really happens. I wonder if this is related or not, but this is my observation.

So first of all, they have to be there (which was broken) and probably have to be moved (my patch basically appends them). If it all makes sense to you, I can't figure out why TupleX disappeared in the first place.

from lslforge.