- 12 Jun, 2012 23 commits
-
-
Ian Lynagh authored
-
Ian Lynagh authored
-
Ian Lynagh authored
This avoid lots of converting back and forth between the two types.
-
Ian Lynagh authored
In particular, this makes life simpler when we want to use a general GHC SDoc in the middle of some LLVM.
-
Ian Lynagh authored
-
Ian Lynagh authored
-
Ian Lynagh authored
-
Ian Lynagh authored
-
Ian Lynagh authored
-
Ian Lynagh authored
It was equivalent to showPpr anyway.
-
Ian Lynagh authored
-
Ian Lynagh authored
-
Ian Lynagh authored
-
Ian Lynagh authored
We also store the Doc still, although we don't have to; we could just convert back again, at the cost of some inefficiency. But we do need the String, so that we can print it in the Show instance.
-
Ian Lynagh authored
-
Ian Lynagh authored
-
Ian Lynagh authored
-
Ian Lynagh authored
-
Ian Lynagh authored
-
Ian Lynagh authored
-
Ian Lynagh authored
-
Ian Lynagh authored
-
Ian Lynagh authored
-
- 11 Jun, 2012 17 commits
-
-
Ian Lynagh authored
To help with this, we now also pass DynFlags around inside the SpecM monad.
-
Ian Lynagh authored
-
Ian Lynagh authored
-
Ian Lynagh authored
-
Ian Lynagh authored
-
Ian Lynagh authored
-
Ian Lynagh authored
This shouldn't be any slower; the FilePaths were already going via String so that we could normalise them. This means that very early on, when printing flag parsing errors, we can do so without having a DynFlags yet.
-
Ian Lynagh authored
-
Ian Lynagh authored
-
Ian Lynagh authored
-
Ian Lynagh authored
-
Ian Lynagh authored
-
Ian Lynagh authored
-
Ian Lynagh authored
A side-effect is that we can no longer use the LogAction in defaultErrorHandler, as we don't have DynFlags at that point. But all that defaultErrorHandler did is to print Strings as SevFatal, so now it takes a 'FatalMessager' instead.
-
Ian Lynagh authored
We don't use it yet
-
Ian Lynagh authored
-
Ian Lynagh authored
We now include the String and the SDoc in the exception, and don't flatten them into a String until near the top-level
-