FYI — I just pushed this; validation went through, and a simple test of the new functionality succeeds.

I put a verbose comment in Note [Initializing globals] in CoreMonad.

commit 163de25813d12764aa5ded1666af7c06fee0d67e
Author: Nicolas Frisby <nicolas.frisby@gmail.com>
Date:   Wed Jul 3 18:23:54 2013 -0500

    include FastString.string_table in CoreMonad.reinitializeGlobals



On Wed, Jul 3, 2013 at 2:26 PM, Simon Peyton-Jones <simonpj@microsoft.com> wrote:

Nick

 

Reading your (excellent) reply to Joachim, that would be an excellent thing to do.

 

Please comment it copiously with a concrete example (a la Joachim).

 

Simon

 

From: ghc-devs-bounces@haskell.org [mailto:ghc-devs-bounces@haskell.org] On Behalf Of Nicolas Frisby
Sent: 03 July 2013 20:22
To: ghc-devs@haskell.org
Subject: add FastString.string_table to CoreMonad.reinitializeGlobals?

 

Was the FastString Table intentionally left out of the reinitializeGlobals mechanism?

 

This has lead to some confusion in using pieces of the front-end within GHC plugins. 

 

Is there foreseeable harm in adding it? I've done so locally, and minimal testing seems fine.

 

Thanks.