Re: Profiled compielr

2010-10-22 Thread Ian Lynagh
On Fri, Oct 22, 2010 at 02:36:11PM +, Simon Peyton-Jones wrote: > > When building a profiled compiler with > > GhcProfiled = YES > GhcLibWays = v p > > I get the error below in the vector library. I guess we can't build 'vector' > with a profiled compiler? Should the build system not kno

RE: Profiled compielr

2010-10-22 Thread Simon Peyton-Jones
Oh I dunno. It serves to keep annotations working! And remind us of where the shoe pinches. But I don’t feel strongly From: Roman Leshchinskiy [mailto:r...@cse.unsw.edu.au] Sent: 22 October 2010 15:42 To: Simon Peyton-Jones Cc: Ian Lynagh; cvs-ghc@haskell.org Subject: Re: Profiled compielr

Re: Profiled compielr

2010-10-22 Thread Roman Leshchinskiy
Should we just give up and use pragmas to drive SpecConstr? Annotations seem to be more trouble than they are worth. Roman On 22 Oct 2010, at 15:36, Simon Peyton-Jones wrote: > Ian > > > > When building a profiled compiler with > > > > GhcProfiled = YES > > GhcLibWays = v p > > >

Profiled compielr

2010-10-22 Thread Simon Peyton-Jones
Ian When building a profiled compiler with GhcProfiled = YES GhcLibWays = v p I get the error below in the vector library. I guess we can't build 'vector' with a profiled compiler? Should the build system not know that? Simon "inplace/bin/ghc-stage2" -H32m -O -Wall -Werror -H64m -O0-