Removing/deprecating -fvia-c

[Dave Bayer]

On Feb 16, 2010, at 12:51 PM, Don Stewart wrote:

> about the time -fvia-C dies?

I wrote more lines of raw C code in my youth than I'd care to remember, before coming to the realization that the most important benchmark is "brain time" not "cpu time". With this revelation, there is no language like Haskell.

Given the choice, I vote for dedicating our precious development resources to the future of Haskell, not the past of Perl and gcc.

Worrying about what we lose is important, but I'd propose that we continue to benchmark the CURRENT version of GHC using -fvia-c against FUTURE versions of GHC, however it evolves.

This doesn't sound fair, but that's my entire point. GHC will get better faster, the sooner that the development team can forget about supporting -fvia-c.