GHCJS now runs Template Haskell on node.js - Any interest in out of process TH for general cross compilation?
stegeman at gmail.com
Sat Jul 5 20:56:40 UTC 2014
I'm not sure I correctly understand your approach, but to have the template
haskell reification work without any runtime communication with the
compiler you'd have to include the entire typechecker state, at least for
all names reachable from the splice (see
, the Quasi class for the required functionality). This would mean
serializing all names with types modules, annotations, instances. I briefly
looked into this for GHCJS but decided that just querying the compiler
would be better.
On Sat, Jul 5, 2014 at 10:39 PM, John Meacham <john at repetae.net> wrote:
> The target compiler would have the TH libraries, which could be made
> to be portable. The external program would just extract the TH bits
> and turn them into a program that spits the TH expanded output to a
> new file to compile, and repeat the process til no TH expansions exist
> and finally that is the result you pass to the compiler.
> On Sat, Jul 5, 2014 at 1:09 PM, Luite Stegeman <stegeman at gmail.com> wrote:
> > How would you do reification with that approach?
> > On Sat, Jul 5, 2014 at 9:59 PM, John Meacham <john at repetae.net> wrote:
> >> Actually, I was looking into it a little, and template haskell could
> >> effectively be implemented by a pre-processor and a portable library
> >> that is compiler independent. If one could get ghc to spit out the
> >> template haskell source after it expands it then that can be fed to
> >> jhc as a quick first pass, but ideally the pre-processor TH would
> >> create programs that can be run under the target compiler. that would
> >> bring TH to every haskell compiler.
> >> John
> >> On Sat, Jul 5, 2014 at 10:38 AM, Brandon Allbery <allbery.b at gmail.com>
> >> wrote:
> >> > On Sat, Jul 5, 2014 at 1:34 PM, Carter Schonwald
> >> > <carter.schonwald at gmail.com> wrote:
> >> >>
> >> >> does JHC support template haskell?
> >> >
> >> >
> >> > Pretty sure TH is too closely tied to ghc.
> >> >
> >> > --
> >> > brandon s allbery kf8nh sine nomine
> >> > associates
> >> > allbery.b at gmail.com
> >> > ballbery at sinenomine.net
> >> > unix, openafs, kerberos, infrastructure, xmonad
> >> > http://sinenomine.net
> >> --
> >> John Meacham - http://notanumber.net/
> >> _______________________________________________
> >> Glasgow-haskell-users mailing list
> >> Glasgow-haskell-users at haskell.org
> >> http://www.haskell.org/mailman/listinfo/glasgow-haskell-users
> John Meacham - http://notanumber.net/
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the ghc-devs