A cursory look at this proposal seems to me that it's beautiful. But I guess it is not easy to give a robust implementation<br><br><div class="gmail_quote">On Tue, Feb 5, 2013 at 8:05 AM, Gabriel Scherer <span dir="ltr"><<a href="mailto:gabriel.scherer@gmail.com" target="_blank">gabriel.scherer@gmail.com</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">There have been very good inputs by Alain, Leo, Hongbo and Jeremy.<br>
I've considered doing a summary as you request (after all, I have a<br>
blog to feed), but Leo promises that he has some more ideas that he<br>
would expose "early this week", so I've mostly been waiting for that.<br>
<br>
My opinionated one-paragraph (in fact one-list) take-away would be:<br>
- nobody apparently objects to Alain's idea of using lightweight<br>
quotations and annotations to mimic most existing extensions, and him<br>
and Leo have been busy discussing concrete syntax for that (see<br>
Alain's examples here :<br>
<a href="https://github.com/gasche/ocaml-syntax-extension-discussion/wiki/Use-Cases-using-light-annotations" target="_blank">https://github.com/gasche/ocaml-syntax-extension-discussion/wiki/Use-Cases-using-light-annotations</a><br>
)<br>
- there is an ongoing discussion on the robustness of global rewrites<br>
and possible restrictions (Alain promotes the immediately available,<br>
low-tech social approach, Jeremy has a nice technical suggestion, Leo<br>
is going to write something, and a compromise *is* probably possible<br>
along the lines of encapsulating common local use cases as a strongly<br>
suggested library (like type-conv did for type-derived boilerplate))<br>
- Hongbo has been asking for a discussion on the AST definition at the<br>
core of syntax manipulations, but so far nobody had the courage to<br>
look in details into the polymorphic variant solution of Fan<br>
- relatedly, there is no consensus on the neededness and<br>
implementation of antiquotations; so far we're heading towards a<br>
solution where they are not well supported<br>
<br>
Hope that helped.<br>
<div class="HOEnZb"><div class="h5"><br>
On Tue, Feb 5, 2013 at 1:43 PM, Yaron Minsky <<a href="mailto:yminsky@janestreet.com">yminsky@janestreet.com</a>> wrote:<br>
> There have been so many detailed and thoughtful emails that it's hard<br>
> to keep up with the deluge. I would humbly submit that this would be<br>
> a good time for one of the organizers (Leo or Alain) to put together a<br>
> summary of the different positions that have been expressed about how<br>
> ppx might work, so that one could read over and think through those.<br>
> Right now, I have a hard time keeping the current state of the<br>
> conversation in my head, and I'm pretty sure I'm not alone.<br>
><br>
> y<br>
> _______________________________________________<br>
> wg-camlp4 mailing list<br>
> <a href="mailto:wg-camlp4@lists.ocaml.org">wg-camlp4@lists.ocaml.org</a><br>
> <a href="http://lists.ocaml.org/listinfo/wg-camlp4" target="_blank">http://lists.ocaml.org/listinfo/wg-camlp4</a><br>
_______________________________________________<br>
wg-camlp4 mailing list<br>
<a href="mailto:wg-camlp4@lists.ocaml.org">wg-camlp4@lists.ocaml.org</a><br>
<a href="http://lists.ocaml.org/listinfo/wg-camlp4" target="_blank">http://lists.ocaml.org/listinfo/wg-camlp4</a><br>
</div></div></blockquote></div><br><br clear="all"><div><br></div>-- <br>-- Regards, Hongbo