Previous week Up Next week

Hello

Here is the latest Caml Weekly News, for the week of April 03 to 10, 2007.

  1. camlp4: let binding

camlp4: let binding

Archive: http://groups.google.com/group/ocaml-developer/browse_frm/thread/60e0ea039394bff6

Hugo Ferreira asked, Martin Jambon said, and Nicolas Pouillard answered:
> > I would like to use a reserved word and substitute that by a function 
> > call. For example, the following: 
> >
> > let  _ = HEAP in 
> > let h1 = HEAP in 
> >
> > would become 
> >
> > let __heap = new_heap () in 
> > let     h1 = new_heap () in 


As always it will be a lot simpler to do that kind of thing using a 
filter in camlp4 3.10. 
... 
match (* do things bottom-up *) super#expr e with 
| <:expr@loc< let _ = HEAP in $e$ >> -> 
     <:expr@loc< let __heap = new_heap () in $e$ >> 
| <:expr@loc< let $p$ = HEAP in $e$ >> -> 
     <:expr@loc< let $p$ = new_heap () in $e$ >> 
| e -> e 
... 

> You shouldn't try to do this because the parser looks only one token ahead 
> to make its decision. If you add a rule that starts from "let" (it has 
> to), the token which enables the parser to select this rule is in position 
> 3, so it comes too late. Camlp4 will not warn you about the conflict but 
> fail during preprocessing because it will choose either the predefined 
> "let" rule or yours without knowing if it's the right one. 

In fact it's wrong camlp4 can takes more than one token of look ahead. 
It will try to match the input with all the firsts terminals of a rule. 
However you're right to discourage him to try that kind of thing. 
Indeed it's highly dependent on the left factorization mechanism 
performed by camlp4.
		
Hugo Ferreira asked and Nicolas Pouillard answered:
> Just a comment: my greatest difficulty with camlp4 is twofold: 
>
> - knowing how the AST maps to the actual Ocaml syntax. 


You can ask to camlp4 itself! 
Using the old one... 

$ cat test.ml 
(* here is some expression that you want to understand *) 
fun e -> <:expr< let x= HEAP in $e$ >> 

$ camlp4o q_MLast.cmo pr_o.cmo test.ml 
let _ = 
  fun e -> 
    MLast.ExLet 
      (_loc, false, [MLast.PaLid (_loc, "x"), MLast.ExUid (_loc, "HEAP")], e) 

> - knowing when/where the camlp4 parser does its transformations so that 
> I may alter these transformations (match and replace expressions). 

That's not really transformations of the AST. 
However there is some grammar tranformations: 
When you add a rule to a grammar entry at a specific level camlp4 
insert it in a tree to factor out the prefix or rules as much as 
possible. 

e1 = SELF; "+"; e2 = SELF -> ... 
e1 = SELF; "*"; e2 = SELF -> ... 

is in fact something like: 

e1 = SELF; [ "+"; e2 = SELF -> ... | "*";  e2 = SELF -> ... ] 

So if you add something make it as much as possible a prefix of existing ones. 

[...] 

> In effect I write Ocaml "expression" which is automatically matched and 
> I only need write an Ocaml expression to introduce the changes I want. 
> All this using simply using the quotation system. The rules and the AST 
> structure itself should, in my humble opinion, be avoided. 

Indeed, extension that are highly dependent on the actual grammar 
should be avoided. 

> This will also allow the underlying camlp4 system to change without unduly 
> affecting user's code. 

Exactly!
		

Using folding to read the cwn in vim 6+

Here is a quick trick to help you read this CWN if you are viewing it using vim (version 6 or greater).

:set foldmethod=expr
:set foldexpr=getline(v:lnum)=~'^=\\{78}$'?'&lt;1':1
zM

If you know of a better way, please let me know.


Old cwn

If you happen to miss a CWN, you can send me a message and I'll mail it to you, or go take a look at the archive or the RSS feed of the archives.

If you also wish to receive it every week by mail, you may subscribe online.


Alan Schmitt