I'm experiencing some unexpected behaviors in my DSL. I've shortened it here to make my point, but it seems that if all my tokens (A,B,C,D) are optional, and I assign onlyone of them, MGrammar assumes that is the last token. I would have expected it to be the first:

Grammar:
module SAMPLE{
language TEST {

syntax Main = a:A? del?
b:B? del?
c:C? del?
d:D?
=>{A=>a,B=>b,C=>c,D=>d} ;


@{Classification["Delimiter"]}
token del = ":";

token A = ("A".."Z" | "a".."z" | "0".."9")+;
token B = ("A".."Z" | "a".."z" | "0".."9")+;
token C = ("A".."Z" | "a".."z" | "0".."9")+;
token D = ("A".."Z" | "a".."z" | "0".."9")+;
}
}

Input (OK): A:B:C:D
Output:
{
A => "A",
B => "B",
C => "C",
D => "D"
}

Input (not OK): A:B
Output:
{
A => null,
B => null,
C => "A",
D => "B"
}

I would have expected:
{
A => "A",
B => "B",
C => null,
D => null
}


Is there anything I can do to make this work as I expected?
//Mikeal

---

If this answers your question, please use the "Answer" button to say so... Mikael - http://blogical.se/blogs/mikael

Hi, Mikael!

The following appears to do what you want, more or less:

syntax Main =
t1:Term t2:Term t3:Term t4:LastTerm
=> {A=>t1,B=>t2,C=>t3,D=>t4}
| t1:Term t2:Term t3:LastTerm => {A=>t1,B=>t2, C=>t3}
| t1:Term t2:LastTerm => {A=>t1,B=>t2}
| t1:LastTerm => {A=>t1} ;

syntax Term = l:Letter del => {l};
syntax LastTerm = l:Letter del? => {l};
@{Classification["Delimiter"]}
token del = ":";
token Letter = ("A".."Z" | "a".."z" | "0".."9")+;

Here's the output it produces, not exactly what you want, but close.

{
A => {
"A"
},
B => {
"B"
}
}

It also requires the presence of non-final delimiters, but allows the last one to be optional, which I suspect is the behavior you wanted.

syntax Main = d:D? del? c:C? del? b:B? del? a:A?

well...closer, but not quite right

Input: A:B
Output:
{
A => "B",
B => "A",
C => null,
D => null
}

Expected:
{
A => "A",
B => "B",
C => null,
D => null
}

---

If this answers your question, please use the "Answer" button to say so... Mikael - http://blogical.se/blogs/mikael

According to the squigglies I might be better off using left or right annotations:

"Annotate the token with left(n) or right(n)"


I tried looking in the help for Production/Term Precedence, but it didn't make me any wiser...:(

//Mikael

---

If this answers your question, please use the "Answer" button to say so... Mikael - http://blogical.se/blogs/mikael

Hi, Mikael!

The following appears to do what you want, more or less:

syntax Main =
t1:Term t2:Term t3:Term t4:LastTerm
=> {A=>t1,B=>t2,C=>t3,D=>t4}
| t1:Term t2:Term t3:LastTerm => {A=>t1,B=>t2, C=>t3}
| t1:Term t2:LastTerm => {A=>t1,B=>t2}
| t1:LastTerm => {A=>t1} ;

syntax Term = l:Letter del => {l};
syntax LastTerm = l:Letter del? => {l};
@{Classification["Delimiter"]}
token del = ":";
token Letter = ("A".."Z" | "a".."z" | "0".."9")+;

Here's the output it produces, not exactly what you want, but close.

{
A => {
"A"
},
B => {
"B"
}
}

It also requires the presence of non-final delimiters, but allows the last one to be optional, which I suspect is the behavior you wanted.