701620 - Various parse node kind disambiguations

Status: RESOLVED FIXED

(Core :: JavaScript Engine, defect)

Description

[Jeff Walden [:Waldo]]

Some parse node kinds have split personalities depending on context.  It would be conceptually simpler to just have distinct kinds in these cases.

Comment 1

[Jeff Walden [:Waldo]]

Attachment: Extract PNK_DEFXMLNS from PNK_DEFAULT

Comment 2

[Jeff Walden [:Waldo]]

Attachment: Extract PNK_ADD and PNK_SUB from PNK_PLUS and PNK_MINUS

Comment 3

[Jeff Walden [:Waldo]]

Attachment: Extract PNK_FORIN from PNK_IN

There's at least an argument that in [x for (y in z)], the for node shouldn't be PNK_FORIN but should be different.  I haven't quite seen value in doing it, so for now there is this reuse.

Comment 4

[Jeff Walden [:Waldo]]

Attachment: Rename PNK_DO to PNK_DOWHILE

I think of them as do-while loops, not do-loops, so I find PNK_DOWHILE more readable than PNK_DO.  No code change, just a name change.

Comment 5

[Jeff Walden [:Waldo]]

Attachment: Split PNK_LC into PNK_STATEMENTLIST and PNK_XMLCURLYEXPR

Comment 6

[Jason Orendorff [:jorendorff]]

Comment on attachment 573951 [details] [diff] [review]
Split PNK_LC into PNK_STATEMENTLIST and PNK_XMLCURLYEXPR

>+      case PNK_XMLCURLYEXPR:
>+#if JS_HAS_XML_SUPPORT
>+        JS_ASSERT(pn->isArity(PN_UNARY));
>+        if (!EmitTree(cx, bce, pn->pn_kid))
>+            return JS_FALSE;
>+        if (Emit1(cx, bce, pn->getOp()) < 0)
>+            return JS_FALSE;
>+#else
>+        JS_NOT_REACHED("XML curly expression without XML support");
>+#endif
>+        break;

Can you just #if JS_HAS_XML_SUPPORT around the whole case, including the break?

All this is great stuff and long overdue.

If you feel like renaming PNK_{RC,LB,RB,LP,RP} please go for it. They are totally cryptic.

Comment 7

[Jeff Walden [:Waldo]]

https://hg.mozilla.org/integration/mozilla-inbound/rev/9e33bed54f9a
https://hg.mozilla.org/integration/mozilla-inbound/rev/6ae5203fe3e1
https://hg.mozilla.org/integration/mozilla-inbound/rev/1030be78cb51
https://hg.mozilla.org/integration/mozilla-inbound/rev/c7099e04ec39
https://hg.mozilla.org/integration/mozilla-inbound/rev/db3280d6199d

I do indeed feel like it.  But it's not as simple as that, yet, because of kind overloading at least for PNK_RB/PNK_RC.  I don't know about the others.  But I do indeed plan to get to it, over time.  :-)

Comment 8

[Marco Bonardo [:mak]]

https://hg.mozilla.org/mozilla-central/rev/9e33bed54f9a
https://hg.mozilla.org/mozilla-central/rev/6ae5203fe3e1
https://hg.mozilla.org/mozilla-central/rev/1030be78cb51
https://hg.mozilla.org/mozilla-central/rev/c7099e04ec39
https://hg.mozilla.org/mozilla-central/rev/db3280d6199d

Comment 9

[Justin Dolske [:Dolske]]

Let's test your knowledge, and see what you've learned so far... What color are the parse node kinds? PNNNNNNK!

Ahem. Sorry.

Comment 10

[Jeff Walden [:Waldo]]

/me smiles