343680 - jsinterp.c warns too much

Status: RESOLVED FIXED

(Core :: JavaScript Engine, defect, P3)

Description

[Blake Kaplan (:mrbkap) (inactive)]

Currently, when compiling jsinterp.c with -pedantic, there are hundreds upon hundreds of warnings by using GCC's __extension__ extension, we can shut these warnings up.

There are two things to note:
-- __extension__ doesn't appear to work in front of 'goto'. I get syntax errors. I was able to work around this by using |__extension__ ({goto foo;})| which is ugly, but appears to work.

-- Since I was here, I decided to nuke the 'ISO C forbids conversion of object pointer to function pointer type' warning that we get casting to/from JSPropertyOp.

Comment 1

[Blake Kaplan (:mrbkap) (inactive)]

Attachment: Fix, v1

The JS_EXTENSION_STMT is so ugly, my eyes bleed, but it works...

Comment 2

[Brendan Eich [:brendan]]

Comment on attachment 228193 [details] [diff] [review]
Fix, v1

How about JS_EXTENSION_ as the one-param macro's name, then?  Reads better, still distinct from JS_EXTENSION as a keyword macro (barely, but who needs to see STMT to differentiate crap from crap :-P).

/be

Comment 3

[Blake Kaplan (:mrbkap) (inactive)]

Attachment: Updated

Comment 4

[Brendan Eich [:brendan]]

Comment on attachment 228331 [details] [diff] [review]
Updated

>-    op = (JSOp) *pc;
>-    goto *normalJumpTable[op];
>+    DO_NEXT_OP(0);
> 
> #else  /* !JS_THREADED_INTERP */

Given that the code you changed to DO_NEXT_OP(0); is #ifdef JS_THREADED_INTERP, I'd rather preserve the explicit two lines that load op and do the goto (with the JS_EXTENSION_(...) of course).

r=me with that and provided you tested the non-gcc case.

/be

Comment 5

[Blake Kaplan (:mrbkap) (inactive)]

Attachment: With that fixed

I'll get someone to test on Windows before checking in.

Comment 6

[Blake Kaplan (:mrbkap) (inactive)]

Attachment: Updated to trunk and compiling on Windows

Seth was nice enough to test my patch on Windows, this is updated to current trunk and actually compiles on Windows.

Comment 7

[(not reading, please use [e-mail] instead)]

Comment on attachment 228338 [details] [diff] [review]
Updated to trunk and compiling on Windows

rs=sspitzer, I've built and smoke tested mrbkap's last patch on win32 / trunk.

Comment 8

[Blake Kaplan (:mrbkap) (inactive)]

Fix checked into trunk.