> function () { <x/>.@* } function () { <x/>[@[*]]; } > function () { <x/>[@[*]]; } function () { <x/>.@[*]; } The fact that it takes more than 1 trip through uneval to settle on a final form confuses my fuzzer in bug 349611. > function () { <x/>.@[*]; } function () { <x/>[@[*]]; } The fact that it *never* settles on a final form (bouncing back and forth between two forms) bothers me even more. These aren't necessarily bugs, but they do interfere with certain methods of testing decompilation.

Need shaver to sub for mrbkap while mrbkap drives back to Rice via L.A. /be

Shaver, please renom for 181 assuming this passes your review. Also, be sure to tell us how it feels to be the new mrbkap!

Comment on attachment 235177 [details] [diff] [review] fix r=shaver. I was the old mrbkap before mrbkap was the new mrbkap.

Fixed on trunk. Checking in jsopcode.c; /cvsroot/mozilla/js/src/jsopcode.c,v <-- jsopcode.c new revision: 3.146; previous revision: 3.145 done /be

Comment on attachment 235177 [details] [diff] [review] fix a=beltzner on behalf of the 181drivers

Fixed on the 1.8 branch. /be

That fixed <x/>.@* to not go into a back-and-forth loop, but the back-and-forth loop is still there. I filed bug 350226 for that.

Checking in regress-349956.js; /cvsroot/mozilla/js/tests/e4x/Regress/regress-349956.js,v <-- regress-349956.js initial revision: 1.1

verified fixed 1.8 20060827 windows/mac*/linux both 1.8.0.7, 1.8 agree expected: function () { <x/>.@*; } actual: function () { <x/>.@*; }

verified fixed 1.9 20060827 windows/mac*/linux