Activity (RSS)

I'm not seeing the above patch in master, and SVG elements aren't created as HTML elements. It seems this ticket should be reopened.

This is reproduced by the test specs/frame in thatcher/env-js/tree/frame-scope as of fa042d516cf1c26bafcd2e10ac61b6295d2d70e1.

The QUnit tests loaded into each ...

This is reproduced by the test specs/frame in thatcher/env-js/tree/frame-scope as of fa042d516cf1c26bafcd2e10ac61b6295d2d70e1.

This test fails with error messag...

ignore me.

The changes seem fine. I'll look into urlparse and adjust as needed.

thanks for doing this!

--nickg

ahhh,just looked at changes

delete foo.bar
delete foo['bar']

are all ok (and in one case necessary)

it's the ones like

delete foo
delete this

that are bad.

I just pushed 41401804df70f9e2de229a82ba081385377aa186 to thatcher/env-js/master. This eliminates all of the delete operation complaints from both jslint and "r...

I declare minor victory in this ticket. The first 4 are due to html5parser. And the later ones are bogus delete operations, which is covered by another ticket.

...

While investigating a ecma5 strict mode error, I raw into

-        getComputedStyle : function(element, pseudoElement){
-            if(CSS2Properties){
-     ...

Sorry, problem I thought I had was actually stupid bug in test. (Any way to just delete a ticket?)

I think each of these could be replaced with 'foo = null;' but I need to
double check.

howdy,

none is in html5 parser....

$ find src -name '*.js' | xargs grep -n 'delete '
// (edited)
src/parser/domparser.js:43:        delete tmpNode;
src/parser/...

checking into this, but generally I understand this issue. I'm hoping none of this is in the html 5 parser.

I think in attempt to prevent memory leaks, a lot of the code involving frames and scopes does this:

delete foo;

In ECMA5, this is deprecated. You can't delet...

ok it's in.

should be 100% invisible/compatible/backwards-compatible unless you are actually using load events

Just filing this in case someone is making release notes.

img.src = XXX and link.href=XXX handling is spotty and does not give the end-user(application) control...

fixed!

I'm just putting this in for the record. I have a patch momentarily.

in parser/htmldocument.js

__extend__(HTMLDocument.prototype,{
    open : function(){
     ...