Greasemonkey/firefox 58 compability #1249
Conversation
| @@ -27,11 +27,11 @@ window.iitcBuildDate = '@@BUILDDATE@@'; | |||
| window.onload = function() {}; | |||
There was a problem hiding this comment.
something like:
unsafeWindow.onload = exportFunction(function() {}, unsafeWindow);
There was a problem hiding this comment.
I've tested on Greasemonkey on FF and Tampermonkey on Chrome. I don't think setting window.onload had any use, it has already run at this point. With GM4.0 it isn't assigning the correct "window.onload". It doesn't fail and has no effect since it is not used anyway.
A way to do this is and stay compatible with different extensions and browsers is:
if(typeof exportFunction == "function")
exportFunction(function() {}, window, {defineAs:'onload'});
else if(typeof window.wrappedJSObject == "object")
window.wrappedJSObject.onload = function() {};
else
window.onload = function() {};There was a problem hiding this comment.
Oh, good point. That's logically dead code, as is the document.body.onload just below it (introduced by 98f6f19) since document.body as a whole is replaced.
main.js
Outdated
| //originally code here parsed the <Script> tags from the page to find the one that defined the PLAYER object | ||
| //however, that's already been executed, so we can just access PLAYER - no messing around needed! | ||
|
|
||
| if (typeof(window.PLAYER)!="object" || typeof(window.PLAYER.nickname) != "string") { | ||
| var PLAYER = window.PLAYER || window.wrappedJSObject.PLAYER; |
There was a problem hiding this comment.
could just be unsafeWindow in either case
There was a problem hiding this comment.
Are we sure window.wrappedJSObject exist in all contexts that run IITC ?
Another way to write the test is
if (!((typeof(window.PLAYER)=="object" && typeof(window.PLAYER.nickname) == "string")
||(typeof(window.wrappedJSObject)=="object" && typeof(window.wrappedJSObject.PLAYER)=="object" && typeof(window.wrappedJSObject.PLAYER.nickname) == "string"))){
// page doesn’t have a script tag with player information.
if(document.getElementById('header_email')) {
// however, we are logged in.
// it used to be regularly common to get temporary 'account not enabled' messages from the intel site.
// however, this is no longer common. more common is users getting account suspended/banned - and this
// currently shows the 'not enabled' message. so it's safer to not repeatedly reload in this case
// setTimeout('location.reload();', 3*1000);
throw("Page doesn't have player data, but you are logged in.");
}
// FIXME: handle nia takedown in progress
throw("Couldn't retrieve player data. Are you logged in?");
}There was a problem hiding this comment.
@Jormund oh, yes, it needs an extra test for wappedJSObject for TM
|
with GM there is no "unsafeWindow" anymore workaround: |
| @@ -27,11 +27,11 @@ window.iitcBuildDate = '@@BUILDDATE@@'; | |||
| window.onload = function() {}; | |||
|
can anyone merge this and create a new release, please? |
| //originally code here parsed the <Script> tags from the page to find the one that defined the PLAYER object | ||
| //however, that's already been executed, so we can just access PLAYER - no messing around needed! | ||
|
|
||
| if (typeof(window.PLAYER)!="object" || typeof(window.PLAYER.nickname) != "string") { | ||
| var PLAYER = window.PLAYER || (unsafeWindow && unsafeWindow.PLAYER); |
There was a problem hiding this comment.
| var PLAYER = window.PLAYER || (unsafeWindow && unsafeWindow.PLAYER); | |
| var PLAYER = window.PLAYER || (typeof unsafeWindow !== 'undefined' && unsafeWindow.PLAYER); |
Otherwise we have TypeError when window.PLAYER is not defined (which is normal when user is not logged in yet).
|
closed because stale project |
GM4.0 / FF 58+ -> Content scripts do not have direct access to webpage Javascript content.
see: greasemonkey/greasemonkey#2653 (comment)
this will (atleast) fix the startup error. So far I didn't recognized any other problems.