Whats is the nature of the auto reload on permissions change

[al_9x]

1) is it equivalent to the default Fx F5? or a cache quickreload? or something else?
2) What is is the reason for the particular choice?

[Giorgio Maone]

1) is it equivalent to the default Fx F5? or a cache quickreload? or something else?

cache quickreload

2) What is is the reason for the particular choice?

Speed & convenience (e.g. in most cases form field filling is kept).

[al_9x]

1) is it equivalent to the default Fx F5? or a cache quickreload? or something else?

cache quickreload

Are you sure? I tried to figure it out but couldn't hit any breakpoints in Main.js (do you have any tips for that?), but it seems that when you temp allow with auto reload, reloadWhereNeeded (LOAD_FLAGS_NONE) is called and not quickReload.

The reason I ask, is that I don't like auto reload and do it manually when needed, but if you are doing the quick and I, regular Fx reload, that's wasteful. So I was going to ask you to expose your reload through a button or something, so one could manually invoke it.

[Giorgio Maone]

1) is it equivalent to the default Fx F5? or a cache quickreload? or something else?

cache quickreload

Are you sure?

Yes. reloadWhereNeeded() reloads every affected tab using the fastest method available for each (i.e. using history-bfcache when possible, cache otherwise and a regular reload if no cache data is available, e.g. after a POST request).

[al_9x]

using the fastest method available

Can you expose your reload so it can be manually invoked on the active tab? You provide an option to disable auto reload, which is good, but doing the regular Fx reload manually wastes time and bandwidth.

Do you have any tips for being able to debug your Main.js? Half the time it doesn't even appear in the file list in venkman, and when it does, breakpoints in it that should be hit are not.

[Giorgio Maone]

using the fastest method available

Can you expose your reload so it can be manually invoked on the active tab? You provide an option to disable auto reload, which is good, but doing the regular Fx reload manually wastes time and bandwidth.

I could, but it's not a priority for obvious reasons.

Do you have any tips for being able to debug your Main.js? Half the time it doesn't even appear in the file list in venkman, and when it does, breakpoints in it that should be hit are not.

Venkman is currently too buggy to debug extensions which export XPCOM components.
Personally I patch my code "live" with debug statement by opening about:config and running this bookmarklet:

Code: Select all

javascript:(function()%20{%20var%20d%20=%20document;%20if%20(d.getElementById("src"))%20return;%20var%20dc%20=%20d.documentElement;%20while(dc.firstChild)%20dc.removeChild(dc.firstChild);%20var%20r%20=%20dc.appendChild(d.createElement("vbox"));%20r.flex=1;%20var%20t%20=%20d.createElement("textbox");%20t.setAttribute("multiline","true");%20t.id="src";t.setAttribute("value",%20'var%20CC=Components.classes,CI=Components.interfaces;\nvar%20ns=CC["@maone.net/noscript-service;1"].getService().wrappedJSObject;\nvar%20fg=CC["@maone.net/flashgot-service;1"].getService().wrappedJSObject;\nvar%20o%20=%20ns;\no.__parent__.eval(function()%20{\n\n\nalert("done")}.toSource())()');%20t.flex="1";t.style.fontFamily="Lucida%20Console,%20Courier";%20r.appendChild(t);%20var%20b%20=%20d.createElement("button");b.setAttribute("label","eval");b.setAttribute("oncommand",%20"eval(document.getElementById('src').value)");%20r.appendChild(b);%20})()

[al_9x]

I could, but it's not a priority for obvious reasons.

Sure, but keep it in mind please.

Personally I patch my code "live"

Thanks for the tip, definitely better than nothing. You'd think they'd have a working debugger by now. However I must be missing something. With your patching I am seeing that reloadWhereNeeded calls browser.webNavigation.reload with LOAD_FLAGS_NONE. Isn't that the same as regular Fx F5 reload?

[Giorgio Maone]

browser.webNavigation.reload with LOAD_FLAGS_NONE. Isn't that the same as regular Fx F5 reload?

Yes it is, but later (in a "http-on-modify-request" observer) the channel flags are forced to VALIDATE_NEVER, meaning that unless stuff had not been cached, the network is not hit. Just use HttpFox to compare a regular reload with a NoScript one.