> The overhead for running a few select extensions, compared to native code, would be minimal.
Only if the browser is designed that way. If the browser has only global settings implementing per-site handling can impose the need of extension reimplementing half of the browser (in extreme case, see how Google made their "extension" of the IE which effectively installed the 90% of the Chrome inside of IE).
And you have to trust extension authors, they can do their own things.
Old Opera had what I needed built in, it worked without the need of even enabling javascript subsystem. I don't know of any other browser able to work that way.
>Only if the browser is designed that way. If the browser has only global settings implementing per-site handling can impose the need of extension reimplementing half of the browser (in extreme case, see how Google made their "extension" of the IE which effectively installed the 90% of the Chrome inside of IE).
That's not an extension running on IE to deactivate JS. It's a completely different rendering engine running inside an IE shell.
The two concepts are so different it's not even worth commenting.
>Old Opera had what I needed built in, it worked without the need of even enabling javascript subsystem. I don't know of any other browser able to work that way.
It's trivial (one setting toggle) to disable Javascript in both Chrome and Safari (and I presume Firefox too).
> It's trivial (one setting toggle) to disable Javascript in both Chrome and Safari (and I presume Firefox too).
Only globally it is trivial, but locally per web site, even in the same tab, certainly not:
You know, one single page can fetch the scripts from different addresses. Opera allowed me to run the script from site A but not from site B while opening the page on site C, from which the script is also not run.
The effect is that you can enable, for example, some real functionality on some site but even if some annoying trackers or ad are included they are never executed. And it's not about being ad-hostile, it's about me as a user selecting which sites I want to ever let run dynamic content and consuming other sites as the static content. Breaks post on Google's "blogger" but the rest of the web is mostly better, faster and much safer.
Only if you never even thought about that, and especially if you've never considered what you would have to do if you would write such an extension can you even argument with the "global toggle" which does something completely different, and really simple.
But what Opera had is something completely different, built in natively.
I recommend you stop commenting on software whose functionality you have never experienced and are judging based on your interpretation of second-hand descriptions.
I'm not sure what software is that. When you assume, etc...
I've started with Mosaic and Netscape (and on Sun OS for that matter), used Mozilla for a few years, then FF on Windows, and finally Safari and then Chrome for as long as both existed on the Mac (and have even used KHTML --in the form of Konqueror-- for light browsing on my dev box for a couple of years in the early 00s). Those, as an end user.
As a web dev for testing, besides IE, I have also fired up Opera for years, including back when it had an ugly QT facade and with its own rendering engine. I've never liked it though, and it never stuck to be used as the main browser.
Being able to toggle Javascript on/off with a single checkbox: all browsers have that.
"Oh, but Opera has it built-in and I can't be bothered to install a third party exception", on the other hand, might be true -- that doesn't make that much of a valid reason to stick to an older browser to me eyes.
As far as I read the source of the extension it's still per tab, not dependent on the origin of the script but just the URL of the whole tab and then it's all or nothing, completely different from Old Opera's behavior?
The overhead for running a few select extensions, compared to native code, would be minimal.
And if Opera implements its native extensions as a scripting API (lots of apps do), then zero.