zetafunction

This is what Chrome already does: it uses a utility process to host the common file dialog. See https://cs.chromium.org/search/?q=CallGetOpenFileName&sq=pac...

Two other reasons that haven't been mentioned yet:

- MSVC had trouble with dead code elimination in some cases: fixing https://crbug.com/684105 reduced binary size by almost 900KB on Windows.

- It lets tooling like https://chromium.googlesource.com/chromium/src/+/master/docs... and the custom clang plugin checks run on Windows-specific code.

Chrome for Android blocks autoplaying videos. To get around this, sites now use JS to decode and render video into a canvas. For example: https://bugs.chromium.org/p/chromium/issues/detail?id=178297.... As one might imagine, this is not so great for power usage…

Actually, both Chromium [1] and Firefox [2] are working on showing plain HTTP as insecure.

[1] https://blog.chromium.org/2016/09/moving-towards-more-secure... [2] https://blog.mozilla.org/tanvi/2016/01/28/no-more-passwords-...

Imagine a more powerful API existed that an extension could use to control other extensions. Now the badware extensions start using the more powerful API and you're back to square one...

From the article:

Are Firefox extensions any better?

To be honest, no.

The interstitial for blocked content is implemented on a per-page level. You can test this yourself by creating a test page with an iframe to http://ianfette.org.

Suppose that Chrome just tried to replace the malicious embedded content. What happens when that embedded content is styled with CSS properties to make it hidden from the user?

As a Chrome developer, I agree that process-per-tab absolutely should not be a replacement for fixing bugs. However, I think it's important to point out that process-per-tab is not just for stability: it's also critical for security.

> What's the mindset that ends up with Chrome going out of its way to swap OK/Cancel when you enable DNT?

Chrome dev here. The way the Chrome settings web UI is written does not lend itself to strong consistency, just eventual, as devs notice it and fix it. Your DNT example was fixed last week in https://codereview.chromium.org/665113003.

Another example of inconsistent button ordering: the overlay for disconnecting a managed profile has its buttons reversed from the usual order, while the overlay for disconnecting an unmanaged profile does not.

This is neat... but why not use the copy/paste event handlers where possible?

Oh oops. I left out the link to the sample in my last post: http://jsfiddle.net/BE6bR/1

Try it out =)

I'm not sure, but several guesses: 1) The clipboard APIs aren't well known. 2) You can't normally set HTML content in the clipboard with IE since it only supports 'Text' and 'URL' types (http://msdn.microsoft.com/en-us/library/ms536744(v=vs.85).as...). 3) The IE interface (via window.clipboardData) might trigger an infobar. I didn't really test this though.

Normal web pages can't trigger copy/cut/paste events. They can access the clipboard via the events (which are automatically fired when the corresponding action is invoked by the user) documented in http://dev.w3.org/2006/webapi/clipops/.

> When you say, deny, to me the user must go to their browser settings and set it manually?

AFAIK: Safari does not expose a setting for this (but I didn't look). Chrome uses Chrome app/extension permissions to control this setting. I'm not sure how Firefox does it.

> If so, what bearing does that have on what we're discussing?

You asked "Is there a way to trigger the copy, cut, and paste events?". You don't need to trigger them in this case since the browser automatically fires them when the user invokes the corresponding edit action.

> As for setting the clipboard data, I have seen no way of setting that through the copy, paste or cut events.

Here is a quick sample that works in everything except IE. I'm too impatient to figure out why it doesn't work in IE but http://msdn.microsoft.com/en-us/library/hh772145(v=vs.85).as... provides IE-centric documentation.

There is a way to trigger the events from JS via document.execCommand('copy'/'cut'/'paste'), but the default is to deny. However, if you just want to alter the behavior of ^C, you simply need to override the oncopy event--there's no need to trigger the event manually. To actually set clipboard data, IE uses window.clipboardData, while Firefox/Chrome/Safari all use event.clipboardData.

For reference, http://www.w3.org/TR/clipboard-apis/ is the W3 draft for clipboard events.

Why aren't the copy/cut/paste events be sufficient for this? There's no need to trigger a programmatic copy event in this context, since you're just modifying the default behavior of copy.

Google is using SSL with ECDHE which provides perfect forward secrecy. Having a copy of the SSL certificate won't do the NSA any good.

A privileged user has nothing to do with it. Running as a non privileged user will not stop keyloggers, cookie theft, privilege escalation exploits, etc.

Multiprocess is also important for security. It means that simply finding a renderer exploit isn't enough to pwn your system if the renderer processes have no privileges to write/execute arbitrary files.