Live data from Hacker News

All modern web-browsers except IE9 are vulnerable to a huge memory leak

code.google.com

61–70 of 71 posts

Re: All modern web-browsers except IE9 are vulnerable to a huge memory leak

#61
Firefox also causes not insignificant pixmap memory growth in the Xorg process (xrestop was used to verify) when running this test: https://code.google.com/p/chromium/issues/detail?id=36142#c6...

Is there a Gecko bug for this issue yet, so I can add myself to the CC list?

Re: All modern web-browsers except IE9 are vulnerable to a huge memory leak

#62

Earlier quoted context omitted.

I've had Ruby processes run wild (think turning off the garbage collector) and grow to 20-30GB on my MacBook and it didn't make my system unusable. Things slowed down, but I could pretty easily open up the Activity Monitor and kill the offending process.

In the highly unlikely scenario where you actually allowed the system to allocate 30GB of swap space, this might not cause the system to become unresponsive since most of the memory in "use" would never be touched, as it comes from discarded but uncollected Ruby objects.

This is a pretty solid hypothesis, but wouldn't the same be true of the exploited cached images? They are essentially the same thing: stale resources that won't be used again, and those regions need to be free'd up.

Re: All modern web-browsers except IE9 are vulnerable to a huge memory leak

#63

Firefox also causes not insignificant pixmap memory growth in the Xorg process (xrestop was used to verify) when running this test: https://code.google.com/p/chromium/issues/detail?id=36142#c6... Is there a Gecko bug for this issue yet, so I can add myself to the CC list?

For which issue? The page in question is loading lots of different (from the browser's point of view) images. They won't be released until sometime later (in particular, for Gecko until after GC runs on the style data structures). So I would expect memory usage to grow here here until said GC happens, then drop, then grow again, etc.

While the images are live, they're stored on the X server, because that's where they need to be drawn; in many cases this will just place them directly into the graphics memory. There _are_ existing Gecko bugs on trying to reduce X server resource usage by dynamically deciding which images to store in the client and which on the server... The bugs are mostly motivated by thin client setups where the X client actually has a lot more in the way of hardware resources than the X server does. Of course those are high-latency X setups (compared to your typical desktop), which makes storing the images you really need on the server all the more important.

Re: All modern web-browsers except IE9 are vulnerable to a huge memory leak

#64

Earlier quoted context omitted.

Hmm, I always prefer Firebug for development, but development and browsing are two separate tasks, so I have no problem using both. Chrome is a solid browser too, though.

Just booted up firebug due to your comment because I hadn't used it in a while. Here is my reaction after using Chromium dev tools: - JavaScript auto complete text doesn't display right, it is a different font size shadow that obscures the typed text. - Pressing tab multiple times will not cycle autocomplete, but instead jump the focus out of the console and into the search prompt. - Really minor, but I prefer a term…

Hmm, I'll have to try the chrome dev tools, I didn't know they were so advanced, thank you!

Re: All modern web-browsers except IE9 are vulnerable to a huge memory leak

#65

Earlier quoted context omitted.

In the highly unlikely scenario where you actually allowed the system to allocate 30GB of swap space, this might not cause the system to become unresponsive since most of the memory in "use" would never be touched, as it comes from discarded but uncollected Ruby objects.

This is a pretty solid hypothesis, but wouldn't the same be true of the exploited cached images? They are essentially the same thing: stale resources that won't be used again, and those regions need to be free'd up.

I suspect there are two major causes for the differences: (1) the faster allocation rate in the browser prevents the system from keeping OS/GUI-critical pages in RAM, and (2) the browser is constantly checking the cache for matching images, so the pages used by the cache are being kept in RAM and/or repeatedly loaded and unloaded.

Re: All modern web-browsers except IE9 are vulnerable to a huge memory leak

#66
post #31

Now, let's see which of these browsers fix this issue the fastest. My bet is that Chrome will be the first.

It looks like the first was IE9 - see http://code.google.com/p/chromium/issues/detail?id=36142#c53 . Chrome doesn't fix the related bug for more than a year - http://code.google.com/p/chromium/issues/detail?id=36142 . Safari wins with more than two years - https://bugs.webkit.org/show_bug.cgi?id=23372 .

Even older (dec 2008) than Webkit: http://code.google.com/p/chromium/issues/detail?id=5688

Re: All modern web-browsers except IE9 are vulnerable to a huge memory leak

#68

Earlier quoted context omitted.

I hate how Opera never gets any love. It's been the most innovative browser for years now, it had tabs before other browsers even had an MDI and was fast even on my PC with 16 MB RAM.

I was a long time Opera user and love its configurability and full keyboard support for general browsing, but the Chromium dev tools won me over from a developer perspective and I disliked switching between the two.

I've tried using Dragonfly several times and I can never even get started with it. It's one of those programs which I simply cannot penetrate the UI. This isn't surprising when it's an app like Blender, but I'm not sure how a browser debugger could be so confusing.

Re: All modern web-browsers except IE9 are vulnerable to a huge memory leak

#69

Earlier quoted context omitted.

Hmm, I always prefer Firebug for development, but development and browsing are two separate tasks, so I have no problem using both. Chrome is a solid browser too, though.

Just booted up firebug due to your comment because I hadn't used it in a while. Here is my reaction after using Chromium dev tools: - JavaScript auto complete text doesn't display right, it is a different font size shadow that obscures the typed text. - Pressing tab multiple times will not cycle autocomplete, but instead jump the focus out of the console and into the search prompt. - Really minor, but I prefer a term…

The console isn't as advanced in Firebug as Chromium or Safari, no doubt. I prefer the CSS and DOM inspection system, however - having to pause JS execution to look at the current state is not nearly as convenient. If you want to do that, though, Firebug offers that as well.

Re: All modern web-browsers except IE9 are vulnerable to a huge memory leak

#70
post #14
post #3

Haha, the irony.

I dont think this is very ironic. IE9 is native to MSWIN and is written by the people who have the deepest knowledge of the operating system. They probably found this issue and fixed it before it was exposed to the other teams like Goog and Mozilla.

The people with the deepest knowledge of the operating system are writing the operating systems (kernel, etc).

I'm not sure what you mean by IE being 'native' to Windows. Regardless of what MS tried to pull 10 years ago, IE being included with Windows does not make it part what the rest of us call the OS. It's an application.

Post reply on HN