WebKit: 4000 memory leaks plugged!

kim kap solkim kap sol
Posted:
in macOS edited January 2014
WebKit developers and open source community fix almost 4000 memory leaks.



The WebKit blog seems to indicate there are only a handful of mysterious leaks left to fix. Once this baby officially hits the street, every app that uses WebKit will instantly gain a benefit. Dashboard widgets memory usage should stop growing in size. Safari will remain fast even after days of usage. Overall, the whole OS will be much healthier.



http://webkit.opendarwin.org/blog/



Want to reap the benefits today using Safari? Download NightShift from VersionTracker or MacUpdate.

Comments

  • Reply 1 of 13
    kim kap solkim kap sol Posts: 2,987member
    If you need a good idea of how much of a difference the latest WebKit provides, I've been using Safari intensively for over 3 hours and it rarely breaks 90MB real memory with a ton of tabs and drops down to as low as 65MB when only one window is open. During normal usage, Safari will hover between 70-80MB.



    The same thing on an older WebKit would make real memory shoot over 120MB+ and stay there, never to come back down until Safari is quit.
  • Reply 2 of 13
    gene cleangene clean Posts: 3,481member
    Opera uses 15-20 MB with one window open. With 13 tabs open, I've seen it go as high as 90MB.



    That's not to say that the WebKit team has not done a great job, they have, but Safari needs more memory leaks plugged if it intends to be the fastest browser on the Mac.
  • Reply 3 of 13
    kickahakickaha Posts: 8,760member
    Since when does small footprint == speed?



    If anything, the two are traditionally a trade off in CS - you can expend space to get speed, or lose speed to save on space.



    Caching uses up space, but speeds up access. Leaks are a whole other matter, of course, but... wondering what you were getting at by tying the two together like that.
  • Reply 4 of 13
    gene cleangene clean Posts: 3,481member
    Quote:

    Originally posted by Kickaha

    [B]Since when does small footprint == speed?



    Since when does big footprint == speed?



    Quote:

    Caching uses up space, but speeds up access. Leaks are a whole other matter, of course, but... wondering what you were getting at by tying the two together like that.



    Speed is a relative issue. If it takes 10 seconds for my Safari window to un-hide, it is not a fast browser. Note that by fast I mean fast as in responsive and generally snappy - not fast in rendering (which depends on the engine, and many other things).



    Memory leaks being plugged will help in this speed/snappiness issue, wether or not it's actual speed or just the perception of it.
  • Reply 5 of 13
    kcmackcmac Posts: 1,051member
    kim kap sol.



    Thanks for this thread. I downloaded Nightshift and read the read me. I didn't really think it would be this easy to use while being non-destructive or risky to my system. It appears that when the new Webkit hits the street we can simply throw away Nightshift and Webkit.app and everything will be just fine.



    I'm not a developer or anything. I just like living out on the edge a little. Gives me something to look at while I wait for Tiger dot three or some other new Apple creation.
  • Reply 6 of 13
    kickahakickaha Posts: 8,760member
    Quote:

    Originally posted by Gene Clean

    Since when does big footprint == speed?



    Google for 'space time tradeoff' or check most any basic CS text worth its salt. There are always exceptions, but the easiest way to make a system faster is throw more memory at it because you can just keep cached data around to speed things up.



    Edit: A couple of top google hits:

    http://www.absoluteastronomy.com/enc...e_tradeoff.htm

    http://www.cprogramming.com/tutorial...timespace.html

    http://en.wikipedia.org/wiki/Space-time_tradeoff



    Your assertion was "Safari needs more memory leaks plugged if it intends to be the fastest browser on the Mac." ie. "Safari needs (reduced memory footprint) if it (will get faster)" ie. (reduced memory footprint) -> (faster)



    Ain't necessarily so, and in most cases it's the opposite. Note that I'm talking about *intentional* memory use. Leaks are bugs, but the above assertion flies in the face of decades of software engineering.



    Quote:

    Speed is a relative issue. If it takes 10 seconds for my Safari window to un-hide, it is not a fast browser. Note that by fast I mean fast as in responsive and generally snappy - not fast in rendering (which depends on the engine, and many other things).



    Memory leaks being plugged will help in this speed/snappiness issue, wether or not it's actual speed or just the perception of it.



    What you're talking about is leaks impacting the VM system, not the speed of Safari itself. In that respect, then yes, plugging the leaks will help with swapping the browser in to physical RAM. That isn't Safari that's getting faster though, and the phrase 'fastest browser on the Mac' is used by marketers to indicate overall speed of connection, rendering, etc. I really doubt that they intend it to mean 'fastest swap in from VM'.



    In any case, plugged leaks == teh good.
  • Reply 7 of 13
    thinkingdifferentthinkingdifferent Posts: 1,604member
    Quote:

    Originally posted by kim kap sol

    WebKit developers and open source community fix almost 4000 memory leaks.





    Will this reduce the demand for memory thus reducing memory prices?
  • Reply 8 of 13
    zozo Posts: 3,117member
    Is there any WebKit based Windows internet browser? Any plans for it?
  • Reply 9 of 13
    andersanders Posts: 6,523member
    Quote:

    Originally posted by ThinkingDifferent

    Will this reduce the demand for memory thus reducing memory prices?



    Yes.



    In the same way as if ten families got rid of their SUVs and the effect rippled into the gasoline marked...
  • Reply 10 of 13
    bwhalerbwhaler Posts: 260member
    Is this the WebKit in the new version of Safari that was just released?



    I've definitely noticed memory problems with Safari since Tiger, and the last patch a couple of weeks ago didn't seem to fix it.



    Can you let us know if this new webkit is in Tiger yet, and if not, any guess if it will make it into 10.4.3?



    Thanks so much.
  • Reply 11 of 13
    kim kap solkim kap sol Posts: 2,987member
    Quote:

    Originally posted by BWhaler

    Is this the WebKit in the new version of Safari that was just released?



    I've definitely noticed memory problems with Safari since Tiger, and the last patch a couple of weeks ago didn't seem to fix it.



    Can you let us know if this new webkit is in Tiger yet, and if not, any guess if it will make it into 10.4.3?



    Thanks so much.



    No...the last few updates via Software Update where extremely minor updates. The current WebKit build is much further ahead than what we have in 10.4.2's Safari. It's faster, has more compliance (CSS 2 and 3 improvements...passes Acid2 test), fixes a huge amount of memory leaks.



    I don't know when it's going to finally be unleashed...but when it does, a lot of people should be happier.
  • Reply 12 of 13
    kcmackcmac Posts: 1,051member
    I have really been enjoying the Webkit so far. The nightly builds app works great. The only problem I seem to have is that the cool gold style Safari icon is now a generic app icon.
  • Reply 13 of 13
    j@ffaj@ffa Posts: 56member
    Quote:

    Originally posted by kcmac

    I have really been enjoying the Webkit so far. The nightly builds app works great. The only problem I seem to have is that the cool gold style Safari icon is now a generic app icon.



    Are you accessing the WebKit app directly? If you are there's no need, just use Safari! If you are using Safari, then try getting info on it in the Finder, clicking on its icon in the Get Info window, and pressing the delete key. That *might* work!
Sign In or Register to comment.