On Wed, Sep 3, 2008 at 11:43 PM, Nathan Cerny [email protected] wrote:

While that was interesting, most of the points were moot. Yes, almost everything in chrome is something someone else has already came up with. But chrome took all the good points from all these other browsers and created a browser that implements them all well.

Well... From a minimalistic, stripped-down perspective it's really nice. However the lack of extensions and add-ons is a HUGE shortcoming. Yes, for the novice user that's not that big a loss, but most of us Linux users are power users, no?

The bookmark system is also primitive as well. Firefox3's is really slick (use the tags, Luke). Firefox3's awesome bar is pretty comparable to chrome's.

And this site seems to imply that chrome should be running far more

resource-dependant than the other browsers out there - I have been running it since it's release yesterday, and I am very impressed. It outperforms IE7 in all aspects...and IE7 outperforms firefox 3 in all aspects (I was shocked too when I realized it). Of course, this is all on my computer at work, because I run only linux at home ;)

You're kidding me, right? When you say IE7 "outperforms", could you be more specific? Have you looked at your memory utilization? Open up more than one tab, please.

Don't get me wrong, Chrome's neat. I can see why Google is doing this, and it will up the ante in the browser wars. That's good. From a UI perspective it's not revolutionary. The backend design is really nice, and ought to lead to innovation on the Mozilla and Microsoft side. The fact that they're using WebKit is interesting as well.

However it's obviously still beta, and it's obviously still not complete. I'm interested in seeing where they go with it, but I hardly think it's worth getting into a fanboy love-fest about it yet.

Jeffrey.

P.S. I'm using Chrome to type this.

On Thu, Sep 4, 2008 at 2:43 AM, Billy Crook [email protected] wrote:

I thought (assuming their speed improvements were real) the JavaScript engine they wrote will probably be useful to other browsers if nothing else. How long will it be until they start using our browser history to target ads... uh, I mean sponsored links!!!

I tested Chrome on a single core 1.8GHz laptop and the V8 engine ran the Google benchmarks at 1200, while FF3 on the same machine ran it at 11-120. At least on the benchmarks they made an order of magnitude jump. Nifty feature-wise I like he lack of status bar by using fade-in tool-tip style balloons at the bottom for link destinations. That saves several pixels of height, which helps the widescreen crowd, esp. thge netbook segment.

Jon.

On 2008-09-04, Justin Dugger [email protected] wrote:

Apparently the Google benchmarks are heavy on recursion, which is a complicated concept and the bane of most programmers.

Whoa, Whoa, Whoa. Hold up. You mean to say you don't implement everything as a single, recursive function!?!?

On Thu, Sep 4, 2008 at 1:17 PM, Luke -Jr [email protected] wrote:

How exactly is recursion in the slightest bit complicated?

Recursion is often analogous to introspection, against which there are strong taboos among certain people who find recursion objectionable.

On Thu, Sep 4, 2008 at 1:17 PM, Luke -Jr [email protected] wrote:

On Thursday 04 September 2008 07:50:24 Justin Dugger wrote:

...

Apparently the Google benchmarks are heavy on recursion, which is a complicated concept and the bane of most programmers.

How exactly is recursion in the slightest bit complicated?

Spoken like someone who's never had to explain it to a novice programmer, or implement it in a compiler. A number of people do find it complicated in practice and avoid it like the plague. I know plenty of people who hate Lisp and Ocaml because one is almost forced to use recursion.

And if that doesn't satisfy you, recursion is complicated because you don't know when the compiler effectively implements tail call recursion without reading the code or analyzing memory usage.

Justin Dugger

Plug-ins were brought up by someone else already. I have to say that on Windows PCs and laptops under my control, Firefox will stay as number one browser for a while due to the plug-ins. I have Adblock Plus and Adblock Filterset.G to block a lot of cruft. Flashblock to stop the blink tags that pass for ads these days. Session Manager so that when I shutdown the browser I can resume with all the tabs I had when I last used it and it helps, but doesn't completely solve the losses from a browser crash. IE tab if something doesn't render as expected in Firefox (I barely use this one anymore).

Firefox 2 and 3 will stay prime for me, butI will keep an eye on Chrome and I hope Mozilla and others do too.

Brian Kelsay

On Thu, Sep 4, 2008 at 2:43 AM, Billy Crook [email protected] wrote:

The one process per tab approach is nice. I just wish the process would be called with the url as the arguments somehow, so you could watch sites by URL in top.

That isn't possible unless you want to do an exec call every time the user goes to a new page. That's why Chrome does its own task manager that knows what each thread is doing.

On Thu, Sep 4, 2008 at 9:00 PM, Hal Duston [email protected] wrote:

...

That isn't possible unless you want to do an exec call every time the

user

...

goes to a new page. That's why Chrome does its own task manager that

knows

...

what each thread is doing.

Postgresql and sendmail manage it.

root 5082 1 0 Jul27 ? 00:52:53 sendmail: accepting connections smmsp 5091 1 0 Jul27 ? 00:00:12 sendmail: Queue runner@01:00:00 for /var/spool/clientmqueue postgres 16733 16667 0 Aug04 ? 00:01:32 postgres: logger process postgres 16737 16667 0 Aug04 ? 00:06:53 postgres: writer process postgres 16738 16667 0 Aug04 ? 00:05:28 postgres: wal writer process postgres 16739 16667 0 Aug04 ? 00:03:30 postgres: autovacuum launcher process postgres 16740 16667 0 Aug04 ? 00:05:07 postgres: stats collector process

They both manage that by recreating the environment of the already running process.

Yes, they exec to spawn these separate processes. But those processes can't change the arguments under which they were invoked without themselves exec-ing all over again. So when you click on a "Heh.", you'd have to change one process from monster 3952 26665 0 10:08 ? /usr/lib/firefox-4.0/firefox-bin -UILocale en-US instapundit.com to monster 3952 26665 0 10:08 ? /usr/lib/firefox-4.0/firefox-bin -UILocale en-US johndoe.blogspot.com/2008/09/snarky-article-title-here.html

The only way I know to do that is to exec firefox-bin, invoking the new URL as an argument. What I don't know is how much overhead that takes. I suppose a stripped-down binary, that inherits a pre-built environment populated by another binary reading configs, etc., might not be doable.

But that's not the way that any existing browser I know of is written. You're talking about a complete re-factoring of how browsers work. That's not to say that it's a bad idea, just that it's a non-trivial effort, which might make things much slower

--- On Fri, 9/5/08, Christofer C. Bell [email protected] wrote:

On Fri, Sep 5, 2008 at 9:27 AM, Monty J. Harder [email protected] wrote:

...

You're talking about a complete re-factoring of how browsers work. That's not to say that it's a bad idea, just that it's a non-trivial effort, which might make things much slower

I realize you're being polite, but I will say that it's a bad idea. I don't mind if the browser has some window that tells me which PID is looking at which page, but I don't want what I'm looking at to show up in a ps(1) listing. While I'm sure that the "average case" is a single user machine sitting on someone's desk as their workstation, that's not the case for everyone. People in computer labs, offices, and even at home can and do have multiple accounts on their machines that can be accessed by others at any time. I don't think they'd be appreciative of their browser coughing up what they're viewing with a simple ps(1).

So in other words it sucks for employees, but its great for employers who want to know exactly what their employees are doing with their Internet?

Sounds like what you don't like about Chrome could be a selling point for managers and instructors.

--- On Fri, 9/5/08, Christofer C. Bell [email protected] wrote:

On Fri, Sep 5, 2008 at 9:56 AM, Leo Mauler [email protected] wrote:

...

Sounds like what you don't like about Chrome could be a selling point for managers and instructors.

a.) That's what corporate firewalls/proxies are for.

Then you should feel lucky you've never worked in some of the places I've worked, where the employer felt a brief browse of the system logs once a month wasn't enough. ;-)

b.) Chrome doesn't do what's being discussed. It was mentioned as a wishlist item.

I thought the web comic explained that Chrome already had a separate process per-tab, such that one tab going down didn't take down the whole browser? Wouldn't that allow a ps to reveal all of the tabs?

On Fri, Sep 05, 2008 at 09:26:16PM -0500, Christofer C. Bell wrote:

What someone mentioned was a desire to see each tab's process displaying the URL as an argument so they could see which process what looking at what (and monitor the memory, CPU use, etc, of that process). What it actually does is something like this:

2467 chrome 2490 chrome 3186 chrome

Rather than this:

2467 chrome http://www.cnn.com/ 2490 chrome http://www.digg.com/ 3186 chrome http://www.goatse.cx/

The actual behavior may give away how many tabs you have open, but says nothing about what you're looking at. The second example can lead to some embarrassment.

Doesn't sendmail display the address to which it is attempting delivery?

Doesn't that at least have the potential to be an embarrassment?

Thanks, -- Hal

On Fri, Sep 05, 2008 at 09:27:13AM -0500, Monty J. Harder wrote:

Yes, they exec to spawn these separate processes. But those processes can't change the arguments under which they were invoked without themselves exec-ing all over again. So when you click on a "Heh.", you'd have to change one process from monster 3952 26665 0 10:08 ? /usr/lib/firefox-4.0/firefox-bin -UILocale en-US instapundit.com to monster 3952 26665 0 10:08 ? /usr/lib/firefox-4.0/firefox-bin -UILocale en-US johndoe.blogspot.com/2008/09/snarky-article-title-here.html

The only way I know to do that is to exec firefox-bin, invoking the new URL as an argument. What I don't know is how much overhead that takes. I suppose a stripped-down binary, that inherits a pre-built environment populated by another binary reading configs, etc., might not be doable.

But that's not the way that any existing browser I know of is written. You're talking about a complete re-factoring of how browsers work. That's not to say that it's a bad idea, just that it's a non-trivial effort, which might make things much slower

A process can change the arguments at any time w/o doing an exec. I have actually read the code that is needed to do this under Linux. It is less than 15 lines. All it does is rebuild the environment which is where the exec'ed command line arguments are kept. Once they are rebuilt ps, top, and everything else just pick them up from /proc/<pid>/cmdline.

Thanks, -- Hal

On Fri, Sep 5, 2008 at 11:14 AM, Hal Duston [email protected] wrote:

...

A process can change the arguments at any time w/o doing an exec. I have actually read the code that is needed to do this under Linux. It is less than 15 lines. All it does is rebuild the environment which is where the exec'ed command line arguments are kept. Once they are rebuilt ps, top, and everything else just pick them up from /proc/<pid>/cmdline.

Wow. I had no idea. Here I thought ps got that info from an internal kernel data structure that wasn't manipulable by userland. That suggests some really ugly possibilities.

On Fri, Sep 5, 2008 at 8:31 PM, Hal Duston [email protected] wrote:

the environment. All the program is doing is modifying the contents of argv[0] after making provision for the rest of the environment.

I wonder if ps has been evaluated for security. It's old enough that surely someone's thought of attacking scripts like that before. I'm sure there's a few scripts out there written to run 'ps' as root.

Justin

On Fri, Sep 5, 2008 at 19:57, Monty J. Harder [email protected] wrote:

Wow. I had no idea. Here I thought ps got that info from an internal kernel data structure that wasn't manipulable by userland. That suggests some really ugly possibilities. Don't know how 'ugly' it would be (aside from processes making it difficult to tell what their PID was if you weren't watching when they were launched)

Here's some code from Hal Dustin (attached). You can demonstrate on your own computer that it is possible for a process to change its process name in ps without having to terminate, and create a new process. So chrome could easily do this. (If it's what the user wanted.) On computers with a single user, I think it could be rather handy to have at least an option for this. Wouldn't it be neat to look, and see 'oh, *that* site has used a good hour of CPU time!'. If it were so trivial to see resource utilization per-website, it would put pressure on web developers to be more efficient, and that is a win for users and browser developers. Side note: It appears gnome's 'System Monitor' still displays the file name.

Download it, then compile it with: [bcrook@One Desktop]$ gcc time.c -o time

Run it and get a shell back: [bcrook@One Desktop]$ ./time & [1] 15146

Then using the PID number bash just gave you, grep ps' output for it, and watch the PID stay the same, but the title (usually the binary's name) change: [bcrook@One Desktop]$ while true; do ps aux | grep 15146 | grep -v grep ; sleep 2; done bcrook 15146 0.0 0.0 3924 516 pts/2 S 22:07 0:00 You would see my binary's name here, but instead, I'll tell you the time is: 2008-09-05 22:08:27 bcrook 15146 0.0 0.0 3924 516 pts/2 S 22:07 0:00 You would see my binary's name here, but instead, I'll tell you the time is: 2008-09-05 22:08:29 bcrook 15146 0.0 0.0 3924 516 pts/2 S 22:07 0:00 You would see my binary's name here, but instead, I'll tell you the time is: 2008-09-05 22:08:31 bcrook 15146 0.0 0.0 3924 516 pts/2 S 22:07 0:00 You would see my binary's name here, but instead, I'll tell you the time is: 2008-09-05 22:08:33 bcrook 15146 0.0 0.0 3924 516 pts/2 S 22:07 0:00 You would see my binary's name here, but instead, I'll tell you the time is: 2008-09-05 22:08:35 ^C

[Ctrl]+[c] when done. Then kill ./time by its PID [bcrook@One Desktop]$ kill 15146

And this site seems to imply that chrome should be running far more

...

resource-dependant than the other browsers out there - I have been running it since it's release yesterday, and I am very impressed. It outperforms IE7 in all aspects...and IE7 outperforms firefox 3 in all aspects (I was shocked too when I realized it). Of course, this is all on my computer at work, because I run only linux at home ;)

You're kidding me, right? When you say IE7 "outperforms", could you be more specific? Have you looked at your memory utilization? Open up more than one tab, please.

My work computer is a fresh install of Windows Vista (it's 2 weeks old today). So I'm running on an IE7 profile/Firefox profile that is 2 weeks old. And a chrome profile that's a day and a half old. With a single tab open, IE7 is in the 60MB range of memory usage. Under my normal load, it's closer to 140MB. Firefox was at over 120MB (with multiple tabs open...since I've since uninstalled firefox I can't give you better numbers than that). Chrome was sitting at 25MB with 2 tabs open. Now in Linux it's completely different - I swear by firefox.

But, numbers aside...when I say IE7 outperforms Firefox 3, I'm referring to overall feel and speed. The same site would load faster in IE7 (or so it felt due to the feel of the browser). I have browser windows open for weeks at a time usually - Firefox's footprint would continue to grow and get to the point where my computer was unusable until I closed it. IE7 doesn't have this same type of leak (or it's much smaller). I haven't had chrome installed long enough to know if it has any major memory leaks or not...I would assume so, but hope not. But so far I really like the lightweight feel to it.

Then again, I've always been a fan of fluxbox ;) Unfortunately I run KDE, but that's just because I'm too lazy to figure out how to get everything to work well in fluxbox...and KDE has apps for everything I want :)

Actually, I upgraded to firefox 3 not the day it was released, but the day after ;) But yes, I had firefox 2 before that... And my experiences may be completely due to the type of sites that I go to at work. I was just giving my 2 cents about my personal experiences :)

On Thu, Sep 4, 2008 at 1:08 PM, Jeffrey Watts [email protected]wrote:

On Thu, Sep 4, 2008 at 8:08 AM, Nathan Cerny [email protected] wrote:

...

...

My work computer is a fresh install of Windows Vista (it's 2 weeks old

today). So I'm running on an IE7 profile/Firefox profile that is 2 weeks old. And a chrome profile that's a day and a half old. With a single tab open, IE7 is in the 60MB range of memory usage. Under my normal load, it's closer to 140MB. Firefox was at over 120MB (with multiple tabs open...since I've since uninstalled firefox I can't give you better numbers than that). Chrome was sitting at 25MB with 2 tabs open. Now in Linux it's completely different - I swear by firefox.

Well, in Linux you don't have many other options. ;-)

...

But, numbers aside...when I say IE7 outperforms Firefox 3, I'm referring to overall feel and speed. The same site would load faster in IE7 (or so it felt due to the feel of the browser). I have browser windows open for weeks at a time usually - Firefox's footprint would continue to grow and get to the point where my computer was unusable until I closed it. IE7 doesn't have this same type of leak (or it's much smaller). I haven't had chrome installed long enough to know if it has any major memory leaks or not...I would assume so, but hope not. But so far I really like the lightweight feel to it.

Well, every benchmark I've seen says that your experience is anecdotal. I'm also wondering if you're talking about Firefox _2_, and not 3. You've said "weeks at a time", and Firefox 3 hasn't been out very long.

Firefox 2 would bloat up, memory-wise, after extensive use. One of Firefox 3's biggest improvements is in the area of speed and memory management.

Jeffrey.

--

"He that would make his own liberty secure must guard even his enemy from oppression; for if he violates this duty he establishes a precedent that will reach to himself." -- Thomas Paine

On 2008-09-04, Kelsay, Brian - Kansas City, MO [email protected] wrote:

Google Chrome should not let one tab w/ a problem bring the whole browser crashing down.

When I was looking at it, one of the first things I tried was loading up Google reader with a big old flash applet embedded in an article, and when i found the process that was powering that monstrosity, and killed it, the tab got its cute little sad face icon like the cartoon said it would. The rest of the browser kept on trucking. I think they should try to auto-reload the page at least once when it crashes, or provide a link to reload it.