Post a Comment
Actually Firefox 1.x and 2.x does not support widgets on Linux for rendering of web-content. Firefox always use the ugly win95-like buttons - no matter which theme you use.
What can be changed is the look of non-website widgets, e.g. Firefox preferences dialogue, download window, the look of tabs and stuff like that. But comboboxes, buttons, text fields and stuff like that are not themed for page content.
Firefox 3 however has support for native look for widgets in content as well. Place a submit-button on your webpage and it will have the look buttons in your gtk-theme. And that at least is a bit cool. Next problem is the memory leaks and out-of-process loading of plug-ins (read flash and java).
I beg to differ, as I said, when on Linux I install new widgets.
http://ubuntuforums.org/showthread.php?t=369596
It's an ubuntu-specific solution and has nothing to do with Firefox or widgets as such.
That solution still doesn't follow the native look, but merely allows for some ugly hacking of the look. Of course it might be better than nothing else, but the solution still doesn't make the widgets follow the native look. And the solution is not a Firefox-solution but merely an ubuntu-solution, since Firefox has no such thing as a widget installer.
The moment such widget-'themes' can be installed in the Firefox 'Extensions'-dialogue we might talk about 'Firefox Widgets'. Until then I'm looking forward to Firefox automagically theming the widgets according to the GTK+-theme in use.
As a matter of fact, I'm not all that keen on native widgets in web pages. As far as style is concerned, there's no telling whether they will match the look and feel of the page. And I do hope they're still going to allow the web page to override via CSS, in which case native widgets are mostly for nothing. I'm not at all sure this was a feature worth putting effort into. It's pure eye-candy. I'm sure there are others more useful. I'd gladly give up native widgets for less consumed RAM or snappier startup when I have many extensions loaded.
Edited 2007-11-21 18:39
Of course it should still be overridden by CSS. That's for sure. But there are many websites that don't theme widgets at all. Those ought to follow the native theme.
But widgets theme by websites should stay untouched. It looks really ugly otherwise (konqueror is a sinner here - it uses the colors from css on widgets from the KDE-theme in use - doesn't fit when playing online RPG
)
Nice to see censorship is live and well, under 5 minutes a point is taken off my score; interesting to see the maturity of some around here is taken to all new pathetic lows. When in doubt, and too lazy to debate, remove a few points to silence the critic(s).
Edited 2007-11-20 19:43
RE[3]: Native widgets?
If the memory thing is out of date, reply to that effect. An outdated comment is not off-topic, it's out-of-date. You should reply indicating current information. If you can make a point, do so.
Then again, maybe the comment isn't out-of-date. Sounds like meianoite and others have seen reports of continued issues with memory usage. In fact, you imply that work in this area is ongoing, so you can't even maintain the assertion that the issue is out of date for 20 words.
200+ fixed memory leaks is a great thing, and the people who identifed, fixed, and verified these leak fixes should all be proud of their work. However, in the final analysis, a lot of people won't care if leaks have been fixed if the product still uses too much memory.
Think of it this way, would you care if 200+ safety problems had been fixed in the new model year car you bought if the tires still came off at speeds over 62mph/100kph? Is it a good thing that the 200+ safety problems had been fixed? Is the car maybe safe for folks that never have to go very fast? Yes and yes. Is it time to declare comments regarding the car's safety problems out-of-date? Maybe you would think so, but many others would disagree.
How about you try not being a jerk? If it helps, here's a post you might have made instead of silently modding kaiwai down:
RE[2]: Native widgets?
If you look, you'll see that there have been 200+ memory leaks fixed, so there should be some reduction in memory usage already. There are some bigger issues, mainly in memory fragmentation, which [we are | the team is] currently working on. It's hard to say now when this work will bear fruit, but you and other users concerned about Firefox's memory usage should know this is on [our | the team's] radar and that it is being worked on.
See, this way, instead of looking like a total anus, you get your point across. As an added bonus, you make [your | the Firefox] team look like what they are, which is a conscientious group of developers, testers, and others working hard to produce the best damn browser on the planet.
wtf? kaiwai is right about Firefox having memory problems. It still has those issues. It has many memory problems solved, but there are still many left to solve.
Personally I prefer webkit-gtk (the 'official' webkit.org port) a lot more and luckily it's been picking up speed recently.
Kaiwai:
Hear, hear.
Cyclops:
It was definitely not off-topic, given the subject of Firefox 3 nearing golden and still hogging memory as always. Bullet points on a release notes document won't change the fact that FF3b as it stands is hardly an improvement over FF2 in memory management:
http://www.techcrunch.com/2007/11/20/firefox-3-beta-1-the-memory-us...
Given that, my opinion is that despite believing otherwise, you had zero reason to mod Kaiwai down, period. And pretending FF3 is the vessel of the divine blood won't change the fact that modding him down for his criticism is akin to sweeping bugs under the mat: people will trip over them. Specially people with constrained hardware resources. Specially 3rd world governments trying to jump into the eeePC/OLPC/Classmate/whatevercomesnext bandwagon of cheap internet-oriented computers.
RE[5]: Native widgets?
According to zdnet memory management is indeed improving (notice that FF3b1 actually beat IE7 by using less memory in the 12 page test).
http://blogs.zdnet.com/hardware/?p=960
Does this now mean you are going to start whining and carrying on about IE7 in the same manner?
If you haven't seen this blog entry, its worth reading. A great description (illustrated!) of why Firefox ends up hogging so much memory-- fairly bad heap fragmentation, and what they are now doing about it.
Great blog comments too, really nice to see dtrace and other analysis tools to benchmark and debug Mozilla. Plenty of useful solutions offered, and the blogger is currently testing OpenBSD's malloc with positive results.
http://blog.pavlov.net/2007/11/10/memory-fragmentation/
I hoped that bug 382267 would be implemented, but looks that is has been left for abstract future (target milestone & priority missing, etc).
https://bugzilla.mozilla.org/show_bug.cgi?id=382267
The issue is about rendering Ogg Theora+Vorbis video directly in browser and that is something that would definitely set new standards. As far it is known also Opera is looking for quick implementation of the same feature.
Those who are interested can see a couple of months old demo here (a web-page implemented in SVG+Theora):
http://www.bluishcoder.co.nz/video_svg_demo.ogg
Sure I could, but I don't think there's anything "elite" about using x86_64 nowadays.
It's most probably the second widespread architecture out there and most open source do already ship prebuilt binaries, so I consider it a bit strange for one of the best-known projects not to offer builds at all.
And no, I'm using Arch64 here.
Its definitely a better browser, and more importantly more responsive.
The two things I haven't liked, although it wouldn't surprise me is the move from scaling the tabs to resizing them, although this is well implemented. The other thing I dislike is that when I open tabs from my bookmarks...it adds them to what is already open in my browser window. Although I'm not sure if I will grow to like the behavior.
I'm busting for this to be released its been too long since a real release.
Yeah... usually, when there's a lot of time between releases, that means the product doesn't have many bugs; most are satisfied with the product, and no new features/fixes need to be added for the product to make the consumer happy.
However, I must say, that's not the case with FF right now. It needs a _lot_ of work, mostly on the resource-usage end. Plus, it's rendering engine is rather slow compared to others, IMO.
Hope FF3 fixes this.
Is Firefox getting better or worse? :O
Phoenix/Firebird/Firefox was special because some guys took the Mozilla Suite and cut it down into JUST a good browser. The Mozilla code has always been a mess, but their goal was to make Firefox fast and good at browsing, and they succeded. Over time though... Firefox seems to be regressing.
[As an aside, I think the commercialization of the Mozilla Foundation (which is now explicitly focused on Firefox rather than Mozilla in general, thus making it more of a Firefox Foundation) might be blamed for both the good and bad advances in Firefox.]
In any case I think we might find that the future has a lot more Webkit in it. A Webkit browser in 2010 could be the Firefox of 2003.
Edited 2007-11-21 02:57 UTC
The font rendering in OSX is expected to change:
http://www.hicksdesign.co.uk/journal/cairo-beats-safari
In summary, gecko used quickdraw for text rendering. The new gecko version uses Cairo with CoreImage, which results in a still different, but a lot cleaner (compared to quickdraw at least), text rendering.
Something that I also noticed is that text rendering used to break when using Calibri (all you see is underscores), but the new gecko handles it fine. I have since switched to Myriad as my sans-serif font but its something nice to know. Too bad Cleartype destroys Myriad with anything less than 16 points.
I suppose I'm coming late to the party. I don't know which are the problems between cyclops and kaiwai if any at all, but I think that both of them have wasted rather enough posting space on this topic.
About this Firefox BETA: I am using it both under GNU/Linux and under Windows XP. I have not suffered the memory problem other posters are relating, so I suppose it depends on the configuration or something else.
What I consider important to remark is the following: this is still a beta (in the web it is said to be for developers), and honestly, when the final release is out, I bet it will be a good product. As any other software in the world, it will have bugs, of course, but despite all of its problems, I still prefer it (even the beta, and that's a good point for it) to another browsers (such as Opera, Seamonkey, Internet Explorer, Konqueror, Lynx, Links or Safari).
Kudos to the developers. I am not asking you for the impossible (bug free software), I just want to thank you for your efforts.
I almost forget this one: please, put the configuration options (when possible) in the same place. In Windows they are at Tools -> Options and in GNU/Linux they are at Edit -> Preferences (Netscape inheritance). If we want Windows users to switch to GNUL/Linux, the same program under GNU/Linux should behave as close as the Windows one. Not that this is very important, but I'd like to see it done.
To anyone who might be interested theres Minefield available for download at the following link. And as for the bug that caused my machine to crawl: I haven't had any such issues anymore. Seems quite stable and usable.
http://ftp.mozilla.org/pub/mozilla.org/firefox/nightly/latest-trunk...










