M11 is out!Tuesday November 16th, 1999M11 binaries are now available on the mozilla.org ftp server in Win32, Linux, and Mac flavors, with more to come. Things new in this release:
This is great! I haven't tried chatzilla yet so lookin' forward to it. Here's a hint to folks who haven't ran mozilla for a while-- delete the prefs-type files (like for me, the Mozilla Registry file in the Mac Preferences folder needed to be deleted. Once I restarted and created a new profile-- it fixed about a hundred "bugs")... Oh... and.. first post! (sorry, always wanted to do that.) W dloading right now Well a few followups. 1. I'm using it now. It's awesome, though there's a small issue w/the default page (the release notes) being slow to display, so the effect is like it freezes for a while. I'm guessing this is due to the incremental reflow not being fine-tuned for my DSL connection. I checked in with #mozilla and others were having this problem too. It looks like this is being addressed and I'll bet by tomorrow it'll be fixed. 2. Chatzilla isn't in the build for mac. Again speaking with the folks at #mozilla, they said it was some kind of a networking issue and asked if any mac developers could help get the make stuff working... 3. the other pages look great! I'm writing this with mozilla M11 now and no problems thusfar! W Just curious, and being an idiot outsider, why are the mac builds and milestones megabytes bigger than the ones other other platforms? Thanx <:3)~~ I think it is simply the hexing of the binary stuffit file. If you look at the .bin file after it decodes the .hqx I don't think it is much larger than the others. Foamy I believe most macs use RISC processors, so their binaries are going to be bigger (fewer instructions in the processor == more in the executable). Yes, the PowerPC processors are RISC processors, and I have noticed that a Mac version of a program will be a tad bigger than the PC version, but we're talking about a 4 meg difference here, which isn't a tad. This is coming from a guy with crude understanding but doesn't RISC reduce the instructions required on the processor by using a base amount of instructions, and the rest is created on the fly by combining the base instructions, so it's still all hardware. Well I read this in an article years ago and not sure if I remember it correctly or if it might be false. <:3)~~ Is it only me or do others also see incremental reflow as a pain? Everything jumps wildly in the page and sites like slashdot are really slow when the browser updates all the gazillion tables. Vote for bug #17325 at http://bugzilla.mozilla.org/show_bug.cgi?id=17325 Bugzilla has some interesting comments on the reflow bug (follow the link above). I'm experiencing the same thing on my own M11-win32 copy, and it sounds to me like reducing the number of reflow calls overall will speed up page loading. But personally I think the reflow idea is wonderful, even if it is at the moment imposing a high-speed penalty. I love the chrome, and I really love the layout engine. There's still lots of work to be done on them, but don't take them out! Implementing reflow fixes the one biggest annoyance in Navigator: not seeing any table content until the entire table and all images without height/width tags are loaded first. Now I have no reason at all to use IE! -Sailor V I feel the same, I like the incremental reflow concept but it's major penality now is speed. I like to watch the page rearrange itself as it loads and I can still easily read the content of the page as it loads but the speed has to and will improve. <:3)~~ I like don't like to watch the page rearrange itself as it loads. Never liked in IE and now will have to dislike it in NC. finally i'm wondering if this incremental reflow is really a Good Thing (tm) ? for me too, it makes it hard to begin to read the text on the loading page, because everything shifts up and down, left and right... it's mostly visible on pages with text in tables. it's a real problem, and i can't see any straightforward solution. -- Hervé It seems to me that much of the problem comes from horizontal jumping, rather than vertical. So a simple compromise is to not display anything on a TR until everything in the TR has been loaded. That way you don't have to wait for the entire table to load, but you avoid the jumpiness. Just my $0.02. --John Keiser The problem with that is the sites that use a two-cell table where one of those cells has a huge embedded table, allowing the other to act as a sidebar. Witness /. and segfault.org. Even in those cases, if you wait for the outer </tr> tag, it won't do any good because the </table> tag follows right after that, and waiting for any inner </tr> tags will still result in some jumping as the new embedded tables in adjacent cells are loaded. It would still be interesting to see, though. Perhaps a solution could be reached by a side-by-side comparison of the two most proposed compromises: a speed-factored delay like the one described in the comments for Bug #17325 http://bugzilla.mozilla.org/show_bug.cgi?id=17325 , and the wait-for-TR proposal here. If such a comparison is possible, they might provide some enlightening results. They may be two cents, but some pennies are made of gold. -Sailor V The fix is simple - don't reflow too often. This is best done dynamically, i.e. time every reflow and wait at least 5 times that time before the next attempt. Unless loading finishes, of course. This ensures that a max of 20% of the cpu is spent on reflowing. Improving the overall efficiency of the reflow process itself will help, which they seem to be doing also since I've noticed a huge improvement between the build where incremental reflow was first introduced to yesterdays build. But there's still more room to fix and they'll get to it. <:3)~~ gerbil What is the Build ID. Thanks. I _really_ hope they get the speed fixed. This browser is now more stable that Netscape 4.7 on my machine (woohoo!), but sadly it's orders of magnitudes slower, both in the user interface (where the slow speed sometimes means that the wrong part of the UI gets messages - e.g. menus instead of text boxes getting keypresses) and in rendering speeds. It's a shame because people are going to see this and say that the new layout engine just isn't fast at all. Fortunately, the same people who would say that are likely not the ones testing M11. ;) I disagree. I think around now there are lots of different web developers testing M11 now, ranging from Windows users to Linux users, IE users, Opera users. All of them, including me a Netscape 4.7 (Linux) user, are going to feel that this is a slow browser right now. I know they are working on it (seeing threads in Bugzilla on it already), but there were huge headlines about how necko was supposed to be a lightning fast renderer - and they're not true. First, the rendering engine is Gecko (NGLayout) and it _is_ lightning fast. Necko is the network library. Gecko is a very light-weight and fast rendering engine but the rest of mozilla is still being written and optimized to take advantage of Gecko. Mozilla uses Gecko to display more than just content. Gecko also renders the UI and this is where top notch engineers are working hard to make changes that will speed things up. I think it's a safe bet that they're up to the task. i believe that they will find a solution, it must be only a tuning problem, but... anyway, the visual effect of seeing things "dancing" on the screen as they are reflowed is a bit devastating, don't you think ? i'm afraid that even if it's a top notch algorithm to render the page gradually, it's maybe too uncomfortable on the visual point of view. just wondering. -- Hervé 2 points: IE does something similar - load a page that has images w/o height/width tags, and the page reflows as the images are downloaded. The implementation in mozilla is not optimal - *everyone knows this* especially the ones who wrote it. Right now, it reflows the page after a certain number of tags are encountered (or something similar to that), which causes a lot of flashing on fast connections. They plan on moving to a timer based mechanism which will 'do the right thing'. Also, THERE IS NO CACHE, ever since Necko landed. This will be coming online in the next week or so; Intel is working on the cache implementation. Well, while Gecko seems to render quite quickly, the fact that Gecko is rendering EVERYTHING from the page to the UI, makes it very slow. While crossplatform development is now quite easy, I'm betting mozilla will be eclipsed by faster browsers w/ native UIs on each platform (i've started using kfm/w3m! ugh!) Make no mistake, this is not a slam on mozilla, I've watching mozilla's progress carefully (downloading daily builds every day for the last 6 months), and the fact that it is a very well done cross platform project strikes awe in me. But, after hundreds of "i fixed feature X, and now its blazingly fast" causing me to waste another 5 megs of bandwidth, for little noticable gain, has left me a bit disillusioned. I wish the developers the best, but with the ever increasing pressure from netscape/media/everyone to just ship it, I don't believe netscape5 will be the tiny/fast/and memory efficent browser its been hyped as. A sad statement from a once rabid lover of the project (my friends make fun of me for downloading it every day... its like an obsessed nerd watching a cute girl he'll never have) ohwell, i'm done venting. back to downloading daily builds.. :) for SuSE 6.2 users : don't despair if the JPEGs don't show up with M11... it's because M11 is compiled so that is searches for libjpeg.so.62 (and SuSE 6.2 ships with libjpeg.so.6.0.1). => download this libjpeg.so.62 (e.g. search a ftp site near you with ftpsearch.lycos.com), become root, put it in /usr/lib and ldconfig -v (libjpeg.so.62 => libjpeg.so.62 should appear somewhere) do it now ! it's safe, it wont break your favorite distribution and the jpegs will show up in M11. Hervé Or you can simply compile from source... although for me with a P133, 96MB RAM, that's an overnight chore :) the subject says it all. explicit proxy settings seem to be ignored?? is there an env var or a cmd line arg that will trace the network layer? Proxies are working, but you still have to painfully edit pref.js - see here: http://bugzilla.mozilla.org/show_bug.cgi?id=8559 Isn't it about time that the Preferences Dialog actually let you set the Proxy Server and have it remember it properly ? I don't think I've ever seen a Mozilla release that let you set proxies from the GUI and we're running out of time before the more public alpha/beta releases are out... Also, call me dumb, but how do you get rid of the infuriating "Sidebar" completely ? I hate that taking up 1/4rd of the browsing area when Mozilla starts up. Toggling off "View -> Sidebar" only brings temporary relief from the space-waster - when I next fire up Mozilla, the sidebar is back again....grrr... ... but they don't take affect. i'm looking for confirmation that proxy info set in either a per user pref file or a global pref file actually works. empirically, it doesn't (under linux). The Fullcircle download doesn't seem to be working. I went to slashdot, voted and the browser crashed. I was expecting Fullcircle to popup, but it never happened. :( M11 crashed for the first time on me at Slashdot also, but the FullCircle did pop-up, sure you got the right download? There is a non-Fullcircle version of M11 also. <:3)~~ Was that the ACLU and Echelonwatch page? Tahts crashed M11 here and Netscape didn't like it too much eitehr. Plus the colors are all messed up compared to normal Slashdot pages. Don't blame M11 for the colours, they are normal. Slashdot now has different colour schemes for different sections, and Your Rights Online has a particularly ugly scheme... I've just tried out M11. I hope Mozilla works out, but agh! the frustration of using this for the first time. a) Is there a way to get rid of the bar on the left? I close "My Panel" but there is still a semi-thin bar there. b) Slow, slow, slow interface (like above posts). Even going through all the preference windows (Edit->Preferences) is slow. Very similar to the slow speed when using Java. The text disappears too (goes invisible) for some reason, while clicking around on the Category menu). c) Slow rendering? Someone mentioned it was "lightning fast" up above. Well, I beg to differ. I put up some sample daily TV listings tables at http://members.tripod.com/~test_page which are similar to the ones I view each day. If you click a time to view, I found that IE5 loads and completely displays each page/table in _less than a second_, Netscape 4.7 in approx 2 seconds, and Mozilla M11 in 7 to 8 seconds. Mozilla M11 also peaks the CPU at 100% during these 8 seconds. Argh. That just isn't fun at all. Obviously this will differ based on inet connection/cpu speed, but that is how it works out here. (mind you IE5 doesn't display the lines in the table properly after the first two vertical bars.. odd) d) The File/Edit menubar is white. I hope someone changes that. It blends in with white pages too easily (especially in javascript popups -- click any TV program on that tv listing page) Well, I've only had 1 crash so far (while moving the scrollbar during the loading of www.firingsquad.com) so I guess that is kind of good? The UI stuff I'm sure is easy to change, but I doubt the world will tolerate the lack of speed of the current product. Just my 2 cents. a) Go to View, and select Sidebar. The checkbox will disapear, and so will the sidebar. b, c) it's not even beta yet, so you can't compare it to a released product such as ie or n4.7, though hopefully you'll be able to soon. On large pages such as the one you mentioned, try using a CSS declaration such as TABLE {layout: fixed;} to increase speed (in both mozilla and ie). d) I agree, it's a bit weird to have white menus. get rid of side bar (mac) view menu->sidebar check mark disappears and so does the side bar done Er, "View Menu->Sidebar" toggling is *very* temporary - it's a known bug (registered in Bugzilla) that this doesn't apply when you open a New Window (i.e. it shows the sidebar in new windows, even when the View toggle is set to off). Even worse than this, the sidebar setting isn't saved from what I can tell, so fire up Mozilla again and the dreaded space-wasting sidebar is back again :-( I, for one, hate losing 25% of my browsing area to a useless sidebar - I hope that the default is to have the sidebar *turned off* when Mozilla goes alpha or beta. i know you can set them via the preferences dialog, but do they actually <b>work</b>? under linux? Just edit the following two lines in the all.js file in the prefs directory. pref("network.proxy.http", "some.proxyserver.com"); pref("network.proxy.http_port", 8080); Naturally you can use an IP instead of a dns name. Port can be whatever port your proxy runs on. You can check out the bug report at http://bugzilla.mozilla.org/show_bug.cgi?id=8559 : bug 8559. I would have mozilla closed while editing this file and then start it up after saving the file. I don't know why everyone sounds so surprised and disappointed in M11's performance. I said many months ago that XUL was a critical mistake. Oh well. XUL is what makes the difference for me. I like the idea of it from a development standpoint as well as it is what makes Mozilla crossplatform. Take that away and I will just stick with Internet Explorer. I think XUL will change the web as we know it. Stop being so pessimistic, Bruce. The poor performance is due to the lack of tuning of the incremental reflow to this point. It'll be fixed. Vidur's working on it. There's got to be another reason for its slowlyness. It takes a minute to render a local file that's only 739K in size. Perhaps network code has a bunch of sleep statements :) The file has no tables -- NC bookmark file format. The local files are trying to reflow incrementally, too. You know the slowdown some people get on DSL and faster connections? Hard drives are much faster, and thus display much more slowly the way the engine is right now. Vidur's working on it; have no fear. :) Perhaps we could have a prefs setting added to disable incremental reflow? This would give us a better picture of the real culprit and a workaround until Vidur has finished tweaking it... Perhaps we could have a prefs setting added to disable incremental reflow? This would give us a better picture of the real culprit and a workaround until Vidur has finished tweaking it... There has been a bug report asking for a preference to turn incremental reflow off, and I forgot the bug number, but it's label as WON'T FIX because they don't want to polute preferences with too many options. They do have a point, in keep preferences simple, but I like lots of options and I know many people who do also, and an advanced preferences section would be VERY VERY good for us. <:3)~~ You sure that's the reason? I would have thought it was because they wanted to make sure the incremental reflow code could be beaten, bloodied, and tortured by the users so it could improve to the point where no one would need to disable it. :) Actually, before the new skin and a lot more features got added into Mozilla zul was pretty damn fast. And its not like my computer is a speed demon. Its just a old Cyrix 6x86MX PR200 underclocked to PR166, so Windows doesn't crash. When the projects reaches alpha maybe even beta then we should see major speed increases. right now there's just too much stuff being added at once. I've been a big supporter of Mozilla since the begining, telling all my friends how it was the next big thing, but I gotta say as Mozilla approaches beta...they really got to speed this thing up. I got two major complaints 1) The damn thing renders too slow! 2) I hate how when it is loading a page, the user interface becomes useless. Like, just now I was loading the poll page on mozillazine, and I couldn't hit stop becuase it was still loading. What, exactly, is the point of a stop button that doesn't work while loading pages? I wonder...does anyone know how, if they will speed this up, or if they even plan to address it? If not...I'll probably keep using 4.7, sad as that sounds. Standards are nice...but people want somethign fast and responsive...noone cares if their car has low emissions, they car if their car can go 0 to 60 in 4 seconds. I hope the people at mozilla realize this. If you read some of the messages above, yes, the mozilla people (sorry for the lack of a better name) are more than aware of the rendering speed and are working on it. M11 is obvioulsy isn't as big of a leap as M10 was over M9 in performance, but the overall stability has increased greatly. Here is a XUL hack to get rid of the sidebar. There were several complaints this is an easy fix. Gotta love XUL! http://geeks.forchrist.org/xul/ In the chrome/navigator/content/default directory there is a file called navigator.xul. Simply back that up and replace with the one listed on the link above. Restart Mozilla and voila no sidebar! Personally I like it, but everyone is entitled to their own opinions. That is what makes OpenSource so wonderful. er... isn't it enough to just do "View -> Sidebar" and hide it ?? maybe i miss the point. regards. -- Hervé Herve, I agree. I simply read it as a literal that they wanted the thing GONE as in permanently. : ) In some builds it would come back on restart even if you had it hidden. I just did a check and it seems to be saving it correctly. So the only need for this hack would be if you permanently wanted to kill the sucker or if you are experiencing issues with it coming back for some reason. And I have got to quit taking things so literally. :) Small annoucement: I've hacked-up a M11 binary for Sun Solaris 2.7 SPARC, available from http://puck.informatik.med.uni-giessen.de/download/mozilla-sparc-sun-solaris2.7-M11.tar.gz (here). Time for me to shift gears for a minute and address this senseless whining regarding the performance of a pre-beta software application. Optimization is one of the *last* items on the agenda during the pre-final release development cycle so it's pointless to compare Mozilla's performance in this stage of the game with IE, Netscape, Opera, etc... Just because I'm not thrilled about XUL doesn't mean I don't have the highest confidence in the rendering engine itself. Thanks to Adam Lock, the XUL debacle is something I don't have to live with. Hi Mark! hehe! :)
Who knows, Bruce? Maybe they would take a native wrapper, if you opened the code (yes, yes, I know you're working on ROM II, but you could always GPL the NavX code :) ). In any case, XUL may yet surprise you. :) >Just because I'm not thrilled >about XUL doesn't mean I don't >have the highest confidence in the >rendering engine itself. If you can have confidence in the rendering engine itself, why not XUL? After all most enhancement done for the rendering engine are done for XUL as well. How would I use ChatZilla to connect to a server such as oz.org? I wanted to try M11 on the Mac but all I can get is the spash screen. (Mac OS 9) I deleted the Mozilla registry and the mozilla folder in the douments folder yet it won't get past the splash. I just moved the Mozilla folder to the trash and didn't delete it. Does this make a difference? Registry was deleted. My mouse cursor will move but I can't force quit or restart via software. Any ideas? Mike S. I tried M11 on an older G3 running MacOS 8.5 at my work and the slash screen was there for a LONG time before I got anything, but it worked. I went to get a drink and when I came back it was still at the splash. Just how long is long? I remember one of the older milestones took a while to launch, a minute or so, but never this long. I don't have a fast Mac (225 603e) but this is nuts. How long should I give it? Thanks, Mike S. I don't remember, sowie, I was so bored at the time I couldn't pay attention to the time except that it took a long time. And let me correct myself, I was using a PowerComputing PowerTower Pro 200. There are some older G3s in the room but I just remember that I was on one of the olde PowerComputing systems. Well try letting it run for a couple minutes and if it still takes long try filling a bug since MacOS 9 is still very new. <:3)~~ I get similar symptoms trying to run M11 on my iMac with System 8.5.1. Splash screen and nothing else, though I can force quit or switch to Finder while Mozilla is just sitting there doing nothing. If you solve the issue Mike, I'd sure like to know about it. Tim The sluggish UI has nothing to do with XUL. It's happening when the page layout is taking up too much time and not allowing native events to be processed. The fact that you blame XUL just shows what a moron you actually are. I'm not blaming XUL for the performance of the rendering engine in M11. XUL will always be sluggish relative to a native UI regardless of the optimizations. And, yes, part of the UI problems are directly related to XUL. If the UI was native then these UI performance problems wouldn't exist. I find it hard to resist calling you a moron... but I won't because it's childish and I would just be lowering myself to your level. . . . Ah... what the hell... you're a MORON!!! "XUL will always be sluggish relative to a native UI regardless of the optimizations." GTK is drawn with a very similar system to gfx, it and XUL could end up being identical speed Yes, XUL may always be slower than native because it is interpreted itself. /But/ I think the original poster was saying that it was /much/ slower only because the main layout was eating cycles from the XUL layout. If the main layout is doing this, then XUL will more even slower than it is supposed to be. I wanted to try M11 on the Mac but all I can get is the spash screen. (Mac OS 9) I deleted the Mozilla registry and the mozilla folder in the douments folder yet it won't get past the splash. I just moved the Mozilla folder to the trash and didn't delete it. Does this make a difference? Registry was deleted. My mouse cursor will move but I can't force quit or restart via software. Any ideas? Mike S. I seem to have a different experience than many of you, since for me R11 is fast. I tried earlier versions, but got tired fast because of the extreme instability and UI slownes. This version seems quite stable and is actually quite fast. Very impressive for software that hasn't even reached alpha level. Indeed it is. It's really sad that a pre-alpha release of Mozilla would be overall faster than the current market releases of other browsers. The only one I've yet used that's faster than M11 is Lynx. M11 is still buggier than a campfire dinner, but then so is any pre-alpha release. I would help with the coding, but I'm still learning C++. ;-) Keep going, Moz! For what it's worth, it's working for me and my mac (OS 9, pb bronze g3). the first time it's a little longer on startup (15 secons maybe) but then subsequent startups the splash screen is on for maybe 7 secons and then maybe 3 seconds of nothing, then the main window comes up. Dunno what to suggest Mike S-- You have enough memory to run it? (I have 192M...) W Some people at another message board have it working on 9 as well. I don't know what to say either :) I tried again and got the same result. I figure maybe I'll try an M12 nightly. Mike S. I'm using a nightly to post this message and they're nice, faster than the milestones. I download a new one everyday. <:3)~~ gerbilpower I download the latest nightly, 19th, and again it just hung. I tried disabling libmoto. Hang. I restarted with the Mac OS 9 base extention set with only my Dayna PCI ethernet card extention. Hang. Either there's something about this 6500 from a hardware perspective it's not liking or the Dayna extention is the problem. I could disable it and find out but I just hate intentionaly freezing my system. Debuging's a B*tch and I can't wait for OS X and it's protected memory. Mike S. to all who thinks mozilla is a tad slow. Just wait until they disable all the debug stuff and start using optimized builds I love the new look, but bookmark text is too small.. Like the mail thing and how that works too. -- Dan :-) What are you pathetic winers running on out there, A 486-66? Get with the program. On this PII-400, it runs great. Good god, top end machines are plowing the field at 733mhz with over 1 gigahertz AMD K-7's due out by the time this thing goes live. And it will only get faster after they remove the debug code and do some performance tweaking. I've been a developer for a long time and believe you me, you don't start a new large project with performance as the driving factor. You build a good stable, modular foundation first.... then worry about speeding things up. heh. although none of that muscle is really necessary. I'm doing XPApps development for Mozilla and I have a K6/166 with a meagre 64Mb RAM ;) I find it a little slow but that's mostly due to rough increflow and areas that haven't been tweaked yet. I fully expect Mozilla to run on any machine capable of running Windows 95 properly (and compatible across platforms) by release. Anybody else having resource leaks under Win32? I'm having some really terrible leaks. After browsing CNN and Slashdot and MozillaZine, my Win95 OSR2 system (with 128MB RAM) is being brought to its knees. Anybody else having this trouble? Its always had memory leaks. They are working on it. It has gone down by huge amounts in the past few weeks. It used to be much worse. No resource leaks of any significance to report here. My setup is a Compaq Presario 1690 notebook: K6-II/400MHz w/64MB RAM and Win98 (pity me). The only leaks I can detect are the ones reported by the Mozilla console window, and that reports two or three webshells leaked on occasion. As for bringing the system to its knees, my performance is staying steady even after heavy surfing with Mozilla. (Of course, since I'm using Win98, that's almost on par with my C64.) I don't know if it's Win95, something only you have running, or something everyone but me has. I have a Win95 system, but it's only a 486/33 with 12MB RAM, so I can't try Mozilla on it. I can tell you, though, that it sometimes leaks resources even when the only thing running is Windows (not Internet) Explorer, and with my integrated Win98 Explorer leaks are even worse with or without Mozilla. If anyone out there finds out I'm wrong, please correct me. I can't think of it being anything else but one of Win95's 3000 known bugs interacting with an integrated Internet Explorer's many known bugs. Hope this helps a little bit. -Sailor V Mike S-- are you using any weird extensions like web-filters or anything? I know there's a bug (I found it :) ) with WebFree, the ad filter... that causes pages to be rendered all fscked up... I'd eliminate all non OS extensions and cps and try that... And make sure you have enough memory free... And be sure to clean out ALL preferences (Mozilla Registry in teh prefs folder + The documents folder in your main directory)... And be patient the first time... It takes about 30 seconds to run.. I dunno if that'll help.. You may wanna check bugzilla for descriptions of your problem and if it's not there, you might wanna file it. Good luck, W I tried running a Mac OS 9 base set + my ethernet card extention and I still hung. I gave it a few minutes to load. Perhaps one of the Mozilla files got flagged invisible some how. I'll have to search. I run my system pretty lean to squeeze what I can out of this aging system. It's really bafling, even an M12 nightly froze. I submitted a bug report but seeing as it works for others I imagine it'll just get passed over. Mike S. I don't know whether this should be the way it is, but when I compare pages loaded with NC 4.6 and Mozilla M11(and earlier), the latter displays extra empty row after each row of text. It's not very pleasant. Also, horizontal lines being displayed on the top of images is not very nice thing either. And the third thing I noticed - pages with many frames where frames nicely continue each others vizual content. So, Mozilla puts scrollbars in some of those frames, and result... Take a look at www.roxette.nl Btw, I forgot to mention I'm using Win32 version. Have A Nice Day ! Ramunas I don't know whether this should be the way it is, but when I compare pages loaded with NC 4.6 and Mozilla M11(and earlier), the latter displays extra empty row after each row of text. It's not very pleasant. Also, horizontal lines being displayed on the top of images is not very nice thing either. And the third thing I noticed - pages with many frames where frames nicely continue each others vizual content. So, Mozilla puts scrollbars in some of those frames, and result... Take a look at www.roxette.nl Btw, I forgot to mention I'm using Win32 version. Have A Nice Day ! Ramunas This appears to be a problem with incremental reflow that was recently introduced into Mozilla. I and several others have already filled out bug reports on this subject. It should get resolved soon. i'm using some CSS2 fixed positioning in style tags in pages which also reference external style sheets, and the lizard has never displayed these right. i mean *bad* errors, like a stray toolbar ghost image plonked down in the web text itself for no reason, and utterly refusing to scroll the webpage in response to the scrollbar. M11 at least isn't quite as bad as earlier versions, which would hash the entire window into white noise garbage. still seems a bit slow, still eats a lot of memory, and little things like this text box just don't work right. making progress though. Mike S-- You sure you have enough memory to run it right? I'm not sure what the req's are, but probably with all the non-fixed bloat it's more than seems normal for a browser (just a guess though) What else could it be... You've tried without the Dayna extension too, so it's just raw OS 9? Maybe post something to netscape.public.mozilla.mac (server: news.mozilla.org) and ask the developers for any suggestions. Maybe there is some kind of hardware issue here.. (?) Are you using a proxy, and if so did you configure the prefs to use it? Are you SURE you've removed ALL the preferences (document folder from main hd + "Mozilla Registry" from preferences)? Are you up to date with the Dayna driver? Damn... I wish I knew what to suggest... I know how you feel cuz for two months it wasn't working right for me and only me, it seemed until thanks to the folks on #mozilla, we narrowed th problem down to the WebFree ad-filter extension... W I have 96 MB of RAM and a 225 Mhz 603ev. Should be plenty. The only thing I didn't try was Dayna off. I didn't see a point since if I can't run the Dayna extention I can't get on the net at all. I ran a Mac OS 9 Base + the Dayna extention. Dayna was bought by Intel and they apperantly have no plans to update the drivers so I'm kind of stuck with what I have. Perhaps it's just fate telling me to wait for an Alpha. :) Anyhow, I can still play with the new iCab beta (end of the month I hear) while I wait for M12 Alpha. If that doesn't work then I'll try it without the Dayna extention and see if it still crashes. If it doesn't I'll submit a new bug to Bugzilla. It's funny how software works. Even with exactly the same variables two people can get different results. Some people were having nightmares with Mac OS 9 while others with the same configs reported no problems. I'm of the latter :) Thanks for your assistance though. Mike S. This is off-topic, but a few postings on opinions about the sidebar caused me to wonder about this -- Does anyone know if the Mozilla sidebar will have any "file manager" support for local directories? (similar to IE5's ability to access local drives?) I can't be sure, but I thought I read about some early Aurora plans for this - could it also be a postbeta feature? Adam Hey, to respond to whomever it was that suggested an "advanced preferences" I think that'd be great or at least a file called like advanced_prefs.txt where we could tweak ALL the settings, not just the ones in the preferences panel. You know, kinda like how Unreal or other games have an .ini file that you can use to change all the technical settings... Of course, I'm presuming this has or hasn't already been built into the architecture already.. Just remember that Mac users can't very easily add parameters to the startup line, cuz it doesn't exist.. There's always prefs.js and while I can't speak about teh Netscape branded product, I expect that any easily switched preferences (from prefs.js) in mozilla will be public. If this doesn't work for the advanced user, then there's always the 'download the source and build it yourself with whatever settings you want' option. There's always prefs.js and while I can't speak about teh Netscape branded product, I expect that any easily switched preferences (from prefs.js) in mozilla will be public. If this doesn't work for the advanced user, then there's always the 'download the source and build it yourself with whatever settings you want' option. Mac 233 G3 minitower w/192meg-system 9.0 w/multiuser on-12 meg to M11. Ran OK (slow). Crashed twice w/error 11 & 3. Text renders as if on a windoz platform (20% larger). MouseOvers did not load untill moused over at a GoLive created site (http://165.121.130.107). Non scrolling frames had scroll bars. This is as far as I got in a 20 minute trial. Is mozilla importing old code from Netscape at all? View source still doesn't let you copy (cut/paste) from it, which has to be one of the most idiotic decisions ever made by Netscape which they continue to this day (surely I can't be the only one that finds this most aggravating?). Why are you complaining about a program that isn't even alpha yet (let alone beta)? Not everything is working yet, but eventually they will be. If everything is working to some degree, then Mozilla is pretty much ready for beta right now, but it isn't yet because not everything works. If you want them to work on certain things sooner, you can file a but at Bugzilla or find an existing bug and vote for it. Patience. <:3)~~ While it has absolutely no relevance at all to Mozilla, you can do that in Netscape by using the copy hot keys for your OS du jour. (Control-C, Command-C, and I think Alt-C for the *n*x versions [because they don't want to screw with emacs bindings]). I've tried to download M11 repeatedly for the past few days and it keeps stalling my connection. I've tried everything I have from NN4 to IE5 and even CuteFTP 2.5. I only have access to one ISP (we use the same one at work) so I can't test if that's the problem. Does anyone know of a mirror where I can try and download M11 so I can see if it's the server? |