Tree Branches for Mozilla 1.3Saturday February 22nd, 2003Yesterday, the tree was branched in preparation for the final release of Mozilla 1.3. Checkins to the 1.3 branch are being tightly controlled by drivers@mozilla.org. Meanwhile, the trunk has been reopened for 1.4 Alpha development. See the Roadmap for more details and tinderbox for the current status of the 1.3 branch and the trunk. A list of bugs blocking the 1.3 release is also available. I am surprised that bug 193038 didn't make the list. I would think that a bug that prevents using the onclick or onmousedown event handlers on a link would be considered a significant impediment to DHTML functioning. The link to the bug page is http://bugzilla.mozilla.org/show_bug.cgi?id=193038 I have posted a page demonstrating the failure of the link function at http://www.oregoncanoesport.com/Link_error.html If you agree with me that this should have more attention, please vote for this bug. Something does not look right. In MSIE 6.0 SP1 and Opera 7.01, I was able to click the link. In the latest trunk build, it's broken. Not if no one runs into it.... Yes, it should be fixed. But unless it's a common problem it should not block the release. Were it not such a fundamental function, the execution of a hyperlink, I might agree (though in so doing, I would, perforce, be relegating myself to "no one" status). Links are the bread and butter and the meat and potatoes of the Internet. For a browser that touts standards compliance to not execute a link in valid markup is, unfortunately, a significant indictment against said browser. There is a major push to get developers to use CSS in place of JavaScript to streamline pages and simplify maintenance. Minimizing the importance of this bug is to minimize the importance of that movement and the standards upon which it is based. The reality is, people may very well be encountering this bug and simply writing off Mozilla and going back to IE. In summation, I believe it is in everyone's best interest that this bug be addressed before a new version of the browser is finalized. I have a hard time imagining splash screens being more important than basic functionality. the splash screen is a legal thing. look at the list of other bugs which have been nominated as blockers and rejected - http://bugzilla.mozilla.org/buglist.cgi?resolution=---&field0-0-0=flagtypes.name&type0-0-0=equals&value0-0-0=blocking1.3- . There's no way that this bug is more important than all of those. There are lots of bugs that will be in 1.3 - print preview has several known crashes and no scrollbars, the master password panel crashes, etc, etc. The idea is to fix the worst ones and then get a release out. If they try and fix everything before a release, we'd be waiting a year between releases - it's not the end of the line, there'll be a 1.4 to fix more stuff up... This bug only manifests itself in cases when a page changes the target node between the mousedown and the mouseup. This is _very_ rare. I don't like leaving this bug in 1.3, but then again I'm also not sure why we're shipping something we consider a "release" so often.... There's a disconnect between people who thing of 1.3 as a "final" version and developers who think of it as just another random meaningless name... One other thing of note... This is a very interesting theoretical situation, since there is no click event involved -- the mouseup and mousedown are on different pieces of content. It's just like mousing down on a button, then dragging your mouse off the button and mousing up -- you do not expect the mouse to trigger. In this case it does lead to rather unexpected behavior. But it's not clear how to fix it while not breaking the button case..... I appreciate what you are saying. If the bug cannot be addressed in a reasonable timeframe, because of the complexity of the issue, there is certainly an argument in favor of releasing 1.3 without resolving it. In reality, though, little separates this implementation of a graphic rollover from a traditional JavaScript rollover. The event handlers are part of the link, as they would be in a trad JS implementation. In the trad approach, graphics are preloaded into JS arrays and images swapped by src attribute manipulation. In the demonstration markup the images are preloaded as CSS objects and swapped by JS manipulation of the visibility attribute. Granted, different parsing is involved, but the net effect should be equivalent. As you can see, the resulting markup and script is compact and executes quickly. The key difference between using visibility and using "src" is that with "src" the mousedown and mouseup happen on the same image. So you get a click on that image and the click bubbles up to the anchor. With visibility, you get a mousedown on one image, then a mouseup on a _different_ image. So there is no click event. The key issue here is that in the DOM event model the "click" event is not actually dispatched at the anchor but at the innermost node being clicked on. The only reason the anchor sees it is that the click event bubbles (and a event handler on any node between the anchor and the image could make the anchor not see it by cancelling the event). So while this may look like a very similar setup for a page designer (I can see that side of things, certainly) from a UA implementation standpoint the two cases are very different.... fixing this may need a significant rewrite of how events are handled in Mozilla. I do think that we should make this case work; just not sure how yet.... Thanks for the explanation and I do see your point. I think the common expectation would be for the event to bubble to the containing element, in this case the "a" element. I appreciate that the act of swapping visibility can confuse the issue regarding the click event and I wish I could offer an easy solution. It would seem that somehow the event needs to be trapped and forced up the DOM tree to the containing node. I note that the examples you have produced differ significantly (at least conceptually) from mine, in that you are placing "div" elements in the same space with separate event handlers. The major departure here is that in my markup the event handlers are all part of a containing block element. That is why I would expect my code to work, though I would expect problems with your implementation. It seems strange that the event handlers, which are part of the containing tag, execute, but the function of the tag itself does not. If the event handlers were on the "img" tags, it would be more like your example and the failure of the event to reach the containing element would make more sense. I certainly appreciate the time you have taken in addressing this issue and hope the solution proves to be less radical than you fear. If there is any assistance I can provide, please don't hesitate to ask. I feel a certain ownership of this bug and am willing to do whatever I can to help resolve it. Unfortunately, my programming knowledge does not extend to C. The issue is that events bubble... I could move the event handlers out to the <body> in my testcases and nothing would change, because the mousedown and mouseup events, while dispatched to the <div>, bubble up to the body... And no, using z-index would not change the basic problem here. :( I'm not sure there's anything useful to be done till someone who's more familiar with the event system looks at the bug. See bug 194291 in the blockers list- replacing the old splash with an ugly black-on-orange bit of text. Terrible. Black on white would be better (no puke-orange color, and looks just fine in any depth all the way down to monochrome). The extremely blurry antialiasing on "Mozilla" in the current replacement splash + ugly orange = vomit-inducing. Maybe it'll help get people interested in decreasing the startup time so that the splash could be removed completely. I know seeing an ugly splash every time I startup mozilla would make me find ways of speeding up startup time (or make me add the nosplash commandline thing). I don't mind the orange, but the fuzzy url and copyright makes me think I need new glasses. I've never liked most of the splash screens I've seen for Mozilla, myself. I created my own that I use on my own PCs. It is designed to look similar to the standard Outlook Express splash screen. This also has the added advantage of making it less frightening to newbies. :-) You're more than welcome to use it: http://johnhaller.com/jh/mozilla/default.asp I'd be happy to give it to Mozilla for inclusion in the builds as well. You can also feel free to add it to your own listings of splash screens. I just ask that it be proper credit be given with a link back to my page and that you drop me a note if you do incude it. Regards, John Stupid question: How do you replace the default splash screen? In windows, just save the mozilla.bmp version to the directory that has your mozilla.exe file in it. In Mac OS, head to http://malletandbare.free.fr/products/csi/index.html to get the custom splash installer. In Linux, you will need to build the rpm source. Go to mozillla/xpfe/bootstrap and change splash.xpm. I like your splash screen. I also like the old splash screen that is unfortunately going to be replaced. I just did a quick search for Mozilla Splash screens on Google, and everything I found was quite ugly. The ugly orange box is better than most of what I saw. I'm surprised it's so hard to come up with a good splash screen. - Eric, http://www.InvisibleRobot.com/ InvisibleRobot.com No matter whether the splash screen is eventually removed. IMO this is a sad example where Mozilla completely lacks any OpenSource spirit. Mozilla.org is clinging to the artificial theory that Mozilla is not for end-users. Can OpenSource browsers not be used by end-users directly? I guess the overwhelming majority of coorporate employees working on Mozilla try to prevent this. Otherwise people might ask: Why install Netscape with all that unnecessary stuff packaged with it and stripped of some functions when I can have Mozilla? So, they try to make it as ugly as possible. Give it an ugly default theme, no splash screen, and so on. Just to make it different from Netscape. Mozilla is still, however, the only browser other than IE which can display an acceptable amount of pages for me. For a long time I've been quite patient with Mozilla but no matter whether its code is GPL'd it lacks a more diverse open source community behind it and you cannot easily change that because such a community must grow and learn the details of the source code from ground up. I know there is a lot of frustration among many former fans of the project. Just go figure yourself and read some of the most requested bugs. They just remain open _for years_ sometimes with an easy solution at hand an nothing happens (Why i.e. is PGP/Gesture-support still not available by default even when lots of users requested this?) Hope Konqueror / Safari make enough progress so they can display some more pages. In speed / configurability / user-friendliness, Konqueror is far ahead already because the KDE team still listens to the community. seems to be made from the people for the people. > completely lacks any OpenSource spirit The open source spirit is in the development process, not in the support offerings.... > no matter whether its code is GPL'd it lacks a more diverse open source community Let's take a quick sample. How many of the layout developers work at Netscape? How many do not? > Why i.e. is PGP/Gesture-support still not available Because no one has produced an acceptable patch. And the people who do gesture support are explicitly _not_ trying to push it into the Mozilla trunk, last I checked. The mozilla process is most definitely hurting the mozilla project. Take projects like phoenix for example. There should be no need for phoenix to exist, since all it does is what the mozilla browser component does, but prettier and easier to use. Why can't mozilla do what phoenix does? Others say, "Why can't Phoenix do what Mozilla does" because it certainly isn't the case that "all it does is what the mozilla browser component does". Where are the installable sidebars? Where is the site navigation bar? Where is the mail integration (not mail itself, mind you, but the "working well with a mail client" items like a send page menu or mailto: link handling)? Where is the Tools menu for quick access to Mozilla's scores of managers? Where is the alternate stylesheet UI? There's plenty in Mozilla that isn't in Phoenix. I happen to think that's a good thing but you're confused if you really think that "all it does iw what the mozilla browser component does". Mozilla does more in some ways and Phoenix does more in some ways. They are different projects with different goals. If there was no need for Phoenix to exist because it _was_ Mozilla then there'd probably be a need for Mozilla to exist because people want Mozilla features that Phoenix doesn't have. I personally think that most of that stuff could live as extensions and we could get to a place where Mozilla was a lot more Phoenix like (or actually Phoenix itself) but that's a different goal than was set out years ago when those involved set out to build an integrated suite of applications. --Asa I am really sad to see this bug. There are artists out there, real artists (and not people who know how to apply filters in Photoshop or PSP), who would happily donate some really cool artwork for Mozilla. With the attitude expressed in that bug Mozilla will never make it on the desktops of ordinary people. No matter what some moz developer may feel, mozilla is definitely a end user product. It's hard enough getting people to use an browser other than the one included with the OS, but then you have developers actively torpedoing any effort to make a good first impression or clean up some of the UI it becomes impossible. The old splash screen was terrible, this new one is just atrocious. Mozillazine should arrange a splash screen contest. I mean look at the Gimp, they have amazing new splash screens for every little release, how cool is that. I'm sure Tigert or someone could make an kick-ass free as in speech splash and not one of these filter monkey creations. I'd arrange the contest myself, create the necessary web pages etc. (I do web design for a living) if I believed for one second that the splash screen would be included in Mozilla. Sorry for the rant... I feel better now ;) You have to realize, the goal is NOT TO HAVE A SPLASH SCREEN and start fast enough not to need one. There's already not splash screen on Linux by default, and iirc Phoenix has no splash screen on Windows either. Investing lots of time and effort in something that needs to be removed from the build ASAP sounds odd to me.... Few things tend to be as permanent as temporary solutions in my experience. Splash screens (on Windows at least) are accepted, it's not something you should feel the need to fight against. Mozilla currently takes around 30 seconds to start up on my 800Mhz/265MB ram laptop, I seriously doubt you have some magic that will cut that down to a few seconds, or even ten. If you can't then Mozilla needs to have a splash screen. Does IE have a splash screen? No. And yes, there are "magic" ways to cut that down. For example, Phoenix starts about twice as fast as Mozilla. Reducing static footprint (a major focus area right now) will also help. More to the point, as Mozilla gets faster average-speed hardware is getting faster. So I suspect that splash screen removal can actually happen for Mozilla within the next year. Well, I guess It depends on what OS you're on. IE on the Macintosh DOES have a splash screen. Pheonix might not have a splash screen, but chimera has a really nice one. Creating a splash screen is a one time deal now. You don't even need programmers making it if it makes you feel better. We would have it for the one year and a half as someone said earlier. Appearence is everything when you want someone to adopt something new. > IE on the Macintosh DOES have a splash screen. IE on the Mac is renowned for being dog slow, even compared to Mozilla. The point is, we should start fast enough to not need a splash screen. Period. I was just trying to make the point that it isn't totally clear cut. Anywho, I just downloaded Mozilla for Mac OS X (A nightly) and it appears not to have a splash screen and it doesn't take too long to load already. Did I miss the boat? I still wanna see the infamous orange splash screen! "Mozilla currently takes around 30 seconds to start up on my 800Mhz/265MB ram laptop" Wow laptops are slow. That's about the same speed as nav 7 loading on my p100/72MB Ram. Definatly a lot slower than my P2 350MHz/256MB ram. > Mozilla currently takes around 30 seconds to start up on my 800Mhz/265MB ram laptop When did you last defragment your hard drive? For Windows programs whose start time is HD seek bound, defragmenting the HD with defrag.exe usually doesn't help much because defrag.exe just makes sure every file is in one piece and then pushes all the files to the front of the disk. It doesn't even try to put files in the same folder physically close together, meaning that some files may get moved to the front while others stay in the back, causing bad thrashing. The amateurish uglyness of the new screen can only be called pathetic. I understand that there is no reason to put too much effort into this, but why make the color clash with both default themes and probably everything on an average user's desktop? Honestly, this looks as if somebody tried hard to make it look as unpleasant an unprofessional as possible. Chances are, they did, since the goal is to remove the damn thing. Anything that pushes people in that direction is good. ;) >> The amateurish uglyness of the new screen can only be called pathetic. [...] Honestly, this looks as if somebody tried hard to make it look as unpleasant an unprofessional as possible. << Ding ding ding ding ding! We have a winner! I think that the new splash does not look bad. it's simple. the colour choosen it's original, and this is good. The colour chosen (roughly) matches http://www.mozilla.org/start/1.0 (Which should be the first page you see when you do a clean install of moz 1.0) and http://www.mozilla.org/party/ yes one could say that. Only blinking would hurt more. The question is: why would one want an "original" color here and not, say, one from the palette of the default (modern) theme or the default throbber? Ah well, it is really not that important (I dont even see this as I am using Linux all the time) - it just amazes me how the quality of bug processing suddenly changes to "childish" when issues of style or design are involved. Another thing is that the splash now includes the (C)-years, making it necessary to update it every year - is this a requirement of US law to include the year or what is the idea behind this? its not supposed to be a beautiful work of art. "Among other things, the plainness will encourage Mozilla distributors to provide a splash of their own." I have Mozilla installed in my wife's second grade classroom. The kinds all love it, not for it's features b ut because of the cool lizzard logo. They especially love the splash screen. Marketing is important. I think having a cool logo works in a projects favor. The splash screen is being changed for legal reasons. mozilla.org does not own the green dinosaur. It is being changed to something plain and simple because: - Experience has shown that anything other than staff diktat will produce months of agonising dispute, debate and hand-wringing - Other possible images have similar legal issues which are still being looked into - Mozilla distributors should be supplying their own splash screens If this is one of the reasons that makes someone pull their finger out and start producing a Mozilla distribution customised for businesses, for consumers or whatever, then it'll have done its job. A plain white placeholder was considered, but decided against because people might file it as a bug. <whine whine Mozilla is a distribution whine whine> If it's so close to being a distribution, then it shouldn't be too hard for you - yes, you - to make one. <waah waah Open Source spirit waah waah competition waah> Go your Big Book Of Open Source Words and look up "benevolent dictatorship". Gerv >> The splash screen is being changed for legal reasons. mozilla.org does not own the green dinosaur. << NO ONE wants the GREEN dinosaur, they want the RED dinosaur. Who owns that? > - Other possible images have similar legal issues which are still being looked into There's your answer, from my original post. Gerv Perhaps, given the growing maturity of the lizard, a dinosaur in a pin-strip suit would be appropriate. Does anyone own a dinosaur in a pin-stripe suit? I think not. #54 Re: May I introduce some clue into this discussion?by jsebrech Monday February 24th, 2003 8:48 AM "If it's so close to being a distribution, then it shouldn't be too hard for you - yes, you - to make one." Or, to turn that around: If it's so close to a distribution, what's stopping mozilla.org from saving everyone the trouble of creating a mozilla-based distribution and doing it themselves? After all, they have the resources (a huge fat server hooked up to a big pipe), and the brainpower. "Or, to turn that around: If it's so close to a distribution, what's stopping mozilla.org from saving everyone the trouble of creating a mozilla-based distribution and doing it themselves? " To date it's been a matter of resources. We barely have the resources to accomodate our development community. Now add to that community of hundreds (thousands if you include our dedicated and consistent testing/bug reporting/triage community) a few million end-users and assume that those actually keeping everything running don't grow in size and resources and you've got a problem. Would you recommend I stop spending my time helping developers and testers and start spending it helping end-users? --Asa For crying out loud people, it's only a SPLASH SCREEN!!! It has no functionality, does not make the product better or worse in any respect, and is only meant to be shown for a short period when you start the browser, so who cares? Hell, it could be puke green with purple polkadots and It wouldn't make a difference to me at all. Why don't you all make noise about some REAL bugs instead of wasting energy on something this trivial. You're all silly!! No, please don't encourage people to whine about real bugs. I think I speak for most developers by saying we'd rather have tons of people whining about a splash screen if it means they aren't whining about some real bug. Am I the only one who *likes* the new splash screen? Most other candidates I've seen 1) are badly composed 2) are dog's breakfasts of blotchy pixellation 3) look like they should be airbrushed on the side of an Econoline van with after-market running boards. ...or combinations thereof. Apparently quite a dispute has started based on my comment about the puke orange/fuzzy text splash. Of course the splash needs to be replaced because of legal issues, Moz.org doesn't want to spend time on finding an 'artistic' splash, etc, etc, etc. I already knew that (I read the related bugs before commenting). I was just saying that the ugly color and fuzz of the slated replacement are annoying. The latest attachment ( http://bugzilla.mozilla.org/attachment.cgi?id=115302&action=view ) is even worse- the over-antialiased L's in "Mozilla" look like something you would get out of one of those diffraction slits in college wave physics class. <rant>And no, this isn't going to motivate anybody to make a new mozilla distribution. Nobody else wants to spend the time to keep such a distribution up-to-date with the latest and greatest, nobody wants to volunteer to spend their days building the latest release on Windows and half a dozen unices, replacing one single image in each build, packaging the builds, uploading them to their server, and then getting their server wiped out because they don't have a worldwide system of mirrors like mozilla.org does. The whole "mozilla.org does not provide a distribution" thing is plainly a bunch of bunk. The whole "distribution" thing never happened and I seriously doubt it ever will. "Mozilla is a platform for development and a template for browser distributions, not a browser application." Who uses Mozilla as a development platform? Beonex. Who uses Mozilla as a template for a browser distribution? Beonex and Netscape. How many people do you know who use Beonex? Isn't the trend of more and more people ignoring Netscape entirely and (as encouraged by rave reviews by all sorts of tech news sites as well as word-of-mouth) going straight to the source- moz.org - extremely obvious? If Mozilla.org doesn't distribute a Mozilla browser, WHO THE HECK WILL???</rant> Are you volunteering to provide the support necessary to make it worthy of being considered a distribution? If so, step up. If not, sit down. "The whole "distribution" thing never happened and I seriously doubt it ever will. " Netscape 7's tens of millions of users didn't happen? " Who uses Mozilla as a development platform? Beonex." No, Beonex is a distro, not a separate development effort using the Mozilla platform. Komodo http://www.activestate.com/Products/Komodo/ Amzon Browser http://www.cfmentor.com/%7Efaser/xulex/amazon_browser/index.html The NewsMonster weblog aggregator http://www.newsmonster.org/ OEOne's HomeBase http://www.oeone.com/ and dozens of projects at mozdev.org are all Mozilla as a development platform efforts. Beonex http://www.beonex.com/ Netscape 7 http://channels.netscape.com/ns/browsers/default.jsp GETax (the Geneva Tax authority) http://www.getax.ch/dyn/ HP/Compaq Mozilla http://www.openvms.compaq.com/new/ IBM Web Browser http://www-3.ibm.com/software/os/warp/browser/ Red Hat Mozilla http://www.redhat.com/ Sun Browser http://wwws.sun.com/software/solaris/netscape/get.html The Ximian Gnome Desktop http://www.ximian.com/about_us/press_center/press_releases/index.html?pr=gnome-14-launch Caldera Open Linux and scores more _are_ Mozilla distributions. Some do heavy branding or add new features (Like Netscape and Beonex) and others make only minor changes like packaging and and config options or custom defaults for items like bookmarks and sidebars (like Red Hat RPMs with XFT enabled and Red Hat specific bookmarks and start page) but they're all distributions and they're going out to millions of users. "Isn't the trend of more and more people ignoring Netscape entirely and (as encouraged by rave reviews by all sorts of tech news sites as well as word-of-mouth) going straight to the source- moz.org - extremely obvious?" Um, no. Netscape 7 got over 18 million downloads http://devedge.netscape.com/ and Mozilla 1.2.1 got less than a million downloads. "If Mozilla.org doesn't distribute a Mozilla browser, WHO THE HECK WILL???" Netscape and all of the other organizations listed above (and probably quite a few I din't know about). --Asa #64 Shutting up-Ignore above post, I've been out of itby jensend Monday February 24th, 2003 2:26 PM Never post when both frustrated and low on sleep
"Never post when both frustrated and low on sleep " I know the feeling. You've been posting at mZ for several years, right? I think it's nearly impossible not to just blow a gasket rant every so often. I figure we're owed at least one or two a year :-) --Asa #68 Re: Shutting up-Ignore above post, I've been out oby jensend Tuesday February 25th, 2003 9:43 AM Exactly. Thanks. Can I then ask: *Why* doesn't mozilla.org provide links to the above distributions? I know of about half of them, but surely the reason that no-one uses Beonex is that you have to be really keen to find out about it. If you/they are so keen to get people onto end-user distributions, the least you could do is move the Mozilla download *off the main page*. Why not ask people to set up RSS feeds on their site stating their latest news & downloads (yes, even for Netscape!), then amalgamate those into a 'downloads' page. It's then the responsibility of those sites to maintain their download information, and Mozilla's doing its bit to ensure that the Mozilla community knows about all the great work people are doing with Mozilla. except for the first paragraph, my previous comment was a bunch of bunk, and I recognized that soon after I posted it (but neglected to reply to myself saying the same). Sorry. I need to get more sleep. |