Another Round of 1.3 Candidate Builds
Tuesday March 11th, 2003
Several fixes have been applied to the 1.3 branch in the last few days and a new set of release candidates is now available. Get the complete details from Asa Dotzler and then grab a build to help with the testing effort.
Will Netscape build Buffy on 1.3 or why are there Candidates? ;)
#2 Yes and No.
by naylor83 <firstname.lastname@example.org>
Tuesday March 11th, 2003 5:06 AM
They are encouraging testing before the release, so as not to repeat the 1.2 mistake of shipping with a big bug.
#3 Linux XFT Builds
by Querty <email@example.com>
Tuesday March 11th, 2003 6:13 AM
I'm sorry to say I'd be more than willing to test if they would only build an XFT enabled version. There is no way I'm going back to non-XFT builds. If there are still XFT issues, this would be a good time to fix them.
I had a quick look at the previous 1.3 build, and all seemed well (and fast).
My pet peeve, however, is still not fixed:
Keyboard scrolling and mousewheel scrolling in overflow: auto elements.
See <http://www.tunfun.nl/> for an example (there are no frames or iframes on the page, it is all CSS)
It seems to be partly a focus bug, since if I use <Tab> to highlight a link within the element the mousewheel works until I deselect the link. IE gets this right, b.t.w. (although it gets a LOT of other things wrong).
Keep up the good work!
#4 Re: Linux XFT Builds
by Querty <firstname.lastname@example.org>
Tuesday March 11th, 2003 6:22 AM
Ok, I was being lazy....
Here is the bug:
there are still XFT-issues, <http://bugzilla.mozilla.org/show_bug.cgi?id=172768> , but that shouldn't be a reason not to have nightly XFT builds - there's a bug on that at <http://bugzilla.mozilla.org/show_bug.cgi?id=176382> , with links to a contributor that's building them. he doesn't seem to be building the 1.3 branch, but you could ask. I'm also not sure if there's a good reason why some or all of his builds couldn't be on the mozilla server as well...
"My pet peeve, however, is still not fixed: Keyboard scrolling and mousewheel scrolling in overflow: auto elements."
If that's your main peeve then you must be one of the happiest users out there! I think that site that you linked to is the first time that I've ever seen a CSS overflow:auto used "in the wild"
We use this to get a table body to scroll. Very handy, under IE we have to use a DIV and the table head is actually another table in order to keep it in view.
Oh, I *am* one of the happiest users! Mozilla/Gecko is the best!
A page like this is indeed rare in the wild.
I got so fed up with frames, iframes and other deprecated "hacks" that we started looking in to doing things exclusively using "XHTML Strict" + CSS2. Although it is beyond the edge of current CSS usage, it is absolutely fun to design and maintain such websites.
It is not for the faint-of-heart though, as getting things to work in other browsers can be a challenge.
I've been using overflow:auto on my site (<http://coda.co.za>) for quite a while now and I'm with you on the scrolling issue, it's really annoying! I've considered switching to iframes.. bleh. ;)
#10 Causes a hang bug too
by mlippert <email@example.com>
Tuesday March 11th, 2003 10:28 AM
On Windows XP with the previous 1.3 candidate (I don't know what other version numbers I could specify here) if I go to the link you specify that has the overflow: auto element: 1) My 1st attempt, it all showed up fine no scrollbars to be seen. So I tried to change the text zoom to 150%...Hang. 2) repeated 1 just to be sure. 3) Came back here, changed the text zoom to 150% (everything looks fine) then click on link, the page comes up with an internal scrollbar which scrolls just fine...now I try to change the text zoom to 100%...Hang.
#11 causes a Hang Bug too
by mlippert <firstname.lastname@example.org>
Tuesday March 11th, 2003 10:30 AM
(resubmitted to get paragraph breaks)
On Windows XP with the previous 1.3 candidate (I don't know what other version numbers I could specify here) if I go to the link you specify that has the overflow: auto element:
1) My 1st attempt, it all showed up fine no scrollbars to be seen. So I tried to change the text zoom to 150%...Hang.
2) repeated 1 just to be sure.
3) Came back here, changed the text zoom to 150% (everything looks fine) then click on link, the page comes up with an internal scrollbar which scrolls just fine...now I try to change the text zoom to 100%...Hang.
#17 Re: causes a Hang Bug too
Tuesday March 11th, 2003 11:28 AM
Yeah, The page displays correctly if you change the text size before going to the page. If you change the text size in any way (larger or smaller) while on the page, the hang happens. The scroll bar works fine with the mouse, but shouldn't the arrow keys work too?
is that there no talkback-enabled tar.gz builds for Linux. Either I want talkback, then I have to use Installer builds, or I want easy, no-intervention install, then I cannot have talkback. This is even more idiotic as the the Win32 ZIPs *do* have talkback enabled. Can anybody explain the logic behind this to me? Anyway, since I hate the Installer, especially when doing installs more frequently, I am stuck with non-talkback anabled builds.
You've got a point there. The installer builds are quite useless... at least when comparing to automated installation of tar.gz builds. Is there any real reason why the tar.gz builds don't have talkback enabled...?
#14 Re: What pisses me off ...
Tuesday March 11th, 2003 11:05 AM
Are you offering a patch to the build and packaging system? No? OK.
Um, I think he was just simply asking why there isn't one. We don't all have to be tech experts to wonder why such-and-such isn't implemented, do we?
#20 Re: Re: What pisses me off ...
Tuesday March 11th, 2003 12:25 PM
no, but there's not much point in asking the question, or being pissed off. 99.9% of the time the answer to "why isn't X implemented" is "because nobody has implemented it yet". possible solutions are to implement it yourself or to pay or persuade someone else to implement it. announcing "I'm pissed off" to the world in general is unlikely to persuade someone that they should do something for you... if you walk into a car show room and shout "it really pisses me off that nobody has given me a free car yet", you are equally unlikely to get a good response...
I think the fact that the issues being raised here are all quite minor is a good sign. :)
this does not exist because it is an effort or complicated to do, rather than because someone decided to do it that way? I am usually not quick with strong language and, yes I have done a few little patches for things that pissed me off. But this just looks very odd and unnecessary to me. Why the difference between the way how it is done for Linux and Windows? I guess what really pisses me off is that nobody ever cared to answer that question, even when I asked politely. Giving me your standard out of the packet "do it yourself" answer is not very helpful really. In another discussion the question has come up why there are so few contributors to Mozilla (well, less than would be needed anyways). Arrogance and unwillingness to help of the core developers are likely causes for this IMO.
To get back to the main topic: you were asking us to test the builds. In other posts we are asked to use talkback enabled builds. I stated that after doing several dozen of Installer installs I am pissed off of the fact that I cannot simply install a talkback build using untar. Well, maybe I should have just kept that fact to myself and simply proceed with not bothering with testing or installing ...
#30 Re: are you telling me ...
Tuesday March 11th, 2003 5:31 PM
"To get back to the main topic: you were asking us to test the builds. In other posts we are asked to use talkback enabled builds. I stated that after doing several dozen of Installer installs I am pissed off of the fact that I cannot simply install a talkback build using untar. Well, maybe I should have just kept that fact to myself and simply proceed with not bothering with testing or installing ..."
Or you could test the million other parts of the build besides stability.
Get a not-talkback build if you don't like the installer. Talkback is good. The more people using talkback builds the better, but if you don't want to use the installer then don't. Having talkback is in no way a pre-requisite for testing a Mozilla build. Even if you did crash this candidate build and submit a talkback report, it probably wouldn't have any impact on the release. We already know what our topcrashers are for this milestone and if we just introduced a new one in the last day or two, we'll see it from the people that do like the installer builds or we'll hear people screaming that the builds are crashing and they don't have talkback. (are the builds crashing a lot for you?)
Thanks asa for the good answer.
I can understand why people would like to give the typical "it is open source, if you don't like it submit a patch" answer. I can also understand why simpler users/testers would get frustrated with getting that answer. Why? What makes them think they can complain at all when they get something (good) for free? Who gives them even the right to think they can complain??
Well, I think it is easily explainable. They think (as I do) that they are part of the team. They think they are part of it, part of a group that is trying to build a browser that doesn't suck. They are told : "Testing is important. Testers are needed. Be part of the team, even if you can't program, submit patches, etc. Be part of the team, do you part, is it needed, good enough and appreciated." Then when they complain they can't do their part well because of this or that, they get an answer which basically says (to them anyway) "you're not doing a part that is big enough. Do more. What you do now is useless. We don't need you unless you do a bigger part.". Now suddenly they aren't part of the team anymore. They are just low testers that are bugging important people with their small issues. If someone feels strong enough about a effort like mozilla to try to be part of it, test it, send bug reports, read mozillazine, etc. like this reader seems to do (he even said he submitted patches), I think he should be treated as part of the team. I really like your last answer because even if you didn't address the core issue (what is blocking like gz build to have talkback, how can this be solved?) you took time to reply something useful leaving the impression that the tester is welcome, and part of an effort, part of the team.
Hope this make sense. I hate it when I see people trying to build something together screaming at each other. I sure hope this little explanation helps in understanding how someone can feel on the other side of the fence.
Long time lurker,
That pretty much sums up how I feel all the time when these sorts of discussions come up. Thanks for your eloquent statement, mathieu. I couldn't agree more, or more loudly.
Not that what *I* think really matters, I'm just a lowly user. ;-)
you again did not tell me if this is an oversight or a deliberate decision. In other words, if I provided a patch for this, would it get checked in? Would you r/sr it? In my view, not having talkback enabled in the tar.gz build is a dumb decision and it is both that decision and the unwillingness to tell a little outside contributor like me why it was reached that annoys me. And the more I am talking about this I am getting the impression that it is NOT an oversight or something that nobody was able to fix yet.
#36 Re: Re: What pisses me off ...
Tuesday March 11th, 2003 11:00 PM
Yes. I am offering to fucking back out the patch that REMOVED this functionality from the Linux builds (it used to be there). Should be pretty easy, right? Think it will get any reviews, Asa? No fucking way -- the build people are very insistent that this will be done their way. That's fine, it's their module. But I wish they'd have some technical reasons for the decision and put them in the relevant bug instead of hiding behind "it's not supported" (with no reasons for WHY it's not supported).
#44 Re: Re: Re: What pisses me off ...
Wednesday March 12th, 2003 3:00 AM
Either it was a pain in the ass to maintain or it made other parts of the automation more difficult or both. I know that when we did have it it was intermittant and I probably got email on average once a week or from someone saying "talkback doesn't work in today's linux build. do something." I know I blew time on several occasions trying to figure out if talkback wasn't packaged with that particular build, was packaged but wasn't working, was working but not for that particular crash, etc. and I imagine folks in build got even more complaints. I don't know why it was a pain and I don't much care because I don't see any huge win in having it for the zipped builds. I remember vaguely sending an email saying "sure, kill it, we've got more important things to worry about" or something like that but I don't remember why it was so much trouble.
We have no shortage of milestone talkback reports on Linux. We could probably get by on half the reports we currently get and still be able to identify and fix every one of the topcrashers we currently identify with talkback. We're not going to open up any floodgates of new information by enabling it in the zipped builds.
That being said, it certainly wouldn't hurt to have talkback for zipped builds on linux. The more the merrier. There may be some bug that's hard to diagnose and the one person with good comments might just happen to be using zipped builds. If you're offering to fix and maintain that piece of the build/packaging system or can convince someone else to maintain it then I'm happy to do what I can to get it reviewed/approved (although I'm not technically capable of doing code review for it). But if it's a pita to keep it working and no one's interested in doing that then I'd rather it not be there at all then it be there only some of the time.
#55 Re: Re: Re: Re: What pisses me off ...
Wednesday March 12th, 2003 9:48 AM
I'm not talking about the topcrashers. I'm talking about all the non-topcrasher crashes we get reported where the reporter says he'd list a talkback ID but getting talkback on Linux is too much of a PITA. I'm talking about faster identification of new crash bugs when they appear. That sort of thing.
I agree that having talkback only intermittently would be worse than not having it at all, though.
#72 Re: Re: What pisses me off ...
Wednesday March 12th, 2003 7:08 PM
Should a million people be involved with Mozilla? Should everyone in the world be programmers and learn how every program works INTIMATELY that they use? Asa, have you ever thought something was annoying about a program but thought it was too much trouble to fix yourself?
I don't know about you but I think user input should be welcome (even if ignored because of more urgent things).
#33 RC builds and Talkback
Tuesday March 11th, 2003 7:29 PM
While this may be somewhat of a legit concern for Moz developers on release builds and perhaps on alphas and betas, this is not a concern non-devel people should have with RC builds. The talkback data from the RC build certainly won't get analyzed before 1.3 is released, and when it is, Moz developers may worry about whether or not they're getting enough talkback data and react accordingly. In short, don't blow a blood vessel because talkback isn't enabled.
Could it be possible to get a suiting talkback "component" or whatever somewhere? A talkback.xpi/xpt or something? E.g. try using the installer to always get a dummy installation of mozilla and copy talkback from there to yout real Mozilla dir. Or maybe you can get the xpi (or whatever it comes in on linux) somewhere on ftp or something?
#58 Re: talkback.xpi or similar?
Wednesday March 12th, 2003 11:44 AM
The problem is _automatically_ getting the XPI that corresponds to your Mozilla version. This is nontrivial.
#79 Where to find talkback.xpi:
Thursday March 13th, 2003 6:06 PM
Well, at least for 1.3 there is <ftp://ftp.mozilla.org/pub….3/linux-xpi/talkback.xpi> and for nightlies there is <ftp://ftp.mozilla.org/pub…nk/linux-xpi/talkback.xpi> and others.
Bug 186898 [ Cookie Manager scrollbar looks messed up ] is a very visible regression from 1.2.1, but no patch exists so its pointless to nominate it?
#23 Re: Visible regression in 1.3
Tuesday March 11th, 2003 12:51 PM
Are you seeing a problem other than when you re-size the window to be too small?
When I install the Mozilla release on Windows 2000, I get the "Welcome to Netscape 6.1!" screen. Anybody else get this, or is this just me?
#13 Re: Welcome to Netscape 6.1! ?
by johnlar <email@example.com>
Tuesday March 11th, 2003 11:04 AM
You need to seriously clean out the clutter in your computer, try removing all moz*.dat files in the windows directory and removing (or moving) your profiles and the program files/mozilla.org directory
#50 Re: Welcome to Netscape 6.1! ?
by webgremlin <firstname.lastname@example.org>
Wednesday March 12th, 2003 6:39 AM
Happens to me too sometimes. I think only on my fiancee's computer.
I seem to recall a similar problem a while back. I located part of my problem in the prefs.js file located in C:\Documents and Settings > Application Data > Mozilla > Profiles etc. You can edit the prefs.js file even though it says it is a generated file. At least I did. If the Welcome to 6.1 is at Netscapes website when you first start up the newest build, look at the bottom of the prefs.js file. You will find a line that says something like "user_pref("startup.homepage_override_url", "<http://www.netscape.com/welcome.html>");". Change the URL to your own prefered homepage. You won't have that problem again.
#54 Re: Re: Welcome to Netscape 6.1! ?
Wednesday March 12th, 2003 9:39 AM
Thank you! That was probably it. My prefs.js did have this property set to the NS6.1 URL. I just went to about:config and clicked "Reset" on the startup.homepage_override_url property, I bet thats the fix. Thanks again!
is anybody else seeing this? open any page in Print Preview. the scrollbar is black, no up/down arrows. although u can scroll the page.
i'm using Mozilla Classic theme and yesterday's 1.3 build.
#18 Re: no scrollbar in Print Preview
Tuesday March 11th, 2003 11:34 AM
I see this with the modern theme. Strangely, when I try to switch to the classic theme I get "You have selected a them which was designed for an earlier version of Mozilla..."
#19 Re: Re: no scrollbar in Print Preview
Tuesday March 11th, 2003 11:55 AM
filed a bug 196908
#21 Re: Re: no scrollbar in Print Preview
Tuesday March 11th, 2003 12:28 PM
you're supposed to check to see if the bug is already filed first...
but anyway, as I've just commented in your report, this is a known problem <http://bugzilla.mozilla.org/show_bug.cgi?id=191938> It's fixed in the trunk (1.4a) code, but unfortunately the fix came too late and was too risky to consider for 1.3.
#24 Re: Re: Re: no scrollbar in Print Preview
Tuesday March 11th, 2003 1:04 PM
"you're supposed to check to see if the bug is already filed first... "
i sure did. it's just it didn't occur to me to search within resolved bugs.
#32 Re: Re: Re: no scrollbar in Print Preview
Tuesday March 11th, 2003 6:02 PM
for future reference, it's a good idea to search in resolved bugs. apart from the fact that the bug may have been resolved between your version and the current time, if you search resolved bugs, you may find duplicates of a bug which match your keywords when the original bug doesn't.
"too late" is relative... The fix at least came before 1.3b. And there is also a different patch by bzbarsky in the original bug about the visual problem (bug 191875) that is even some days older and does not do the risky rewrite. This patch only changes 12 lines in a simple way. Did somebody think of getting that one for 1.3? Now I guess it's too late...
drivers reconsidered the decision based on the fact that it was getting more attention than they'd expected, so the patch is in (and the whole thing, not just the 21 lines :)
I still cannot understand how somebody can expect that such an obvious problem in the UI does not get much attention. Attention and bad publicity would've been even worse after 1.3 *final*. There aren't only testers who download Mozilla final releases, no matter if this is welcome or not... Even my girlfriend said: "that's only a scrollbar, it can't be too complicated to fix it!" This may not be right, but this is what most users would've thought. And they would've concluded that the quality of Mozilla in general must be bad if even such obvious errors are in a stable release.
But it's over and fixed. Thank you! Thank you all who have made this happen: Josť Jeria and Rick DeBay who did not stop bugging the owners of the bug, Boris Zbarsky, who did not stay as decided against fixing it on the 1.3 branch as he was, Robert O'Callahan, who finally decided to ask drivers about it and - last but not least - drivers, who let this in very late in the game as probably the second to last fix. Thanks. ;-)
P.S.: the patch does not have approval1.3+. Who do you think you are to check this in?!? ;-) Just kiddin, but since drivers agreed, approval 1.3+ can be set, I guess.
#25 mozilla.org still announces 1.3 beta
Tuesday March 11th, 2003 2:02 PM
It would probably be a good idea to remove the 1.3beta links on mozilla.org and add the links to the latest release candidate. Whats the point in letting someone test the old builds? And there might be more users testing the release candidates too.
A colleague has found a worrying problem with Mail. He starts Mail then selects a random message from his inbox and attempts to delete it (del key, menu, whatever). The message instantly vanishes from the inbox, but Mozilla then sticks in the moving message to trash state. Attempts to delete further messages in that session are all ignored. This is using POP, with mail stored on a network drive. This was with todays release candidate and clean installs (but keeping profile). We tried the previous RC and got the same results. When we went back to 1.2.1, all was fine. I haven't had a chance to properly test this, and so am slightly reluctant to put up a new blocking bug. I am concerned that this is a regression related to 142196 (an old friend of ours). IMAP works fine, and also POP with local mail. I haven't yet had a chance to reproduce it with a different user, also using POP and network drives, but will do tomorrow. If this does prove reproducable it would be a disaster. POP with network drives only became useable at 1.2, and this would send it right back again. A big problem for enterprise. :-(
Can anyone offer any suggestions on how to proceed here? Is it already too late?
This isn't a fix, but it is a workaround for many network related problems: Map the network share containing your data to an actual drive letter. Then, make sure the drive is mapped before starting Mozilla and have Mozilla look at that drive letter for the account/profile information.
#43 Re: Re: Problems deleting emails
Wednesday March 12th, 2003 2:16 AM
I've now had chance to do some proper testing and things are looking fine. My colleague still has the problem, but I can only assume that this is down to something specific about his setup. For a start, his is the only machine in the tests that stores his mail on a UNIX net share; all the others come from windows. Still a bug, but not the showstopper I thought it might have been.
#60 It happens to me a lot prior to 1.3
Wednesday March 12th, 2003 12:05 PM
I'm still on Moz 1.2.1. I've been seeing behaviour like this for the best part of a year. I don't use the mail mainly because this problem destroys my trust before I even get going. When Mozilla imports my Netscape 4.79 profile, it has trouble listing all the messages (2,200) in my inbox. When I start deleting messages and folders, it might work, or it might just copy them to the trash. Once it starts copying instead of moving, I have to close Mozilla, delete the summary files, and try again. It even gets to the point where empty the trash doesn't work without a restart.
Maybe it's just me, but it seems like it's very slow to open new pages and tabs. Otherwise good.
#35 html source editor of my choice --- gvim
Tuesday March 11th, 2003 10:22 PM
Is it possible to add such feature? I would like to use vim to view html source. It is possible for IE, but not mozilla, yet. Thanks
#37 Re: html source editor of my choice --- gvim
Tuesday March 11th, 2003 11:01 PM
This is trivial to add with one line of codechange and optimoz. Feel free to do so.
#63 Re: Re: html source editor of my choice --- gvim
Wednesday March 12th, 2003 1:19 PM
bzbarsky: any idea why no one has bothered to implement this? Every web developer I know who runs on windows with IE would use Mozilla, has tried to, but has gone back to using IE because they want to be able to edit the page via view source. What is with this? Why wouldn't Mozilla.org add so many users to its codebase for such a simple feature?
#68 Re: Re: html source editor of my choice --- gvim
Wednesday March 12th, 2003 3:42 PM
Please vote for bug 8589. <http://bugzilla.mozilla.org/show_bug.cgi?id=8589>
#71 Re: Re: Re: html source editor of my choice --- gv
Wednesday March 12th, 2003 5:38 PM
Um... That's not the same as viewing the source in an editor. You're asking to edit the original file, which is a totally different kettle of fish.
As for why it has not been added, it's because there is no "default editor" on a lot of the platforms Mozilla uses. So one needs to be a little careful about making this change.... and also because I know far more people who hate the IE view-source thing precisely because it uses Notepad. I doubt that any change here will significantly affect Mozilla usage
The rendering for this site is whacked in the latest candidate build.
Nevermind. I just had text zoom on. Ignore and move on.
Who should I kill to make a patch with r+ & sr+ land in CVS for branch&trunk? <http://bugzilla.mozilla.org/show_bug.cgi?id=182997>
for the branch, you would need to get approval from drivers, and I think it's too late now for 1.3.
for the trunk, all you need is someone with CVS permissions to do the checkin - if nobody spots it as a result of your post here, you could ask timeless, who did the review, or another developer, by email or go onto IRC...
Bug 196628 ( <http://bugzilla.mozilla.org/show_bug.cgi?id=196628> ) is set as a blocker to 1.3, and no listed activity for a couple of days. So it seems that other bug fixes still have time to get approved for inclusion.
#62 Re: There is still a blocker
Wednesday March 12th, 2003 12:32 PM
I think they're aiming to tag the release today, so I guess that bug won't make it, especially as there doesn't seem to be a quick solution on hand...
#56 Is this build supposed to open a blank window,
Wednesday March 12th, 2003 9:59 AM
For all downloads?? It seems that with many downloads when you click on the file to download a blank window opens up and then you finally get a download box. In contrast Phoenix 12-28-2002 just correctly gives you the download box. With this latest 1.3 build I end up having to close blank windows ever time I download a file. I been using Phoenix for a long time now so maybe I forgot how Mozilla deals with downloads, but this behavior certainly doesn't seem right.
#66 Anyone having a problem with imported Favorites?
by kberk <email@example.com>
Wednesday March 12th, 2003 3:26 PM
Favorites did not work for me, but I'm using a beta version of Windows. Is anyone else having this problem?
#73 Re: Anyone having a problem with imported Favorite
Wednesday March 12th, 2003 7:13 PM
Yes. Just search Bugzilla.
#67 Wow. Look at the Tools menu.
by kberk <firstname.lastname@example.org>
Wednesday March 12th, 2003 3:37 PM
I use Netscape 7 so much that any time I download Mozilla I'm surprised by something. I just noticed the Popup manager, and the Image manager. If I where dialup, that Image Manager would be awesome.
the image manager still does not allow what was already in 4.x: there you could suppress the download of images, but then deliberately view individual images in a page by right-clicking and selecting "view this image". That was extremely useful for slow connections, but is impossible to do in Mozilla, unfortunately.
I just crashed at nationalpost.com <http://www.nationalpost.com/> after a long browsing session. No Talkback ID. Windows 98. I was clicking on a few different things. As the front page opened in a tab, Mozilla crashed.
I agree, this is important. Go to bug 47475 <http://bugzilla.mozilla.org/show_bug.cgi?id=47475> and vote for it.
It has an old patch, maybe anyone could revive it?