[lotuseclat79]

Hi tomdkat,

What do you mean people's resistance to simply unchecking "work offline" - we all tried that and it reverted to the opposite behavior we all expected?

If that action had actually worked, there would have been no problem, but, as you can see - there was a Big problem.

In this case, clearly, the FF developers made the wrong call and a portion of the user base suffered the problem.

-- Tom

P.S. I will check to see what default Network Manager version I am running and whether 0.7 fixes the problem. Ok, it looks like I am running 0.6.6 version of the Network Manager for Ubuntu 8.04 LTS - I have all of the repositories in my apt sources.list and do not see any available 0.7 Network Manager there nor at the Gnome website.

[tomdkat]

Originally Posted by lotuseclat79

What do you mean people's resistance to simply unchecking "work offline" - we all tried that and it reverted to the opposite behavior we all expected?

What I mean is people would rather not deal with the issue of an incorrect network connection state being reported to Firefox by simply unchecking "work offline" in the File menu, which addresses the issue directly. You can read comments in this bug report to see what I mean. People acknowledge they "could" uncheck "work offline" to resolve the issue for the duration of that Firefox session but their insistance that Firefox, which is NOT the problem, be "fixed" is resistance to focusing their energy on getting the actual problem fixed.

Additionally, I think you're misunderstanding the "work offline" switch in the File menu. That switch is to configure Firefox's mode of operation at that time. If you're online surfing the web with Firefox, you can check the "work offline" option and it will stop connecting to the Internet. You can then uncheck it and it will connect to the Internet, provided the network connection is still there. That File menu setting isn't intended to be a "saved" option that is set with other browser options, like the size of the browser cache or whether to have Firefox check to see if it's the default browser or not.

If that action had actually worked, there would have been no problem, but, as you can see - there was a Big problem.

The thing is, people acknowledge manually unchecking "work offline" DOES work for them. They just don't want to have to uncheck that (at least until Network Manager is fixed).

In this case, clearly, the FF developers made the wrong call and a portion of the user base suffered the problem.

Not at all. In the bug report I posted the link to above, one person suggests Ubuntu made the "wrong call" by including Network Manager in the distro, considering the version they included is "broken". Keep in mind, uninstalling Network Manager solves the problem. The method you posted above also solves the problem by disabling the incorrect reporting Network Manager is doing. That is a Network Manager problem not a Firefox problem.

P.S. I will check to see what default Network Manager version I am running and whether 0.7 fixes the problem. Ok, it looks like I am running 0.6.6 version of the Network Manager for Ubuntu 8.04 LTS - I have all of the repositories in my apt sources.list and do not see any available 0.7 Network Manager there nor at the Gnome website.

Yep, I'm running the same version as you and that's the problem. In fact, one of the links you posted above points to a bug report that has a link to version 0.7 of Network Manager that could be downloaded, I believe.

Peace...

[lotuseclat79]

Hi tomdkat,

You have your opinion and I have mine.

What I experienced directly is that by clicking Work Offline to an unchecked state prior to Quit from FF was insufficient to cause FF to launch in an online state - i.e. the specific bahavior of the preceeding FF 3.0 Release Candidate 1 and 2 and Beta 3.0b5 did not have the behavior that FF launched into the Work Offline mode - so, why the change for FF 3.0?

The FF developers made a change that caused the subsequent behavior of FF 3.0 to be different. I did not experience the same thing you appear to have read in the comments. In my case, FF 3.0 never failed to launch in Work Offline mode no matter what I did, until I applied the workaround to the Network Manager code. That it worked is good enough for me to give me the launch behavior of FF 3.0 - remember, I am not working from an installed OS here - only a Live CD with very few updates.

Even with all of my repositories updated, I did not find a Network Manager 0.7 nor at the Gnome primary development website.

Your assertion about Network Manager is speculative. If it is not available from the prime development web site at Gnome, it is not available as a general release.

Just to note: I do have a network card, but I have no connection for it - only a 56k dialup phone line. If and when I choose to get FiOS, I will be able to see a connection and will not need the dialup modem. I am certain Network Manager will work as advertised at that time. I see no reason to uninstall Network Manager even though I do not make use of it (and even though it has a side effect to my dialup setup). My dialup setup for the pppd is a customized setup that I built for my original Linux installation (still bootable) that I now use for storage instead of Internet use.

At present the network symbol upon mouse over presents the text - No network connection.

-- Tom

[tomdkat]

Originally Posted by lotuseclat79

What I experienced directly is that by clicking Work Offline to an unchecked state prior to Quit from FF was insufficient to cause FF to launch in an online state

Of course it won't do this. Why? That's NOT the purpose of that switch!!!!!!!!!!!!!!!!!!!!! If you would have unchecked "work offline" after starting Firefox, that will change the mode Firefox started in. It won't change any "default" setting because there is NO "setting" to change! You're changing Firefox's mode of operation and Firefox is responding to the network connection state the system is reporting, plain and simple. This isn't my opinion, this is fact. (I mean the part about Firefox responding to the network connection state being reported by Network Manager).

- i.e. the specific bahavior of the preceeding FF 3.0 Release Candidate 1 and 2 and Beta 3.0b5 did not have the behavior that FF launched into the Work Offline mode - so, why the change for FF 3.0?

Ah yes, let's talk about FF 3b5 for a moment. First, you DO realize then when you use beta software, the behavior of that software isn't "guaranteed" or "set in stone" because development is still active on it. Why Ubuntu included the beta version of Firefox in the 8.04 release, I have no idea and I was actually quite shocked they would do that. The same applies to the release candidates except the release candidates will undergo fewer changes since the developers are in bug fixing mode, fixing bugs that they consider serious enough to hold up the release. No major functional changes will go into a release candidate release unless a major function is severely broken. If anything, the Firefox developers finalized the planned upon new feature of using D-bus (I believe) on Linux to better communicate with the underlying system. Network Manager also uses D-bus and I believe this is how Network Manager comes into the picture. I'm still a bit sketchy on how all of these pieces fit together.

Generally, if you use beta software it's at your own risk and the developers make NO guarantees or assurances the betas will even work, let alone be stable. To base your position on the behavior of beta software doesn't make sense to me but I guess you'll consider that "my opinion".

One last thing I would like to add, people have reported, in some of the BugZilla bug reports, that they have experienced this issue with Firefox 3b3 and 3b4.

I did not experience the same thing you appear to have read in the comments.

So, you're saying after you started Firefox and unchecked "work offline" you STILL couldn't access the Internet? In your previous posts, you never indicated if you actually unchecked "work offline" after starting Firefox.

In my case, FF 3.0 never failed to launch in Work Offline mode no matter what I did, until I applied the workaround to the Network Manager code. That it worked is good enough for me to give me the launch behavior of FF 3.0 - remember, I am not working from an installed OS here - only a Live CD with very few updates.

I think we've established this is NOT environmental (in terms of you running from a liveCD vs me running from a full installation). We're BOTH running Network Manager 0.6.6 and that version of Network Manager has the issue. We're on a level playing field.

Even with all of my repositories updated, I did not find a Network Manager 0.7 nor at the Gnome primary development website.

Your assertion about Network Manager is speculative. If it is not available from the prime development web site at Gnome, it is not available as a general release.

*Ahem* I stated Network Manager 0.7 is expected to have this issue resolved. I never stated it was released yet but in the threads you posted that had links to bug reports on the issue, people were discussing their experiences with using Network Manager 0.7 to resolve this issue. So, when it's available this whole thread becomes mute. For the time being, check out these links:

http://ubuntuforums.org/showthread.php?t=676992
https://bugs.launchpad.net/ubuntu/+s...er/+bug/219906

In my Google search results, I believe I found comments indicating Network Manager 0.7 might already be part of some other distros, even though it's still in beta or under development. I guess technically speaking, the version of Network Manager we're using now is in the same status and 0.7. Here are some relevant links:

How to: Install Network Manager 0.7 SVN on Hardy
http://ubuntuforums.org/showthread.php?t=797059

Of course, I suggest waiting until Network Manager 0.7 is officially released.

https://bugs.launchpad.net/ubuntu/+s....0/+bug/191889
https://bugs.launchpad.net/ubuntu/+s...er/+bug/191889
https://bugs.launchpad.net/ubuntu/ha...er/+bug/147119

I was mistaken previously. The comments I found regarding people using Network Manager 0.7 came from the bug report links I posted above, not that you had posted.

Just to note: I do have a network card, but I have no connection for it - only a 56k dialup phone line. If and when I choose to get FiOS, I will be able to see a connection and will not need the dialup modem. I am certain Network Manager will work as advertised at that time.

Yep. The comments I'm reading indicate Network Manager 0.6.6 doesn't support PPP well, yet version 0.7 does a better job.

I see no reason to uninstall Network Manager even though I do not make use of it (and even though it has a side effect to my dialup setup). My dialup setup for the pppd is a customized setup that I built for my original Linux installation (still bootable) that I now use for storage instead of Internet use.

You have effectively uninstalled it through the workaround you implemented above. Firefox can run without Network Manager being installed. Given your liveCD environment, it will probably be harder for you to uninstall Network Manager and that won't be worth the effort.

You should actually read through the links I've posted above. Some people are using wvdial, as you are, as well as other tools. Some are using dial-up connections, some are using 3G connections, and so on. If you're looking at pointing a finger at someone, point it at the Ubuntu maintainers. One of those bug report links actually identifies this Network Manager problem before 8.04 was released (at least I think the discussion took place before 8.04 came out). So, this isn't anything new and isn't something "discovered" by Firefox users. Firefox is definitely affected but this isn't a Firefox issue.

Peace...

[lotuseclat79]

Hi tomdkat,

Excuse me, but in about:config there is a switch, browser.offline, that gets toggled every time you click on Work Offline. Clicking on the switch the first time the event occurred did indeed cause FF to go into an online state - but, that only lasts for the session, i.e. it is insufficent to last between sessions (which is the behavior that I expected, but did not get)!

No previous versions of FF ever caused the new default behavior of Work Offline upon launch.

Never said I couldn't access Internet after unchecking Work Offline.

What you said about Network Manager in post #17 was:

Yep, I'm running the same version as you and that's the problem. In fact, one of the links you posted above points to a bug report that has a link to version 0.7 of Network Manager that could be downloaded, I believe.

When it is necessary, I will followup - thanks for all of your comments!

-- Tom

[tomdkat]

Originally Posted by lotuseclat79

Excuse me, but in about:config there is a switch, browser.offline, that gets toggled every time you click on Work Offline. Clicking on the switch the first time the event occurred did indeed cause FF to go into an online state - but, that only lasts for the session, i.e. it is insufficent to last between sessions (which is the behavior that I expected, but did not get)!

Great! The behavior you describe is a good thing. That means Firefox is working correctly in that unchecking "work offline" enables you to get online. As I have posted above, the "work offline" switch controls Firefox's behavior for that session. It's not intended to be a "permanent" setting, like the others I mentioned above. For a moment there, I was starting to think you couldn't access the web at all (regardless of unchecking "work offline" or not) without mucking with Network Manager. Thanks for confirming the behavior. What you experienced during the "uncheck" test is mostly what I experienced when I unplugged my network cable.

No previous versions of FF ever caused the new default behavior of Work Offline upon launch.

Firefox 3 is the first version of Firefox to "use" D-bus messages or otherwise interact with Network Manager. This choice to interact with D-bus isn't necessarily a bad thing. It's just Network Manager is reporting bogus info and that is causing problems with several network aware apps that also "use" D-bus.

Never said I couldn't access Internet after unchecking Work Offline.

Yep, you just never said anything about unchecking "work offline". You focused on what happened when Firefox started up, which is why I had to ask.

What you said about Network Manager in post #17 was:


When it is necessary, I will followup - thanks for all of your comments!

Yep, I was mistaken in post #17. I corrected myself on that in post #19. I thought I found that info in the links you posted and the info I was referring to was actually in the bug report links I had posted links to.

Peace...

[lotuseclat79]

Hi tomdkat,

The new Release Notes of Firefox 3.0.1 say:
Fixed a Linux issues where, for users on a PPP connection (dialup or DSL) Firefox always started in "Offline" mode (bug 424626)

The comments make interesting reading among the developers.

Looks like I should now be able to backoff my changes to NM. I have just now installed FF 3.0.1 and still need to make the backoff changes.

-- Tom

P.S. I have now backed off my original changes to the NetworkManager.conf file, and rebooted - all is well again.

[tomdkat]

Yep, I saw that and a lot of that commentary was in the bug report links I posted above.

Dan Williams was right on the money with his comments.

Peace...

[lotuseclat79]

Hi tomdkat,

Comment #90 Martin Steigerwald 2008-07-04 10:34:45 PDT says it all.

-- Tom

[tomdkat]

Comment #90 says a lot, but definitely leaves even MORE out.

Obviously, that person doesn't understand software development. The statements made about Network Manager clearly show he doesn't understand or realize that Network Manager is still in active development, hence the 0.x version numbering. I don't think Network Manager developers "advertise" that it supports ALL kinds of network interfaces or connections 100% correctly, yet. I'm sure that poster hasn't even been to the Network Manager site to find out what it is and what it's actually trying to do.

His points about the software "trying to be smarter than the user" are completely idiotic since the very "creature comforts" he might appreciate in some applications are the result of those applications "trying to be smarter than the user".

The problem here is this: people are frustrated that Firefox didn't "just work" when they used it. They have no clue as to why it didn't but it bugged the hell out of them that it didn't. All they wanted is for Firefox to work and really don't care how or why but only that it works. Sure, a simple and easy "workaround" is to uncheck a setting but apparently doing that is a bit too much to ask. So, they want Firefox "fixed" because Firefox appears "not to work".

Now, at face value I can certainly understand that point of view and can see why someone would be frustrated. However to attempt to base a technical argument on an emotional reaction is (what's a nice way to phrase this...) not the best approach to take. Even in that bug report, the bulk of the discussion was about Network Manager and why? Because Network Manager was the problem and the Firefox developers were gracious enough to accomodate the frustrated users.

Think about the various browsers you've used over the years. How many had configuration options relating directly to the underlying networking of the operating environment? Let's see, Internet Explorer has the "Internet Options" configuration dialog but that's really at the Windows level since you can access it through Control Panel. Ok, just about all non-IE browsers have some kind of proxy setting where you can configure a direct connection to the network or configure a proxy server or have it detect a proxy, if necessary. Firefox provides the "about:config" interface to allow you to tweak a lot of the internal settings of the browser for special cases. If those were all "common" kinds of settings, they would be available in the main configuration interface. But I digress....

I think post #95 is more appropriate. A "band-aid" was provided by the Firefox developers. As people choose to either stop using Network Manager on their system or migrate to the newer versions as they come out, the need for this "band-aid" will go away.

Peace...

[lotuseclat79]

The original problem still stands - i.e. FF developers lack a strong QA and decision matrix in the area of what features and dependencies to include without understanding the side effects on the total user population of their decision.

No self respecting competent QA organization would have let FF out the door under the circumstances of depending on a non-existent future version of software (NM v7).

Needless to say it is a daunting open-software project that carries with it the pressures of release schedules.

Knowing when to back-off is as important as knowing when to include - neither of which met a minimum standard IMHO for good sound software engineering - and I've been doing SE for several decades!

But, don't take my word for it - but, do take the user's word for it - they are the lifeline of any really good software product, and will always let you know one way or the other.

Case of developers not in touch with users - one fact stands out, and that is that the very last FF 3.0 RC2 did not have that feature, which begs the question of why they included it under the mistaken assumptions they made without exploring the consequences and missing the opportunity to get meaningful user feedback.

Chalk this one up to experience gained (hopefully) by inexperienced developers/managers making the decisions.

-- Tom

[tomdkat]

Originally Posted by lotuseclat79

The original problem still stands - i.e. FF developers lack a strong QA and decision matrix in the area of what features and dependencies to include without understanding the side effects on the total user population of their decision.

This is certainly NOT the original problem but it certainly is more of a broader spectrum issue. You claim Firefox developers lack a "strong QA" yet you probably know very little about it. I don't know much about the internals of their QA process or procedures but from snipets of comments I've read here and there (mainly in some developer blogs I stumbled across in links on various forums) the Firefox QA process is stronger than you might think. Let's keep things in perspective here: we're talking about an obscure case of people on dial-up or other PPP-oriented Internet connections using Network Manager (through their Linux distribution of choice) which is interfering with Firefox's ability to go online due to a bug in Network Manager. You're saying you would have preferred the ENTIRE Firefox 3 release be held up because of this? What about Windows and Mac users who don't fit into this category? Are the Linux users impacted by this in the majority of Linux network configurations? Of course, I have NO idea but I would venture to speculate there are far MORE Linux users with broadband Internet connections than without. So, the Firefox QA process is "weak" because an obscure minority of users was hit by a bug in another software component? Sounds like an emotional argument to me.....

No self respecting competent QA organization would have let FF out the door under the circumstances of depending on a non-existent future version of software (NM v7).

The same could be said about Microsoft, Apple, Adobe, and just about any other software vendor that develops an app in use today. What I find most ironic about this entire discussion we've been having is your refusal to place blame where it truly lies: the Linux distribution maintainer. Firefox can run fine without Network Manager installed. Firefox does not require Network Manager to function. This is a fact. Here is another fact: the Ubuntu maintainers chose to include Network Manager in their distribution. Other distro maintainers did the same. Since we're talking about Firefox on Ubuntu, let's focus on Ubuntu. The Ubuntu maintainers are the ones who need to perform the system and integration testing to make sure their "package" works as intended. Why aren't you asking why the Ubuntu maintainers didn't more thoroughly test the distro before releasing it? Here is another fact: the Ubuntu maintainers included a beta version of Firefox as the default browser in their GA release of the Ubuntu. Why would they do that? I can see making FF3b available as an additional browser, like Epiphany, but to make it the default? Doesn't make much sense to me.

So, if you're wanting to get to the "heart" or the "root" of the matter, you really need to focus your energy toward the Ubuntu maintainers (in this case). They are the ones who need to make sure ALL of the software they bundle works correctly since they provide the package (the distro) that we, the end users, will use. Why aren't you questioning the QA process of Ubuntu?

Needless to say it is a daunting open-software project that carries with it the pressures of release schedules.

Yep, to stay competitive they must make their dates since we, the users, don't have the patience to wait. Fortunately, the Firefox development and release schedule is pretty transparent. We don't have access to everything but we can see their time lines and know when slips occur, etc.

Knowing when to back-off is as important as knowing when to include - neither of which met a minimum standard IMHO for good sound software engineering - and I've been doing SE for several decades!

I do agree that it's NOT good "cram stuff in" for the sake of getting it in. This makes me question some of the decisions the Ubuntu maintainers made.

But, don't take my word for it - but, do take the user's word for it - they are the lifeline of any really good software product, and will always let you know one way or the other.

I do take the users' word for it. Tons of them love FF3 because it works and works very well. Others have issues with stability on their particular systems. Others have issues with extentions causing grief.

Case of developers not in touch with users - one fact stands out, and that is that the very last FF 3.0 RC2 did not have that feature, which begs the question of why they included it under the mistaken assumptions they made without exploring the consequences and missing the opportunity to get meaningful user feedback.

I don't believe this is a fact. Go back and read the links I posted above. People impacted by this Network Manager problem encountered is with FFb3 and b4. Even you stated you had the problem with b5. The inclusion of D-bus support is not something that was added after RC2 came out. I think that is a fact and one that you've supported with your previous comments in this thread.

As for developers not being in touch with users, I think the Firefox developers are very much in touch with its user base. They might not agree with everything the users have to say and they realize the users really understand far less about how Firefox is developed or even works than the users themselves realize but they tend to be responsive to the user community and I think they do a great job of it!

Chalk this one up to experience gained (hopefully) by inexperienced developers/managers making the decisions.

It certainly will be interesting to see how the Ubuntu maintainers deal with this kind of thing in future releases. I certainly hope the inexperienced users also learn from this and do a more to learn about the software they are using and the system configuration they are using so they can use their systems in ways that suit them the best.

Peace...

[lotuseclat79]

Hi tomdkat,

Apparently, whatever fix they installed in FF 3.0.1- it does not work on my setup. Back to the work-around for me.

Never implied entire FF release should be held up, just that whatever changes they made between RC2 and 3.0 final release should have been backed off until they understood all the ramifications of releasing not-fully-developed future NM software capability (which they "thought" they could - basically, an incomplete set of assumptions must have been in play, or a lack of testing in QC/QA to not spot the problems, or the pressure of schedules drove them to it).

Also, we are not talking about "an obsure dialup use case" here - what we are talking about is the great malaise that all developers at one time or another suffer from until they gain real expertise and experience "at the system level" - i.e. hubris - they "think" they either know it all, or no one else does, even before they have learned to think. It is no wonder that software development is as poor as it is these days.

It is a simple matter of human failure that if there had been a formal specification and means of verification that it was correct, complete and the implementation met the specification - but, I digress - the water has fallen over the dam, so, the most rational thing the developers could have done is to back off the change, the locus of which was centrally located around the use of dbus and NM signals that they clearly did not understand the full ramifications about.

Really well-run development projects in software engineering do at least two major things:
1) They document design decisions (i.e. the reason for them), and if the reason gets stuffed by some lame decision later on, then the design decision can be reconsidered in light of reality - usually changing specifications.
2) They do not make decisions by committee (as it appears from the discussion, not that the discussion does not represent valuable points-of-view) - someone has to have a wider perspective - usually, the system engineer of a project.

-- Tom

[tomdkat]

Originally Posted by lotuseclat79

Apparently, whatever fix they installed in FF 3.0.1- it does not work on my setup. Back to the work-around for me.

Wait a minute, above when you posted the link to the bug report indicating the Firefox side-effect had been dealt with in 3.0.1, you indicated that you backed off your Network Manager change and "all is well again", to quote you directly. Was it working at some point and then it stopped? Also, there was an Ubuntu update of Firefox (or something Firefox related) the other day. How do you pick up Ubuntu updates, given your configuration? This is the UA string for the version I'm running right now:

Mozilla/5.0 (X11; U; Linux x86_64; en-US; rv:1.9.0.1) Gecko/2008071717 Firefox/3.0.1

Ubuntu released two Firefox 3.0.1 updates. One came out right before 3.0.1 went GA and then one went out after 3.0.1 was GA.

Never implied entire FF release should be held up, just that whatever changes they made between RC2 and 3.0 final release should have been backed off until they understood all the ramifications of releasing not-fully-developed future NM software capability (which they "thought" they could - basically, an incomplete set of assumptions must have been in play, or a lack of testing in QC/QA to not spot the problems, or the pressure of schedules drove them to it).

"not-fully-developed future NM software capability"? What the hell does *that* mean? NM's deficiencies are not the responsibility of the Firefox developers. It's the responsibility of the NM developers and NM is part of the GNOME project. As I read more about this, it looks like Firefox 3 introduced support of D-Bus on Linux. You can also read this. Sounds reasonable. Now, according to this:

nd since the NetworkManager components communicate with each other using dbus, it's easy to to build network-aware applications in languages like C, C++, and Python.

I understand Firefox to be simply asking the "system", via D-Bus, what the current network connection state is. Network Manager responds via D-Bus with an incorrect state for those with PPP connections. Now, we know Firefox is completely capable of running WITHOUT Network Manager being installed so we know Network Manager itself is not a requirement. Network Manager is "getting in the way" for PPP connection users, obviously. The comments you make above apply directly to the Ubuntu maintainers, not the Firefox developers. The Ubuntu maintainers provide the software its users can use to support dial-up Internet connectivity. The Ubuntu maintainers are the ones who integrate the software and get the various parts communicating with each other. Your comments are completely misdirected.

Also, we are not talking about "an obsure dialup use case" here - what we are talking about is the great malaise that all developers at one time or another suffer from until they gain real expertise and experience "at the system level" - i.e. hubris - they "think" they either know it all, or no one else does, even before they have learned to think. It is no wonder that software development is as poor as it is these days.

Looking at the specifics of this case, we certainly ARE talking about an obscure minority of Linux users. Look at the Firefox install base and the numbers of people with various Linux distros who complained about it. In the grand scheme, these numbers will minuscule. Taking a step back and looking at this from a broader perspective, I think you're right. Some developers think they "know it all" and that's not necessarily a good thing. I also do agree software development quality has suffered over the years but for a different reason or set of reasons.

It is a simple matter of human failure that if there had been a formal specification and means of verification that it was correct, complete and the implementation met the specification - but, I digress - the water has fallen over the dam, so, the most rational thing the developers could have done is to back off the change, the locus of which was centrally located around the use of dbus and NM signals that they clearly did not understand the full ramifications about.

I agree with your comments except the responsibility for doing what you describe falls on the Ubuntu maintainers, not the Firefox developers. Now, for some bizarre reason you're convinced the D-Bus support was "added" or something post Firefox 3rc2. You yourself have indicated YOU experienced the problem with Firefox 3b5. I've provided documentation of OTHERS reporting the problem in Firefox 3b3 and Firefox 3b4. It's clear to me that the D-Bus support was in place during the BETA cycles of Firefox 3 and was LEFT IN during the release candidate cycles. What makes you think something "changed" in relation to this post FFrc2? What change do you think was made?

Really well-run development projects in software engineering do at least two major things:
1) They document design decisions (i.e. the reason for them), and if the reason gets stuffed by some lame decision later on, then the design decision can be reconsidered in light of reality - usually changing specifications.
2) They do not make decisions by committee (as it appears from the discussion, not that the discussion does not represent valuable points-of-view) - someone has to have a wider perspective - usually, the system engineer of a project.

I agree with both of these things. I do believe the Firefox development team implements both of these things. I know they publicly publish a roadmap and they publicly publish info on their QA processes and so forth. I'm not sure if they publicly publish any of the internal design documentation. I guess you could subscribe to a developer mailing list to see what kind of information might be disseminated there.

One last thing, given the fact the Network Manager issue plagues other applications, and not "just" Firefox, why are you refusing to look at the Ubuntu distro maintainers? In the end, it's their choice to include a "broken" Network Manager that is impacting software distributed by them.

EDIT: I just found this on this page:

It should be easy to configure dial-up and ADSL out of the box, we should use Network Manager to at least bring the connections up and down.

Peace...

[lotuseclat79]

Hi tomdkat,

NM 6.6 is not necessarily broken - just incomplete in terms of development as a version.

Ubuntu distributed the latest version of the software for NM. Obviously, the FF developers may have concluded it to be full featured (when in fact it was not).

Yes, I spoke too soon about all being well - as after I backed off the workaround and rebooted, the plague of offline reoccurred.

I suspect it has more to do with my Live CD environment which is not a full featured Ubuntu installation.

Yes, I have seen the FF roadmap, but strangely, was never able to find a schedule to which they would commit until it was announced about a month or so before release.

Ubuntu 8.04 release preceded the FF 3.0 release.

Not surprising - most software is released with bugs in it.

I've gotten over it - have you?

BTW, if I experienced the problem during FF 3.0b5 (as you say), I probably had the workaround in place as soon as I found it and put it in place - looks like June 2, 2008 by the date on the file I have with the workaround.

You talk about an obscure minority of FF users as if you have access to data that says so - if so, what's your data? Sounds like an assumption to me, so I will call it that. Just so you know, I call 'em as I see them!

-- Tom