Actually, that's the general policy at this point for support from the Mozilla Corporation. They will support product X until 6 months after product X + 1 ships.
This is not to say that other providers of Firefox can't do longer-term support. Just that the Mozilla Corporation will not be doing the QA+triage+release work on fixes past that point. The bug database is open; various distros have write access to the revision control system Firefox uses, and they have in fact been maintaining older Firefox releases (1.5, say) past the Mozilla Corporation end-of-support.
If you think the constitution should be amended, that's fine. Let's do it. I might even agree.
But paying lip service to it while disregarding it in practice is really not called for, especially if you're going to teach the original text to all those poor high school students who will then have no idea how their government really works.
This would absolutely make sense. In fact, a number of countries are much closer to this than the US is (e.g. in many clergy cannot actually handle the government-related paperwork, unlike in the US).
So basically, the government should get out of the marriage business altogether. There is indeed a need for some sort of "license of union" thing, but it's not obvious to me that:
1) All marriages (in the religious/traditional/whatever sense) need qualify 2) It needs to have some of the restrictions that marriage currently does (e.g. requiring
that the people involved not be too closely related, in addition to the gender issues
under discussion). 3) It needs to be restricted to two-person partnerships.
There is the obvious question of why the government should issue any such licenses at all. What is the inheritance special-case, and why (and when!) is it in society's benefit to special-case it? What are the tax benefits, and why should they exist? What are the decision benefits, and how do they differ from explicitly granted limited powers of attorney?
A lot of the current treatment of marriage basically revolves around treating the married couple as a single economic unit, and there are obvious benefits to that approach where child-rearing is concerned. Historically, of course, such economic integration is the whole point of marriage. It's worth reexamining when we want the government to encourage it nowadays.
Really? Say that's true. You make 100k a year and have a family of 4. You're like to have a 3-bedroom residence for said family (so each of the kids can have a separate room). You live in the Boston area.
You have 50k after taxes. Your rent, electricity, gas, heat, water add up to $2500 a month (not at all unreasonable in the non-crime-ridden parts of the Boston area for a 3-bedroom apartment). Your phone+cable+internet is about $100 per month. Your food expenses are $500 a month (at current food prices, that need not involve any eating out). You drive, most likely, and almost certainly so does your spouse. Figuring 150 miles per week (which is a low estimate for most commutes, and assumes your wife isn't spending on gas), and assuming you get 50 miles per gallon, you're looking at $10 or so a week, or $500 or the year. You're probably looking at $2000 year or more for car insurance. You're probably also looking at some amount of time actually buying and maintaining said cars; let's say you buy slightly used (not new) at a decent price every 7 years, and do no maintenance, for about $1000 a year.
Let's total it up: $30,000 housing and utilities $6000 food $1200 phone+internet+cable $2000 car insurance $500 gas $1000 car purchasing
Total: $40,700
Things we didn't include here: saving for retirement, saving for college for the kids, saving to buy a house if you rent, purchasing clothing (if you buy kids clothes new figure at least another $1000 or more a year, since they grow out of it), health insurance (easily another several grand a year for employer-subsidized plans; 10+ grand if you pay it all yourself), visiting grandparents (plane flight for four is $1000+; driving is cheaper, of course), any sort of vacations, any sort of music lessons for the kids or the like. I assumed no medical bills, no dental bills, no orthodontia for the kids. No furniture purchases. No daycare for the kids, if they're young. No presents you buy. No church/synagogue/whatever dues.
$100k is great if you're single, but there are already parts of the country where it's pretty hard to maintain a middle-class lifestyle on that sort of household income even now, with taxes being closer to 30% effective (20% federal, 7% FICA, some state/local) than your 50% number.
Of course if your housing is not insane (equivalent housing in other parts of the country could easily run 10k a year instead of 30k), things are a lot better...
It's a good trade. You do that, the U.S. pulls out of Germany and drops this missile shield in Poland boondoggle. And pulls out of South Korea too.
Oh, wait, some of those folks keep wanting the U.S. to stay, and people here keep muttering about treaty commitments.
It's nice to have it both ways: get someone else to spend money on your defense and shit on them at the same time. Too bad the U.S. lets people get away with this.
That said, a lot of the things young liberals push for are seen as "weird" by older folks, more or less for inertia reasons. That probably contributes to the effect you see.
I sure hope that in 50 years people will be amazed that the government was in any way involved with the marriage-as-romantic-involvement business at all. Involvement in setting up an economic unit for the purpose of raising children is a different issue entirely, of course, but that one need not have any romantic overtones.
That is, have a civil setup for raising kids and a whatever-the-heck-you-want-to-do-privately setup for romantic commitment.
Anyone running for president nowadays is pretty much batshit insane by definition (at best they're a power-hungry sociopath).
The thing we should try to make sure of is that the same party never controls both congress and the white house. That was the real problem with most of Bush's presidency, in my opinion, just like it was a problem at the beginning of Clinton's. The fewer laws Washington can add, the better.
It would be even better if they'd start to repeal some laws (I'm looking at http://en.wikipedia.org/wiki/Guano_Islands_Act as a good place to start here), but there's slim hope of that. Heck, the US tax code grew by something like 50% in the last 20 years... as if it had not been complex enough in 1988.
> Which as I said means it's still there and needs to be fixed.
That's fine, but a hard-to-trigger bug automatically gets lower priority than an easy-to-trigger bug. Surely you agree with that?
> So I just happened to be the guy...right?
All I said is that neither I, nor any other developer I've talked to, has been able to reproduce it recently. We're well aware that a number of people (small in the grand scheme of things, but certainly far from zero) are still able to reproduce the problem. I'd love any pointers you may have on doing so.
The cause is obviously that the mouseup event gets delivered to one of the menu items. That much is clear. What's not clear, from code inspection, is how this could happen: the top-left of the menu is positioned below and to the right of the right-click point before the menu is mapped, last I checked.
Again, this is a real problem, we'd love to fix it, we haven't been able to figure out a way it could possibly happen from code inspection, and we haven't been able to reproduce so as to be able to debug.
> Reproduce it reliably? Different setups? Please. Just do a right click at most ten times > in a row, it kicked in.
I did just that. No bug.
Seriously, we've been trying to actually reproduce this for months, with very little luck.
> What do you mean it "went away for you"?
I mean that it stopped appearing. In late alphas I'd get it about once a day of use. After the first beta, I never saw it.
> If you saw it in the betas, then it was a known bug.
Indeed.
> And since it is a SERIOUS bug, the release should have been held up until it was > reproduced and fixed.
Except all the people who used to be able to reproduce it could no longer do so with the later builds. It's really hard to hold a release on a bug that no one can actually get to happen...
> And now they want to replace the entire JavaScript engine? How do you figure they're > going to catch all THOSE bugs if they can't catch a simple right click bug?
Javascript engines, unlike UI, are very amenable to automated correctness testing.
> How the HELL do you claim to do ANY QA when you release a "final" with that kind of easily > found, major bug?
Because you don't figure out a way to reproduce it reliably, and because it doesn't seem to appear in all setups or affect all users...
I know that I was hitting it for a while in the betas, then it went away for me. Which was unfortunate, because it's really hard to fix a bug you can't reproduce.
> that emit real native code (not CALL streams) for entire functions (or even larger chunks > of code, with inlining).
A small nit: Tracemonkey doesn't compile entire functions. It compiles instruction streams, which means that if your function has a branch that's never taken that part of the branch will never get compiled.
> This is necessary to enable traditional optimizations like register allocation, > instruction scheduling, constant folding, loop unrolling etc
There's a tradeoff here, though, which is that you want the compilation to be fast (as in, avoiding the Java "let's be slow while we start up and jit all this code" syndrome). The faster your compiler needs to be, the less room you have to fancy optimizations. Register allocation, yes (helped in tracemonkey's case by the fact that it's all in SSA form already). Constant folding, yes (actually happens on the interpreter level too). Instruction scheduling, maybe not so much.
I agree with your main point that V8 and Tracemonkey can get faster on these benchmarks. I'm just not convinced they'll necessarily end up that much faster than SFX.
> What country does IE have less then 50% of the market?
As of June 2008 (before the Firefox 3 launch), Firefox had >50% market share in Indonesia.
As of March 2008, Firefox had 46% market share in Finland, 44% in Poland, 43% in Slovenia, 43% in Slovakia, 41% in Hungary. I don't know how much (if at all) those numbers have gone up since, but if Firefox is at 46% it doesn't take much safari/opera/whatever usage to push IE below 50%.
There's no need to strip the code down to file a bug report. If you do, it's very much appreciated, but if you don't someone else might. That's the beauty of an open bug database.;)
Thanks for the links; it looks like at least the -anim version of that raytracer was mentioned on Slashdot before and is covered by https://bugzilla.mozilla.org/show_bug.cgi?id=460964. I've added pointers to the other two pages to that bug report.
It might not be a good plan, but nevertheless people do it. Or more precisely they create a baseline user experience that works for all users and enhance it as they can. Sometimes that means IE-only features, sometimes not.
> IE is the most common browser and that is just a fact we all have to deal with.
That strongly depends on your user base. There are plenty of sites where this is in fact no longer the case. Heck, there are entire countries where this is no longer the case.
The problem with that bug is that the Xorg graphics performance stuff... sucks. A lot. A lot of rendering is actually faster done entirely in software than trying to go through X's "accelerated" stuff. Some of this is due to Render shipping stone-age versions of pixman and actually doing its own software rasterization when you'd think it would use your graphics hardware.
And best of all, the response of the Xorg developers to all this is "once we finish our new acceleration architecture in a few years, all this should be better". They've said it for quite a while now, with at least 2 different acceleration architectures.
So yes, Mozilla does give "a flying...." about Linux, but that doesn't help much in this situation.
> if you type something in the address field, it gets sent off to Google as a search entry
This is preference-controlled, of course. Both the behavior and the search engine used. The default preferences are set the way they are because that's what most people seem to want (in the US; the preferences differ in other locales). If you want something else, you can change it.
> "Block popups (except from Google, etc...)"
Uh... say what? The popup blocker exception list ships empty by default. If yours is nonempty, then you either didn't get your browser from mozilla.org (and whoever you got it from changed the defaults), or at some point you unblocked popups from some sites.
> make me feel that they are being dishonest and that is not a nice feeling
I think mozilla.org has been very upfront about everything here... I'm sorry you feel the way you do, but I think you have a very wrong impression of the way things are working.
> they could have said up front that this is a vehicle to promote Google without losing > anything
The x86-64 port was done in a summer by an intern...
So yeah, it's harder than just a recompile, but not that hard.
I should also note that lack of jit support for an ISA just means no jit and using the interpreter all the time and therefore you're no worse than the (jitless) status quo. Pure interpreter performance continues to be a priority, since not everything is being jitted.
> grinds to inexplicable halts kinda randomly whenever my code does anything repetitive and > strenuous,
Would you mind filing a bug with a pointer to a web page that shows this problem? That sounds like something we should fix, but it's hard to do that given the information presented here.
You do realize that Google maps uses SVG for some of its stuff, right? In IE they just use VRML or something like that that IE has support for instead...
Other webapps use SVG in UAs that support it and just deliver a degraded user experience in IE.
Slashdot Mirror
Actually, that's the general policy at this point for support from the Mozilla Corporation. They will support product X until 6 months after product X + 1 ships.
This is not to say that other providers of Firefox can't do longer-term support. Just that the Mozilla Corporation will not be doing the QA+triage+release work on fixes past that point. The bug database is open; various distros have write access to the revision control system Firefox uses, and they have in fact been maintaining older Firefox releases (1.5, say) past the Mozilla Corporation end-of-support.
> Now while I don't expect the 2.x branch to have any security compromising problems
You probably should. It's shipping security fixes on a regular schedule so far, no?
If you think the constitution should be amended, that's fine. Let's do it. I might even agree.
But paying lip service to it while disregarding it in practice is really not called for, especially if you're going to teach the original text to all those poor high school students who will then have no idea how their government really works.
This would absolutely make sense. In fact, a number of countries are much closer to this than the US is (e.g. in many clergy cannot actually handle the government-related paperwork, unlike in the US).
So basically, the government should get out of the marriage business altogether. There is indeed a need for some sort of "license of union" thing, but it's not obvious to me that:
1) All marriages (in the religious/traditional/whatever sense) need qualify
2) It needs to have some of the restrictions that marriage currently does (e.g. requiring
that the people involved not be too closely related, in addition to the gender issues
under discussion).
3) It needs to be restricted to two-person partnerships.
There is the obvious question of why the government should issue any such licenses at all. What is the inheritance special-case, and why (and when!) is it in society's benefit to special-case it? What are the tax benefits, and why should they exist? What are the decision benefits, and how do they differ from explicitly granted limited powers of attorney?
A lot of the current treatment of marriage basically revolves around treating the married couple as a single economic unit, and there are obvious benefits to that approach where child-rearing is concerned. Historically, of course, such economic integration is the whole point of marriage. It's worth reexamining when we want the government to encourage it nowadays.
Really? Say that's true. You make 100k a year and have a family of 4. You're like to have a 3-bedroom residence for said family (so each of the kids can have a separate room). You live in the Boston area.
You have 50k after taxes. Your rent, electricity, gas, heat, water add up to $2500 a month (not at all unreasonable in the non-crime-ridden parts of the Boston area for a 3-bedroom apartment). Your phone+cable+internet is about $100 per month. Your food expenses are $500 a month (at current food prices, that need not involve any eating out). You drive, most likely, and almost certainly so does your spouse. Figuring 150 miles per week (which is a low estimate for most commutes, and assumes your wife isn't spending on gas), and assuming you get 50 miles per gallon, you're looking at $10 or so a week, or $500 or the year. You're probably looking at $2000 year or more for car insurance. You're probably also looking at some amount of time actually buying and maintaining said cars; let's say you buy slightly used (not new) at a decent price every 7 years, and do no maintenance, for about $1000 a year.
Let's total it up:
$30,000 housing and utilities
$6000 food
$1200 phone+internet+cable
$2000 car insurance
$500 gas
$1000 car purchasing
Total: $40,700
Things we didn't include here: saving for retirement, saving for college for the kids, saving to buy a house if you rent, purchasing clothing (if you buy kids clothes new figure at least another $1000 or more a year, since they grow out of it), health insurance (easily another several grand a year for employer-subsidized plans; 10+ grand if you pay it all yourself), visiting grandparents (plane flight for four is $1000+; driving is cheaper, of course), any sort of vacations, any sort of music lessons for the kids or the like. I assumed no medical bills, no dental bills, no orthodontia for the kids. No furniture purchases. No daycare for the kids, if they're young. No presents you buy. No church/synagogue/whatever dues.
$100k is great if you're single, but there are already parts of the country where it's pretty hard to maintain a middle-class lifestyle on that sort of household income even now, with taxes being closer to 30% effective (20% federal, 7% FICA, some state/local) than your 50% number.
Of course if your housing is not insane (equivalent housing in other parts of the country could easily run 10k a year instead of 30k), things are a lot better...
It's a good trade. You do that, the U.S. pulls out of Germany and drops this missile shield in Poland boondoggle. And pulls out of South Korea too.
Oh, wait, some of those folks keep wanting the U.S. to stay, and people here keep muttering about treaty commitments.
It's nice to have it both ways: get someone else to spend money on your defense and shit on them at the same time. Too bad the U.S. lets people get away with this.
Perhaps they're gaining life experience? ;)
That said, a lot of the things young liberals push for are seen as "weird" by older folks, more or less for inertia reasons. That probably contributes to the effect you see.
I sure hope that in 50 years people will be amazed that the government was in any way involved with the marriage-as-romantic-involvement business at all. Involvement in setting up an economic unit for the purpose of raising children is a different issue entirely, of course, but that one need not have any romantic overtones.
That is, have a civil setup for raising kids and a whatever-the-heck-you-want-to-do-privately setup for romantic commitment.
This is the slot machines that they're going to put at tracks and such, so as to try to target people who already have gambling problems?
Anyone running for president nowadays is pretty much batshit insane by definition (at best they're a power-hungry sociopath).
The thing we should try to make sure of is that the same party never controls both congress and the white house. That was the real problem with most of Bush's presidency, in my opinion, just like it was a problem at the beginning of Clinton's. The fewer laws Washington can add, the better.
It would be even better if they'd start to repeal some laws (I'm looking at http://en.wikipedia.org/wiki/Guano_Islands_Act as a good place to start here), but there's slim hope of that. Heck, the US tax code grew by something like 50% in the last 20 years... as if it had not been complex enough in 1988.
> No - severity of effect is what matters.
Sure. Equal severity bugs, of course.
> If the Ubuntu build of 3.0.3 reproduces it repeatedly, you could start from there.
It doesn't in our testing...
> Which as I said means it's still there and needs to be fixed.
That's fine, but a hard-to-trigger bug automatically gets lower priority than an easy-to-trigger bug. Surely you agree with that?
> So I just happened to be the guy...right?
All I said is that neither I, nor any other developer I've talked to, has been able to reproduce it recently. We're well aware that a number of people (small in the grand scheme of things, but certainly far from zero) are still able to reproduce the problem. I'd love any pointers you may have on doing so.
The cause is obviously that the mouseup event gets delivered to one of the menu items. That much is clear. What's not clear, from code inspection, is how this could happen: the top-left of the menu is positioned below and to the right of the right-click point before the menu is mapped, last I checked.
Again, this is a real problem, we'd love to fix it, we haven't been able to figure out a way it could possibly happen from code inspection, and we haven't been able to reproduce so as to be able to debug.
> Reproduce it reliably? Different setups? Please. Just do a right click at most ten times
> in a row, it kicked in.
I did just that. No bug.
Seriously, we've been trying to actually reproduce this for months, with very little luck.
> What do you mean it "went away for you"?
I mean that it stopped appearing. In late alphas I'd get it about once a day of use. After the first beta, I never saw it.
> If you saw it in the betas, then it was a known bug.
Indeed.
> And since it is a SERIOUS bug, the release should have been held up until it was
> reproduced and fixed.
Except all the people who used to be able to reproduce it could no longer do so with the later builds. It's really hard to hold a release on a bug that no one can actually get to happen...
> And now they want to replace the entire JavaScript engine? How do you figure they're
> going to catch all THOSE bugs if they can't catch a simple right click bug?
Javascript engines, unlike UI, are very amenable to automated correctness testing.
> Sorry - but Firefox has become as bad as IE 5.
Thinking that is your prerogative, of course.
Yeah, ARM is for the "run on your cell phone" use case.
> How the HELL do you claim to do ANY QA when you release a "final" with that kind of easily
> found, major bug?
Because you don't figure out a way to reproduce it reliably, and because it doesn't seem to appear in all setups or affect all users...
I know that I was hitting it for a while in the betas, then it went away for me. Which was unfortunate, because it's really hard to fix a bug you can't reproduce.
> that emit real native code (not CALL streams) for entire functions (or even larger chunks
> of code, with inlining).
A small nit: Tracemonkey doesn't compile entire functions. It compiles instruction streams, which means that if your function has a branch that's never taken that part of the branch will never get compiled.
> This is necessary to enable traditional optimizations like register allocation,
> instruction scheduling, constant folding, loop unrolling etc
There's a tradeoff here, though, which is that you want the compilation to be fast (as in, avoiding the Java "let's be slow while we start up and jit all this code" syndrome). The faster your compiler needs to be, the less room you have to fancy optimizations. Register allocation, yes (helped in tracemonkey's case by the fact that it's all in SSA form already). Constant folding, yes (actually happens on the interpreter level too). Instruction scheduling, maybe not so much.
I agree with your main point that V8 and Tracemonkey can get faster on these benchmarks. I'm just not convinced they'll necessarily end up that much faster than SFX.
Uh... How so? The tracing jit only traces loops, so non-loopy straight-line code is interpreted no matter what.
So you can't "drop the interpreter".
> What country does IE have less then 50% of the market?
As of June 2008 (before the Firefox 3 launch), Firefox had >50% market share in Indonesia.
As of March 2008, Firefox had 46% market share in Finland, 44% in Poland, 43% in Slovenia, 43% in Slovakia, 41% in Hungary. I don't know how much (if at all) those numbers have gone up since, but if Firefox is at 46% it doesn't take much safari/opera/whatever usage to push IE below 50%.
There's no need to strip the code down to file a bug report. If you do, it's very much appreciated, but if you don't someone else might. That's the beauty of an open bug database. ;)
Thanks for the links; it looks like at least the -anim version of that raytracer was mentioned on Slashdot before and is covered by https://bugzilla.mozilla.org/show_bug.cgi?id=460964. I've added pointers to the other two pages to that bug report.
It might not be a good plan, but nevertheless people do it. Or more precisely they create a baseline user experience that works for all users and enhance it as they can. Sometimes that means IE-only features, sometimes not.
> IE is the most common browser and that is just a fact we all have to deal with.
That strongly depends on your user base. There are plenty of sites where this is in fact no longer the case. Heck, there are entire countries where this is no longer the case.
The problem with that bug is that the Xorg graphics performance stuff... sucks. A lot. A lot of rendering is actually faster done entirely in software than trying to go through X's "accelerated" stuff. Some of this is due to Render shipping stone-age versions of pixman and actually doing its own software rasterization when you'd think it would use your graphics hardware.
And best of all, the response of the Xorg developers to all this is "once we finish our new acceleration architecture in a few years, all this should be better". They've said it for quite a while now, with at least 2 different acceleration architectures.
So yes, Mozilla does give "a flying ...." about Linux, but that doesn't help much in this situation.
> if you type something in the address field, it gets sent off to Google as a search entry
This is preference-controlled, of course. Both the behavior and the search engine used. The default preferences are set the way they are because that's what most people seem to want (in the US; the preferences differ in other locales). If you want something else, you can change it.
> "Block popups (except from Google, etc ...)"
Uh... say what? The popup blocker exception list ships empty by default. If yours is nonempty, then you either didn't get your browser from mozilla.org (and whoever you got it from changed the defaults), or at some point you unblocked popups from some sites.
> make me feel that they are being dishonest and that is not a nice feeling
I think mozilla.org has been very upfront about everything here... I'm sorry you feel the way you do, but I think you have a very wrong impression of the way things are working.
> they could have said up front that this is a vehicle to promote Google without losing
> anything
Except of course that it's not such a vehicle.
The x86-64 port was done in a summer by an intern...
So yeah, it's harder than just a recompile, but not that hard.
I should also note that lack of jit support for an ISA just means no jit and using the interpreter all the time and therefore you're no worse than the (jitless) status quo. Pure interpreter performance continues to be a priority, since not everything is being jitted.
> grinds to inexplicable halts kinda randomly whenever my code does anything repetitive and
> strenuous,
Would you mind filing a bug with a pointer to a web page that shows this problem? That sounds like something we should fix, but it's hard to do that given the information presented here.
You do realize that Google maps uses SVG for some of its stuff, right? In IE they just use VRML or something like that that IE has support for instead...
Other webapps use SVG in UAs that support it and just deliver a degraded user experience in IE.