Unfortunately, doing that would sort of imply Mozilla would need to vouch for the extension developers (hey, they're letting them use a cert; that's what it's for, right?). As it is they barely have enough people to just try installing extensions before approving for the main site...
If it's just extension updates anyway, and extensions already act as a part of Firefox (i.e. they're not sandboxed... which they can't be in the current architecture)... They might as well just require SSL for updates, and people who don't use the Mozilla update service can just ship their own (self-signed) cert with the extension. Of course, some authors will still work around that by doing their own thing anyway. (There were, at one point, very, very insecure extensions that... load the whole toolbar at runtime using eval() by pulling data from unsecured sites.)
Hmm, are you talking about login information (mainly, session and permanent cookies)? All tabs/windows share the cookies, so that sort of sounds like it. So, for example, you can't log in into two different instances of Gmail at once. But I have no idea how this could ever have been broken before - cookies shouldn't be global, ever...
(I'd happily take a bugzilla.mozilla.org bug number, or a usenet post, or anything of the sort I can use to learn more about this, if you'd be kind enough to oblige.)
To be more correct, Firefox (well, Gecko) is largely single-threaded. All of the UI, DOM, and JavaScript [*] happen on the same thread. This particular case sounds like it's trying to lay out the page in the non-visible tabs, in which case there is indeed nothing you can do other than start a new process, due to Gecko limitations.
[*] For internal code, it is possible to use threads in a very small subset of JS (in particular, XPCOM components that don't have any interaction at all with the UI). That won't help in this case.
Hmm, can you describe the shared namespace thing in a bit more detail, preferably with a test case (two pages on different domains I can open in tabs to see the interaction)? My understanding was that each tab was basically its own content window, and are completely unrelated...
There was something about assigning extra things to prototypes, but that doesn't actually happen as far as I can tell - I've only seen it do that when using Firebug on a site with prototype (i.e. things leak into Firebug), and that'd be an (unconfirmed, since I didn't look very hard at it) Firebug bug, not a Web 2.0 one.
keyconfig should be able to do what you want. Of course, staying with Opera is a valid option too; choice is good. But Firefox is not Opera, nor is Opera Firefox.:) (So "make it be exactly the same as Opera" isn't exactly something they can do)
The scrollbar thing, I think, is just a Mozilla bug and doesn't really have much to do with native form widgets. I believe it's supposed to be a bit better in Firefox 3 too.
If you want to force native widgets to go away, use the vendor-specific CSS -moz-appearance: none !important;
Mozilla's always had a little less love than Win32 and Linux, but it is slowly getting better - I guess having a bunch of developers run Macs helped:)
Around the IE4 time frame, an equivalent feature to bookmark keywords did exist - description. I have no idea where to get that anymore though, neither do I know if it still works...
What are the warnings on herbs and vitamins? None! If you take too much the most you'll get is a tummyache.
Is definitely incorrect. Chinese medicine taken irresponsibly certainly can cause things like death. Of course, to do it properly you would need to get it prescribed by a doctor (trained in this stuff rather than the western stuff, of course), who will typically do things like ask about your condition, and take your pulse. (The pulse part involves more than just frequency, by the way.)
The stuff that can get you up to a tummyache? Some of it might just be... random herbs. Will they cure you? Probably not.
/3GB doesn't do that. It lets apps allocate up to 3GB of data (and the kernel 1GB instead), but that's a virtual address thing - how much physical memory the OS sees isn't related at all.
I hope the GP had verified in some way that the machine, somewhere, did see the full 4GB. (Probably during POST.) As opposed to one stick just being bad...
If your router supports UPNP, that might be it - it broadcasts stuff to let the computers know that there's a UPNP capable device there. In Wireshark they seem to show up as SSDP.
I've also had one that would randomly ping computers - I guess it's checking if they're alive or something.
application/x-mplayer2 is both the old WMP64 and the new WMP9 plugin; mplayerplug-in (Linux) also supports that content type. If you want to force the new plugin only, application/x-ms-wmp. There's no need though, of course.
Actually, the plugin itself (np-mswmp.dll) works just fine with WMP9. (WinXP SP2 machine here.) Of course, I don't like installers and extracted the file manually (7zip+WiX)... So I don't even know if the installer does any checking.
Well, the release notes (the text file) mentioned that Firefox proxy settings won't take effect, you need to adjust the IE proxy settings for the plugin to see it...
Umm, Firefox 3 will be using Cairo, but I have no idea where the "fully based on Vectors" comes from. (Also, Firefox 1.5.x is Gecko 1.8.0; Firefox 2.x is Gecko 1.8.1)
If there's native support for OpenID, I haven't seen it yet:)
And in the improving-the-web direction, you basically want to look at WHATWG anyway - at least Mozilla, Opera, and Apple are behind it, so even if IE isn't there the other major desktop players are. And the new-ish HTML WG at W3C...
Re:First to include the other as a plugin:
on
Firefox 3.0 Preview
·
· Score: 3, Informative
1) Umm, that killer app must also run in Windows, or it's a bug in ReactOS (since they're trying to recreate Windows), no?
2) Windows 2000+ does have a Unix compat subsystem. It seems more NetBSD than Linux (uses pkg_add etc, NetBSD has Interix packages that you can use), and no X server just libs. I have no idea if they also emulate fork() or not though.
Or they actual people working on it know DRM doesn't work, but gets told by the boss to do it anyway because somebody else (content providers) want it. And since it's useless anyway, they're fine with writing a half-assed implementation because the extra effort isn't worth it - anything beyond barely passing isn't worth it.
But that doesn't actually give you revenue, it just reduces expenses.
Instead, just forward the user to a page explaining that ad blocking is not supported and link to instructions on whitelisting the ads in yahoo. Then the user at least knows how to fix it, and has the choice of sticking with you...
You need to add 20 to do it in Malbolge and find that language to be perfectly safe because nobody could claim to have something that would possibly ever be deployed, right?:p
I think I have you beat, Firefox (trunk) can even compile on wine with some patches:) (... with make 3.80; there's some bad interaction with make 3.81)
Hmm, can you tell me how to implement multiple inheritance? I've wondered about that... And yes, I do want to keep instanceof working, so just copying the properties/methods over, or even calling the base constructor in the derived constructor, will not work:(
Also, Netscape (now Mozilla) does still keep evolving the language. For example, JS1.7 (Firefox 2 / Mozilla 1.8.1 branch and later). And there's work on ECMA262 ed 4. (Mostly seems to be Mozilla and Adobe folks behind it).
Except that when you visit site B, the browser discards all JS from site A before going to site B. Site B never sees any JS objects from A anyway. Think of the browser-supplied things (e.g. the XMLHttpRequest constructor) as a template; if you modify it you just get a copy of the template for yourself.
If touching prototypes of built-in objects would persist across sites there simply could not have been more than one JS framework system. And nobody would have had scripting enabled...:)
Slashdot Mirror
Unfortunately, doing that would sort of imply Mozilla would need to vouch for the extension developers (hey, they're letting them use a cert; that's what it's for, right?). As it is they barely have enough people to just try installing extensions before approving for the main site...
If it's just extension updates anyway, and extensions already act as a part of Firefox (i.e. they're not sandboxed... which they can't be in the current architecture)... They might as well just require SSL for updates, and people who don't use the Mozilla update service can just ship their own (self-signed) cert with the extension. Of course, some authors will still work around that by doing their own thing anyway. (There were, at one point, very, very insecure extensions that... load the whole toolbar at runtime using eval() by pulling data from unsecured sites.)
Hmm, are you talking about login information (mainly, session and permanent cookies)? All tabs/windows share the cookies, so that sort of sounds like it. So, for example, you can't log in into two different instances of Gmail at once. But I have no idea how this could ever have been broken before - cookies shouldn't be global, ever...
(I'd happily take a bugzilla.mozilla.org bug number, or a usenet post, or anything of the sort I can use to learn more about this, if you'd be kind enough to oblige.)
To be more correct, Firefox (well, Gecko) is largely single-threaded. All of the UI, DOM, and JavaScript [*] happen on the same thread. This particular case sounds like it's trying to lay out the page in the non-visible tabs, in which case there is indeed nothing you can do other than start a new process, due to Gecko limitations.
[*] For internal code, it is possible to use threads in a very small subset of JS (in particular, XPCOM components that don't have any interaction at all with the UI). That won't help in this case.
Hmm, can you describe the shared namespace thing in a bit more detail, preferably with a test case (two pages on different domains I can open in tabs to see the interaction)? My understanding was that each tab was basically its own content window, and are completely unrelated...
There was something about assigning extra things to prototypes, but that doesn't actually happen as far as I can tell - I've only seen it do that when using Firebug on a site with prototype (i.e. things leak into Firebug), and that'd be an (unconfirmed, since I didn't look very hard at it) Firebug bug, not a Web 2.0 one.
keyconfig should be able to do what you want. Of course, staying with Opera is a valid option too; choice is good. But Firefox is not Opera, nor is Opera Firefox. :) (So "make it be exactly the same as Opera" isn't exactly something they can do)
The scrollbar thing, I think, is just a Mozilla bug and doesn't really have much to do with native form widgets. I believe it's supposed to be a bit better in Firefox 3 too.
:)
If you want to force native widgets to go away, use the vendor-specific CSS -moz-appearance: none !important;
Mozilla's always had a little less love than Win32 and Linux, but it is slowly getting better - I guess having a bunch of developers run Macs helped
Around the IE4 time frame, an equivalent feature to bookmark keywords did exist - description. I have no idea where to get that anymore though, neither do I know if it still works...
Is definitely incorrect. Chinese medicine taken irresponsibly certainly can cause things like death. Of course, to do it properly you would need to get it prescribed by a doctor (trained in this stuff rather than the western stuff, of course), who will typically do things like ask about your condition, and take your pulse. (The pulse part involves more than just frequency, by the way.)
The stuff that can get you up to a tummyache? Some of it might just be... random herbs. Will they cure you? Probably not.
/3GB doesn't do that. It lets apps allocate up to 3GB of data (and the kernel 1GB instead), but that's a virtual address thing - how much physical memory the OS sees isn't related at all.
I hope the GP had verified in some way that the machine, somewhere, did see the full 4GB. (Probably during POST.) As opposed to one stick just being bad...
If your router supports UPNP, that might be it - it broadcasts stuff to let the computers know that there's a UPNP capable device there. In Wireshark they seem to show up as SSDP.
I've also had one that would randomly ping computers - I guess it's checking if they're alive or something.
application/x-mplayer2 is both the old WMP64 and the new WMP9 plugin; mplayerplug-in (Linux) also supports that content type. If you want to force the new plugin only, application/x-ms-wmp. There's no need though, of course.
Actually, the plugin itself (np-mswmp.dll) works just fine with WMP9. (WinXP SP2 machine here.) Of course, I don't like installers and extracted the file manually (7zip+WiX)... So I don't even know if the installer does any checking.
Wine + ies4linux + Firefox/win32 + WMP9 = yes it runs. Not exactly performant, of course...
;)
Oh, and the DRM stuff won't install in WMP9
Well, the release notes (the text file) mentioned that Firefox proxy settings won't take effect, you need to adjust the IE proxy settings for the plugin to see it...
Umm, Firefox 3 will be using Cairo, but I have no idea where the "fully based on Vectors" comes from. (Also, Firefox 1.5.x is Gecko 1.8.0; Firefox 2.x is Gecko 1.8.1)
:)
If there's native support for OpenID, I haven't seen it yet
And in the improving-the-web direction, you basically want to look at WHATWG anyway - at least Mozilla, Opera, and Apple are behind it, so even if IE isn't there the other major desktop players are. And the new-ish HTML WG at W3C...
Eclipse won. ATF
1) Umm, that killer app must also run in Windows, or it's a bug in ReactOS (since they're trying to recreate Windows), no?
2) Windows 2000+ does have a Unix compat subsystem. It seems more NetBSD than Linux (uses pkg_add etc, NetBSD has Interix packages that you can use), and no X server just libs. I have no idea if they also emulate fork() or not though.
In the same vein, does it run Interix (Services for Unix / Subsystem for Unix-based Applications)? (It's like cygwin, but runs on top of the NT core itself rather than Win32)
That'd be interesting, even though as far as I can tell it's more NetBSD than Linux.
Or they actual people working on it know DRM doesn't work, but gets told by the boss to do it anyway because somebody else (content providers) want it. And since it's useless anyway, they're fine with writing a half-assed implementation because the extra effort isn't worth it - anything beyond barely passing isn't worth it.
But that doesn't actually give you revenue, it just reduces expenses.
Instead, just forward the user to a page explaining that ad blocking is not supported and link to instructions on whitelisting the ads in yahoo. Then the user at least knows how to fix it, and has the choice of sticking with you...
1) inline-block? (which, btw, Firefox doesn't do yet; might get it in Firefox 3 due to reflow branch landing)
You need to add 20 to do it in Malbolge and find that language to be perfectly safe because nobody could claim to have something that would possibly ever be deployed, right? :p
I think I have you beat, Firefox (trunk) can even compile on wine with some patches :)
(... with make 3.80; there's some bad interaction with make 3.81)
Hmm, can you tell me how to implement multiple inheritance? I've wondered about that... And yes, I do want to keep instanceof working, so just copying the properties/methods over, or even calling the base constructor in the derived constructor, will not work :(
Also, Netscape (now Mozilla) does still keep evolving the language. For example, JS1.7 (Firefox 2 / Mozilla 1.8.1 branch and later). And there's work on ECMA262 ed 4. (Mostly seems to be Mozilla and Adobe folks behind it).
Except that when you visit site B, the browser discards all JS from site A before going to site B. Site B never sees any JS objects from A anyway. Think of the browser-supplied things (e.g. the XMLHttpRequest constructor) as a template; if you modify it you just get a copy of the template for yourself.
:)
If touching prototypes of built-in objects would persist across sites there simply could not have been more than one JS framework system. And nobody would have had scripting enabled...