Yes this is a known bug with GTK+ integration. I agree it's pretty embarassing for them - Firefox and Mozilla had always been stable as a rock and then at the last minute, a multiple-crashes-a-day bug was introduced. I'd have expected an extremely fast bugfix from them but they're apparently too busy printing adverts.
Your university clearly bears no resemblence to mine, where I know only two people who have a Mac (both laptops). One is a friend of mine, and one is a lecturer. I've met more like 12 people who use Linux systems while I've been here.
As already pointed out, this seems to be a US/non-US thing. MacOS has little to no serious desktop penetration in Europe. iPod is a different story.
Final thing - I did not say "Macs are only used in design", please don't misquote me again in such a blatant fashion. I said that if I was in a design environment instead of a CS degree at university, I would probably see more Macs. Where I live this is true.
There are now over 12 million Mac OS X systems in use (source: 23:40 of WWDC keynote). According to Apple, this eclipses shipments by all other UNIX/UNIX-like system vendors.
So what? Linux distros are compatible enough that only the most old-school care about whether Red Hat or Debian or Novell or whoever have more "market share". The only thing that's interesting is how many people are using Linux vs XYZ platform.
I've not seen any hard statistics on this because there aren't any. You cannot count Linux installations, so it's pointless trying to use statistics to prove a point here. In my world (a university) I'd say I know more people using Linux than using Macs, just, but that's probably because it's a university. If I worked at a design house I'd probably have to say the opposite.
Actually I always loved Encarta. The thing is, that it's got a lot of good articles on a range of very interesting topics and you can spend literally days just flicking through and reading them. People say "but the internet!!" which is fine but on the web you don't get lots of videos, interactive 3D models, animations, an interactive atlas, sounds and so on. Believe it or not, these really do make a difference. Encarta was always one of my favourite products from Microsoft.
And, sooner or later, Adobe will port PS to Linux, and GIMP will die out, except for the extreamists. PS is a fine tool.
That's clearly not true, as Photoshop already works well enough on Linux via Codeweavers Crossover for DreamWorks to use professionally and is the GIMP dead? No, it's stronger than ever, and had several hundred submissions for its splash screen contest.
Re:But does it work on Windows?
on
RAD with Ruby
·
· Score: 1
That's not quite what he meant. KDE is not ported to Windows and Qt has a hefty cost especially if you're just doing a one-off custom app thing. Python + Glade is quite a powerful combination because it includes the GUI environment as well, not just the runtime.
Well I'd hope they did reuse Anaconda as much as possible, there's a *huge* amount of effort gone into that program and it'd be a crying shame to reinvent the wheel.
Still, I fully expect them to do exactly that because it's always cooler to write your own than reuse somebody elses, and especially because Red Hat wrote it and it's associated with RPM.
It wasn't so much Xing leaking a key that undid CSS, it was the fact that the keys were mathematically related and so could be brute forced in about a week with a fast computer.
If the key generation algorithm had been done correctly, DeCSS may not have been possible.
Oh come on. Rose tinted glasses anybody? Delphi was great, and I used it for years, but let's not be blind to its faults. In many areas the GTK+/Glade/[Python|GtkMM] combination beats the snot out of it:
Proper containment based layout. Later in its life Delphi got "anchors" but these never worked well for me, and were hard to get right.
Great stock artwork: sorry, the Delphi stock icons and artwork was incredibly limited, and what was there sucked balls. You can spot a Delphi app miles off by its Windows 3.1 era graphics.
Event handling model. Delphi events were simply fancy function pointers, but you could only connect one method/function to an event. In GTK+ (and Qt) you have more flexible signals, which allow multiple connections.
Language limitations. Yes, Object Pascal had a lot of upsides, for instance its properties system, but it lacked support for things like templates and AFAIK they were never added. Delphi required you to use Pascal. GTK+ supports nearly every language going, even obscure ones like Haskell or Lisp. About the only one it doesn't support AFAIK is Visual Basic.
IDE limitations. Delphi worked great as long as you only used Delphi. Forget about trying to use emacs to write Object Pascal code, it just wasn't happening. The IDE did a lot of magic behind the scenes to avoid ugly generated code and it worked well, but tied you to the Delphi editing component which was, and still is, a crude childs toy compared to emacs. About the only feature it has which emacs lacks is great symbolic completion, and they're even working on adding that now.
COM integration was ropey. I dunno about you but I often wasted a great deal of time trying to figure out WTF was going on underneath the hood. Delphis abstractions and wizards were well designed but if they messed up you were SOL.
Online help. This worked well in earlier versions and progressively got worse with later versions in Delphi. GTK+ has rather poor online help as well, but it's all available on the web and is indexed by Google, and DevHelp can be used if you want an interactive indexer/help viewer.
GTK+ has a ton of annoying quirks but if I really try and remember Delphi well as opposed to simply recalling the best bits, I remember stupid stuff about the VCL too. GTK+ is also more portable, Kylix/Qt never worked well.
I was under the impression metacity was written because sawfish did not meet the projects usability goals, and also because Sawfish was partly written in the authors own specific dialect of Lisp. A language per author was deemed a bit much.
The problem with migrating to Linux isn't people not knowing how to use a command line. The latest revisions of the desktop Linux distros hardly require that at all anymore.
The problems are things like hardware and software compatibility. Sorry, but Apple gets us nowhere with this: their hardware AND software is closed. Their APIs are every bit as proprietary as Windows' are.
Migrating to MacOS would simply lock you into Apple instead of Microsoft, it wouldn't actually accomplish anything. It certainly would not make a migration to Linux any easier, as at least there is an implementation of the Windows APIs (Wine) which works, whereas there is no complete or useful implementation of the Mac APIs. Oh sure there's GNUstep which nobody uses, isn't binary compatible, and is massively behind Apple, but there's no implementation of Carbon which is what many of the interesting apps use.
So no, I think that'd be a pointless thing to do, and I as a Linux user will certainly not be recommending to anybody that they should buy Macs.
Tough. You're wrong. Go search bugzilla, there is a bug there over 2 years old that basically says "Gecko does not release memory when you close windows or tabs". At all. That's one hell of a leak and it's never even been looked at.
Seriously, trust me on this one, I know the internals of Windows pretty well. Until you actually run IE there are no significant parts of it loaded into memory. No MSHTML, no BROWSEUI, no SHDOCVW etc. You can't see it because it's not there, not because Microsoft cunningly hid it.
Mozilla has always been a memory pig, and the developers apparently are no longer interested in that sort of profiling.
MSFT have been trying to kill Win32 for years. That's what.NET is all about, and there were plenty of initiatives before that. They failed every single time: there's just too much damn software already using it out there.
The integration work for GNOME and KDE has been largely funded by Novell and Red Hat. There are apparently no such companies funding OS X integration, probably because Apple values its relationship with Microsoft too much.
Yes, it's by design. It sounds like you're trying to simulate a tiled MDI layout in the absence of MDI (which is also by design). A better solution may be a large window with splitters and/or dock widgets rather than attempting to fight the WM, which is resisting your abuse of it.
I posted that example because in the past people have said that their users could not install or run custom software as their home directory was mounted noexec. Obviously if there are no restrictions like that in place you can just run it, point being that even if they are in place you can still run it.
The grand-parent was suggesting that allowing people to install any software they liked into their home dir and run it was a bad idea. But it cannot be prevented, so it doesn't matter if it's a bad idea or not. That was my point.
Actually you can do this, but it's appallingly documented (like a lot of Linux binary compatibility stuff). Put this on your link line:
-Wl,--rpath,'${ORIGIN}'
and now the linker will search the directory the binary is in for libraries first. You can use any path relative to ${ORIGIN} eg '${ORIGIN}/../lib' is appended automatically by apbuild, the autopackage binary portability tool.
Hope that helps. Feel free to ask the autopackage developers any questions you have about binary distribution on Linux.
It's not the stability of the code, and it's not that packaging Wine is hard (it's not). It's that the packagers are either clueless or deliberately ignore upstream policy in favour of distro policy. This can and will lead to obscure breakage (ie the software doesn't run correctly).
I have witnessed this sort of thing many times, with many different programs. I'm afraid it's a systematic problem.
You don't realise they're broken because you're not a developer on those projects. I deal with broken Wine packages every week. I'm sure most of the people apt-getting and emerging away believe they are getting a quality tested package, but they're actually downloading crap-in-a-tarball.
To be fair, recently both Gentoo and Debian got new Wine maintainers after a long period of neglect. Maybe things will get better now. Let's hope so. Unfortunately it doesn't solve the fundamental weaknesses of the system.
Apparently you're not reading what I wrote. There is no sane security system that can prevent the user running arbitrary code. I won't go into all the ways you can do it here, suffice it to say there are a lot.
Investigate what I said to investigate. grsecurity does not stop users from running code, it's not a FS permissions issue. It's a design-of-the-system issue, and basically impossible to work around.
App developers can use binreloc to make their apps binary relocatable (ie can install to any location). This is mandatory for autopackages and is easy to add in. binreloc is designed for minimal source disruption.
Slashdot Mirror
Yes this is a known bug with GTK+ integration. I agree it's pretty embarassing for them - Firefox and Mozilla had always been stable as a rock and then at the last minute, a multiple-crashes-a-day bug was introduced. I'd have expected an extremely fast bugfix from them but they're apparently too busy printing adverts.
As already pointed out, this seems to be a US/non-US thing. MacOS has little to no serious desktop penetration in Europe. iPod is a different story.
Final thing - I did not say "Macs are only used in design", please don't misquote me again in such a blatant fashion. I said that if I was in a design environment instead of a CS degree at university, I would probably see more Macs. Where I live this is true.
So what? Linux distros are compatible enough that only the most old-school care about whether Red Hat or Debian or Novell or whoever have more "market share". The only thing that's interesting is how many people are using Linux vs XYZ platform.
I've not seen any hard statistics on this because there aren't any. You cannot count Linux installations, so it's pointless trying to use statistics to prove a point here. In my world (a university) I'd say I know more people using Linux than using Macs, just, but that's probably because it's a university. If I worked at a design house I'd probably have to say the opposite.
Actually I always loved Encarta. The thing is, that it's got a lot of good articles on a range of very interesting topics and you can spend literally days just flicking through and reading them. People say "but the internet!!" which is fine but on the web you don't get lots of videos, interactive 3D models, animations, an interactive atlas, sounds and so on. Believe it or not, these really do make a difference. Encarta was always one of my favourite products from Microsoft.
That's clearly not true, as Photoshop already works well enough on Linux via Codeweavers Crossover for DreamWorks to use professionally and is the GIMP dead? No, it's stronger than ever, and had several hundred submissions for its splash screen contest.
That's not quite what he meant. KDE is not ported to Windows and Qt has a hefty cost especially if you're just doing a one-off custom app thing. Python + Glade is quite a powerful combination because it includes the GUI environment as well, not just the runtime.
Still, I fully expect them to do exactly that because it's always cooler to write your own than reuse somebody elses, and especially because Red Hat wrote it and it's associated with RPM.
If the key generation algorithm had been done correctly, DeCSS may not have been possible.
GTK+ has a ton of annoying quirks but if I really try and remember Delphi well as opposed to simply recalling the best bits, I remember stupid stuff about the VCL too. GTK+ is also more portable, Kylix/Qt never worked well.
I was under the impression metacity was written because sawfish did not meet the projects usability goals, and also because Sawfish was partly written in the authors own specific dialect of Lisp. A language per author was deemed a bit much.
Yikes! You think bizarre hacks around InstallShield are a better solution than something like Red Carpet server or the cheaper alternative:
for m in $( cat machines.list ); do ssh root@$m 'rpm -Uvh http://deployment/foo.rpm' done
Just because a tool exists for Windows doesn't mean the same thing is needed on Linux.
The Active Directory point I will give you, though some people are working on this.
The problem with migrating to Linux isn't people not knowing how to use a command line. The latest revisions of the desktop Linux distros hardly require that at all anymore.
The problems are things like hardware and software compatibility. Sorry, but Apple gets us nowhere with this: their hardware AND software is closed. Their APIs are every bit as proprietary as Windows' are.
Migrating to MacOS would simply lock you into Apple instead of Microsoft, it wouldn't actually accomplish anything. It certainly would not make a migration to Linux any easier, as at least there is an implementation of the Windows APIs (Wine) which works, whereas there is no complete or useful implementation of the Mac APIs. Oh sure there's GNUstep which nobody uses, isn't binary compatible, and is massively behind Apple, but there's no implementation of Carbon which is what many of the interesting apps use.
So no, I think that'd be a pointless thing to do, and I as a Linux user will certainly not be recommending to anybody that they should buy Macs.
Tough. You're wrong. Go search bugzilla, there is a bug there over 2 years old that basically says "Gecko does not release memory when you close windows or tabs". At all. That's one hell of a leak and it's never even been looked at.
Seriously, trust me on this one, I know the internals of Windows pretty well. Until you actually run IE there are no significant parts of it loaded into memory. No MSHTML, no BROWSEUI, no SHDOCVW etc. You can't see it because it's not there, not because Microsoft cunningly hid it.
Mozilla has always been a memory pig, and the developers apparently are no longer interested in that sort of profiling.
Open source isn't a panacea.
MSFT have been trying to kill Win32 for years. That's what .NET is all about, and there were plenty of initiatives before that. They failed every single time: there's just too much damn software already using it out there.
The integration work for GNOME and KDE has been largely funded by Novell and Red Hat. There are apparently no such companies funding OS X integration, probably because Apple values its relationship with Microsoft too much.
Yes, it's by design. It sounds like you're trying to simulate a tiled MDI layout in the absence of MDI (which is also by design). A better solution may be a large window with splitters and/or dock widgets rather than attempting to fight the WM, which is resisting your abuse of it.
That's called "lock in", don'tcha know.
I posted that example because in the past people have said that their users could not install or run custom software as their home directory was mounted noexec. Obviously if there are no restrictions like that in place you can just run it, point being that even if they are in place you can still run it.
The grand-parent was suggesting that allowing people to install any software they liked into their home dir and run it was a bad idea. But it cannot be prevented, so it doesn't matter if it's a bad idea or not. That was my point.
-Wl,--rpath,'${ORIGIN}'
and now the linker will search the directory the binary is in for libraries first. You can use any path relative to ${ORIGIN} eg '${ORIGIN}/../lib' is appended automatically by apbuild, the autopackage binary portability tool.
Hope that helps. Feel free to ask the autopackage developers any questions you have about binary distribution on Linux.
I have witnessed this sort of thing many times, with many different programs. I'm afraid it's a systematic problem.
To be fair, recently both Gentoo and Debian got new Wine maintainers after a long period of neglect. Maybe things will get better now. Let's hope so. Unfortunately it doesn't solve the fundamental weaknesses of the system.
Apparently you're not reading what I wrote. There is no sane security system that can prevent the user running arbitrary code. I won't go into all the ways you can do it here, suffice it to say there are a lot.
Investigate what I said to investigate. grsecurity does not stop users from running code, it's not a FS permissions issue. It's a design-of-the-system issue, and basically impossible to work around.
App developers can use binreloc to make their apps binary relocatable (ie can install to any location). This is mandatory for autopackages and is easy to add in. binreloc is designed for minimal source disruption.