> There are different ways to deal with it. Having written a 150 page thesis recently, I can assure you that I had it in a subversion repository and had several checkouts in different places.
I remember the good old ext2 days, where my linux box could just die when the power was cut during disk activity, while my freebsd one was rebooting without issues. Of course, the bsd box was journalled, but linux guys were basically saying "if you care about your data, use an UPS" (that was waaay before ext3).
So, my point is that not everyone will use subversion and backup to different devices just to guard against a low occurrence event (power failure) that is generally harmless. But ext4 just changed from "generally harmless" to "potentially lethal" by growing the window of data loss from 5 seconds to 120 seconds (of course, Theodore T'so is already working on a patch, so I guess it means that, yes, it is a bug...)
> In addition, most text editors rename the old file and write to a new one
Renaming the old file to write a new one kills hard links, so I know at least some editors that used to avoid that...
> At most the last edit is lost, and that is acceptable in my book.
Unless you save twice in 120 seconds, of course...
> I think you did not understand my posting at all.
I think I did, but we can agree to disagree. The point is that your "startup-critical files" are actually not critical at all, compared to the 50 pages thesis that the user is working on. So, if you want to argue that fsync() should be used for gnome/kde configuration files, then it should also be used for cp...
Yep. The overall attitude in that article have a tremendously high suckage ratio.
10 years ago, slashdot was full of discussion about using virus to remove malware, and is now filled with wannabe lawyers whining as soon as someone does anything remotely constructive, with sentences like:
"How is changing the desktop not modification of computer material?"
"I am surprised, [...] that the BBC is not already feeling the ramifications of its actions"
"[...] this [...] actually risk becoming an international incident?"
"This is both highly illegal and unethical"
"Regardless of intent it is illegal"
bla-bla. All modded +5. Depressing.
And of course, journalists always had to do borderline legal (or plain illegal) stuff to do their job. So, of course, their lawyers said it was fine. Because it is.
> I agree. This is not way to treat startup-critical configuration files.
This is bull. Most files are critical to someone. This would means that most processes that write data must use fsync.
Are you arguing that cp should use fsync for every file it copies ? In that case, you'd better tell the maintainers of coreutils-7.1, because copy_internal (used by cp.c) does not. (And you'll be laughted at)
So, right, now, on ext4, the sequence:
> cp/disk1/file1.data/disk2/file1.data
wait a few seconds
> rm/disk1/file1.data
crash
will probably cause the file to be lost. That you choose to blame it on cp is funny, but most of the rest of the world will blame it on ext4.
> [Icelanders] is a whole country of smart, literate people
I was under the impression that Iceland was an island populated by morons that let their own government sell their DNA data to the world (http://www.actionbioscience.org/genomic/hlodan.html) Those very same people, also started to live on credit, without any concern about the realities of the economy of their island.
> Fucking over Icelanders is not a generally good strategy.
My general feeling is that fucking over Icelanders by politician and business should be seen as an advanced warning about what is going to happen about everywhere else in the world. I should add that, so far, it seems to be a pretty good strategy for the people doing the fucking.
> I switched from Parallels Desktop to VirtualBox and it has one feature which I really like; the ability to run for over a week without causing a kernel panic.
Same here.
> Since I don't like paying for bug fixes, I never found out if the new version actually did fix the problem, but there's no way I'd give the company any money after that.
Exactly. I bought Parallels v2 at the very beginning, and even updated to v3 before I really used v2. Then I had two troubles with them:
1) The product is buggy, and I don't like paying for bug fixes (automatic upgrades didn't really improved the situation). I am ok to pay for a new version of a product I already have if I am happy with it, which is not the case.
2) They never sent me the wine source code I asked for.
In general their support team waves hands, and pretend everything is ok in their side and that the issue is on your side (you apps or even your ISP (!))
> I'm sure it doesn't help things that Stroussoup made this explicit [att.com] in C++. So if your view is that C is a subset of C++, you'll get these trivia wrong. Unfortunately, C and C++ will penalize you for getting trivia questions wrong with great zeal.
You are wrong on two counts: first, his name is Stroustrup. Second, in C++, like in C, the literal 0 in a pointer context will be turned into the NULL pointer by the compiler.
void *p = (void*)0;
p will be a NULL pointer
int a = 0; void *p = (void *)a;
p may not be a NULL pointer, as the 0 was not a literal in a pointer context.
Hence,
if (p)
means:
if (p==0)
so, if p is a pointer type, it means:
if (p==(void*)0)
which means
if (p is the NULL pointer)
The NULL pointer can be represented internally by whatever bitpattern or trick that the compiler writer wants.
(And of course, it was a stupid position to begin with, but clever people are pretty good at rationalizing anything. I guess that in 2000 he would have rejected git on the ground that it makes too easy to add stuff to the kernel)
> I just installed Fedora Core 10 on a spare box with a bc4306 wireless card, and > after screwing around for a few hours hooked up an RJ45 to the back because it > was just too much a pain in the ass to get wireless to work on it. I would have > tried Ubuntu but its support for SATA RAID sucks. I work with Linux and Unix for > my job, but don't have to worry about user interface because it is mostly > command line there. But at home I want a GUI thank you very much. I also want it > to just work. I work on computer systems during the day... when I want to use my > machine at night I don't want to 'work' on it. I wish it were the case that this > will drive people to Linux but it won't until it just works, including every day > things that are done very easily on Windows like setting up SATA RAID and > wireless. No blaming vendors or users... it just needs to work. If anything, it > will drive people to Apple who in my opinion aren't any better than Microsoft > when it comes to 'lock in' and DRM shite. Fist the things in Linux that should > be easy and people will use it. Otherwise they won't. No, I don't want to fix > it. I don't need to fix it. I am willing to pay for MS Windows if I need > something to work. I don't rip CDs or other crap so this won't bother me. I'm > just tired of hearing how MS does something else that is bad and of course this > will drive people to Linux. No it won't. It will never drive people to Linux > until Linux works as easily as Windows does. And don't get in a huff over this, > the proof is in the pudding. People ignore programs that are hard to use and use > programs that are easy or easier to use. [...]
And people ignore posts that are hard to read, and read posts that are easy or easier to use.
So, what about learning about that big key labelled "return" that is located on the center right of your keyboard ?
You got that wrong. Microsoft was the new Apple. And Apple is now getting back to the old Apple.
I remember the time when the gcc license specifically prohibited you to run it on Apple hardware (that was at A/UX time), because "apple limited your freedom". Apple was the most greedy, most control freak and most bullying company at that time.
(just my opinion here. if you were a BeOS fan, skip it)
I was a NeXTstep developer at that time, an I did quite a lot of BeOS developement too.
Let me give it to you straight:
BeOS sucked balls. The APIs were horrendous C++ kludges. For a design that was done 8 years later than NeXT, it didn't make sense. The UI was ugly (for instance, windows minimisation left the small title bar in the middle of the screen).
At the end, it really looked like a bastard C++ clone of MacOS to me (which was already doomed at that time), with a multitasking OS. You see, a bit like if a group of Mac developers wanted to rebuild Mac OS "right", without seeing that the world had moved since...
> Microsoft can't even sneeze without getting in trouble with the courts. What makes you think they could pull something like this off? It has "antitrust" plastered all over it.
No it doesn't. They should just add a framework for registering versions, urls to find new versions and upgrade procedures, so they could integrate ALL the updates in a single service and a single UI. With, of course, some sort of crypto infrastructure (saying that, for instance, an app that register for updates should provide a public key, and all upgrade must be signed with that key).
The protocols could be totally open (ie: just HTTP).
They would NOT control anything, just gives a shared set of tools to perform upgrades. No antitrust there. Just user friendliness...
First, that is not the issue. The issue is that they ADDED that requirement without warning. If you don't accept that new requirement, you cannot run the new software.
Second, I don't want every fucking software company from the earth to add services for various kind of updaters. If they want to check for updates, they can check at app startup. Better yet, as I actually "grasp how network apps like google earth work", there should be version numbers in the data protocol.
We all know how this will end: you'll end up having hundred of processes calling home to search for update for software you've uninstalled ages ago. Each software company will jump on the idea of having their own updater sucking cycles on you machine, because everyone think that their software is the most important thing in the world...
I read that Google Earth 5.0 had some new features. As I did not update Google Earth in a long time, I decided to give it a go.
I downloaded Google Earth 5.0, opened up the dmg. I then proceeded to a normal Mac upgrade: I COPIED THE NEW VERSION ON TOP OF THE OTHER.
I then launched the new software. I was greeted with two options:
1) Accept that the Google Updater gets installed. 2) Quit
There was no warning of any kind when I downloaded the software that it will remove functionality unless I accept some sort of spyware/nagware/bloatware on my machine.
So, I have a few hundreds of "bookmarks" in my Google Earth application. What am I supposed to do ?
For the record, I accepted the install, because I thought that the update app would be easy to disable. Wrong. There is no UI for that app. I had to go the hard way to disable it. I will now go back to Google Earth 4.3.
Anyway, Google finally crossed the line for me. "Do no evil" ? My ass...
> They couldn't because a decision was made to make KDE 4 more compatible with the future is to redesign it now and go from there.
Been there, done that. "We break things now so we don't break them later. This thing is now future-proof". Same excuse that was made for KDE2=>KDE3 transition. Where is the "futureproofness" of KDE3, now ? "Oh, there was none, sorry. But THIS time, it IS futureproof, and the foundation are sane, and bla, bla, bla". Until next time.
All that is pure rationalization. The truth is that they broke things and alienated users. It is not due to any technical reason, only due to the way they manage the project. Same people. Same process. Same mistake.
"And they probably redesigned the whole sickbay too! I know engineers. They love to change things."
While your answer is indeed somewhat funny, I don't exactly understand what is your problem with my post. I just reported a bad experience that was due to a badly designed USB hub from Belkin. It may be uninteresting, but I fail to see how it can be irrelevant to the discussion at hand. Care to enlighten me ?
> So you are supposed to say "a USB port" and "an AMD motherboard".
Good. So, I should that you are *an* ignorant and pompous a-hole that is correcting the spelling of someone that isn't even a native English speaker ?
> I am so sick of hearing this bloody retarded English being peddled by so-called English speakers.
Ceci dit, on peux comparer mon Anglais avec ton Français. J'ai comme à l'idée que tu ne vas pas pouvoir aligner plus de deux mots sans te ridiculiser. Mais bon, ca n'a pas l'air de te faire peur, alors allons-y...
Slashdot Mirror
> There are different ways to deal with it. Having written a 150 page thesis recently, I can assure you that I had it in a subversion repository and had several checkouts in different places.
I remember the good old ext2 days, where my linux box could just die when the power was cut during disk activity, while my freebsd one was rebooting without issues. Of course, the bsd box was journalled, but linux guys were basically saying "if you care about your data, use an UPS" (that was waaay before ext3).
So, my point is that not everyone will use subversion and backup to different devices just to guard against a low occurrence event (power failure) that is generally harmless. But ext4 just changed from "generally harmless" to "potentially lethal" by growing the window of data loss from 5 seconds to 120 seconds (of course, Theodore T'so is already working on a patch, so I guess it means that, yes, it is a bug...)
> In addition, most text editors rename the old file and write to a new one
Renaming the old file to write a new one kills hard links, so I know at least some editors that used to avoid that...
> At most the last edit is lost, and that is acceptable in my book.
Unless you save twice in 120 seconds, of course...
Have a nice day.
> I think you did not understand my posting at all.
I think I did, but we can agree to disagree. The point is that your "startup-critical files" are actually not critical at all, compared to the 50 pages thesis that the user is working on. So, if you want to argue that fsync() should be used for gnome/kde configuration files, then it should also be used for cp...
Yep. The overall attitude in that article have a tremendously high suckage ratio.
10 years ago, slashdot was full of discussion about using virus to remove malware, and is now filled with wannabe lawyers whining as soon as someone does anything remotely constructive, with sentences like:
"How is changing the desktop not modification of computer material?"
"I am surprised, [...] that the BBC is not already feeling the ramifications of its actions"
"[...] this [...] actually risk becoming an international incident?"
"This is both highly illegal and unethical"
"Regardless of intent it is illegal"
bla-bla. All modded +5. Depressing.
And of course, journalists always had to do borderline legal (or plain illegal) stuff to do their job. So, of course, their lawyers said it was fine. Because it is.
> I agree. This is not way to treat startup-critical configuration files.
This is bull. Most files are critical to someone. This would means that most processes that write data must use fsync.
Are you arguing that cp should use fsync for every file it copies ? In that case, you'd better tell the maintainers of coreutils-7.1, because copy_internal (used by cp.c) does not. (And you'll be laughted at)
So, right, now, on ext4, the sequence:
> cp /disk1/file1.data /disk2/file1.data
wait a few seconds
> rm /disk1/file1.data
crash
will probably cause the file to be lost. That you choose to blame it on cp is funny, but most of the rest of the world will blame it on ext4.
Yes, yes, yes, yes, and yes.
Wish I had mod points today...
> [Icelanders] is a whole country of smart, literate people
I was under the impression that Iceland was an island populated by morons that let their own government sell their DNA data to the world (http://www.actionbioscience.org/genomic/hlodan.html)
Those very same people, also started to live on credit, without any concern about the realities of the economy of their island.
> Fucking over Icelanders is not a generally good strategy.
My general feeling is that fucking over Icelanders by politician and business should be seen as an advanced warning about what is going to happen about everywhere else in the world. I should add that, so far, it seems to be a pretty good strategy for the people doing the fucking.
Of course, it is firebug. I have to restart firefox every 30 minutes due to it, but I can't complain: it is such an invaluable tool.
> I switched from Parallels Desktop to VirtualBox and it has one feature which I really like; the ability to run for over a week without causing a kernel panic.
Same here.
> Since I don't like paying for bug fixes, I never found out if the new version actually did fix the problem, but there's no way I'd give the company any money after that.
Exactly. I bought Parallels v2 at the very beginning, and even updated to v3 before I really used v2. Then I had two troubles with them:
1) The product is buggy, and I don't like paying for bug fixes (automatic upgrades didn't really improved the situation). I am ok to pay for a new version of a product I already have if I am happy with it, which is not the case.
2) They never sent me the wine source code I asked for.
In general their support team waves hands, and pretend everything is ok in their side and that the issue is on your side (you apps or even your ISP (!))
They are not going to get any more money from me.
> I'm sure it doesn't help things that Stroussoup made this explicit [att.com] in C++. So if your view is that C is a subset of C++, you'll get these trivia wrong. Unfortunately, C and C++ will penalize you for getting trivia questions wrong with great zeal.
You are wrong on two counts: first, his name is Stroustrup. Second, in C++, like in C, the literal 0 in a pointer context will be turned into the NULL pointer by the compiler.
void *p = (void*)0;
p will be a NULL pointer
int a = 0;
void *p = (void *)a;
p may not be a NULL pointer, as the 0 was not a literal in a pointer context.
Hence,
if (p)
means:
if (p==0)
so, if p is a pointer type, it means:
if (p==(void*)0)
which means
if (p is the NULL pointer)
The NULL pointer can be represented internally by whatever bitpattern or trick that the compiler writer wants.
> > should they focus more on usability?
> Errr... yes?
> How can you possibly answer "no" to that question?
Well, it is open source we are talking about.
Here is Linus Torvald back in 2000 explaining that adding features that make development easier in the kernel is not a good idea:
http://lkml.indiana.edu/hypermail/linux/kernel/0009.0/1148.html
(And of course, it was a stupid position to begin with, but clever people are pretty good at rationalizing anything. I guess that in 2000 he would have rejected git on the ground that it makes too easy to add stuff to the kernel)
> I just installed Fedora Core 10 on a spare box with a bc4306 wireless card, and
> after screwing around for a few hours hooked up an RJ45 to the back because it
> was just too much a pain in the ass to get wireless to work on it. I would have
> tried Ubuntu but its support for SATA RAID sucks. I work with Linux and Unix for
> my job, but don't have to worry about user interface because it is mostly
> command line there. But at home I want a GUI thank you very much. I also want it
> to just work. I work on computer systems during the day... when I want to use my
> machine at night I don't want to 'work' on it. I wish it were the case that this
> will drive people to Linux but it won't until it just works, including every day
> things that are done very easily on Windows like setting up SATA RAID and
> wireless. No blaming vendors or users... it just needs to work. If anything, it
> will drive people to Apple who in my opinion aren't any better than Microsoft
> when it comes to 'lock in' and DRM shite. Fist the things in Linux that should
> be easy and people will use it. Otherwise they won't. No, I don't want to fix
> it. I don't need to fix it. I am willing to pay for MS Windows if I need
> something to work. I don't rip CDs or other crap so this won't bother me. I'm
> just tired of hearing how MS does something else that is bad and of course this
> will drive people to Linux. No it won't. It will never drive people to Linux
> until Linux works as easily as Windows does. And don't get in a huff over this,
> the proof is in the pudding. People ignore programs that are hard to use and use
> programs that are easy or easier to use. [...]
And people ignore posts that are hard to read, and read posts that are easy or easier to use.
So, what about learning about that big key labelled "return" that is located on the center right of your keyboard ?
> echo -e 'global _start \n _start: \n mov eax, 2 \n int 80h \n jmp _start' > a.asm; nasm a.asm -f elf; ld a.o -o a;
An assembly fork bomb. You just forgot to execute it, but that's cute anyway...
> Apple is the new Microsoft.
You got that wrong. Microsoft was the new Apple. And Apple is now getting back to the old Apple.
I remember the time when the gcc license specifically prohibited you to run it on Apple hardware (that was at A/UX time), because "apple limited your freedom". Apple was the most greedy, most control freak and most bullying company at that time.
What about reading the blurb before posting ?
"This would only go into affect after enough states totaling 270 electoral votes (enough to elect a president) adopted similar resolutions."
That means the vote will count, until the US president is elected on popular vote.
(just my opinion here. if you were a BeOS fan, skip it)
I was a NeXTstep developer at that time, an I did quite a lot of BeOS developement too.
Let me give it to you straight:
BeOS sucked balls. The APIs were horrendous C++ kludges. For a design that was done 8 years later than NeXT, it didn't make sense. The UI was ugly (for instance, windows minimisation left the small title bar in the middle of the screen).
At the end, it really looked like a bastard C++ clone of MacOS to me (which was already doomed at that time), with a multitasking OS. You see, a bit like if a group of Mac developers wanted to rebuild Mac OS "right", without seeing that the world had moved since...
I was so waiting for someone to post that...
For the clueless mods that marked you as Troll: http://www.youtube.com/watch?v=UDPqB9i1ScY
That "cyborg" post is a great post. I'll keep that handy: I just started to need to remove my glasses to read...
And, kuro5hin still exist ? Amazing...
> Microsoft can't even sneeze without getting in trouble with the courts. What makes you think they could pull something like this off? It has "antitrust" plastered all over it.
No it doesn't. They should just add a framework for registering versions, urls to find new versions and upgrade procedures, so they could integrate ALL the updates in a single service and a single UI. With, of course, some sort of crypto infrastructure (saying that, for instance, an app that register for updates should provide a public key, and all upgrade must be signed with that key).
The protocols could be totally open (ie: just HTTP).
They would NOT control anything, just gives a shared set of tools to perform upgrades. No antitrust there. Just user friendliness...
First, that is not the issue. The issue is that they ADDED that requirement without warning. If you don't accept that new requirement, you cannot run the new software.
Second, I don't want every fucking software company from the earth to add services for various kind of updaters. If they want to check for updates, they can check at app startup. Better yet, as I actually "grasp how network apps like google earth work", there should be version numbers in the data protocol.
We all know how this will end: you'll end up having hundred of processes calling home to search for update for software you've uninstalled ages ago. Each software company will jump on the idea of having their own updater sucking cycles on you machine, because everyone think that their software is the most important thing in the world...
It is worse than that.
I read that Google Earth 5.0 had some new features. As I did not update Google Earth in a long time, I decided to give it a go.
I downloaded Google Earth 5.0, opened up the dmg. I then proceeded to a normal Mac upgrade: I COPIED THE NEW VERSION ON TOP OF THE OTHER.
I then launched the new software. I was greeted with two options:
1) Accept that the Google Updater gets installed.
2) Quit
There was no warning of any kind when I downloaded the software that it will remove functionality unless I accept some sort of spyware/nagware/bloatware on my machine.
So, I have a few hundreds of "bookmarks" in my Google Earth application. What am I supposed to do ?
For the record, I accepted the install, because I thought that the update app would be easy to disable. Wrong. There is no UI for that app. I had to go the hard way to disable it. I will now go back to Google Earth 4.3.
Anyway, Google finally crossed the line for me. "Do no evil" ? My ass...
In Soviet France also... (link in French, about smart ads in subways...)
http://www.lemonde.fr/societe/article/2009/01/28/des-elus-reclament-le-retrait-des-ecrans-de-pub-animes-dans-le-metro_1147803_3224.html#ens_id=1147813
> They couldn't because a decision was made to make KDE 4 more compatible with the future is to redesign it now and go from there.
Been there, done that. "We break things now so we don't break them later. This thing is now future-proof". Same excuse that was made for KDE2=>KDE3 transition. Where is the "futureproofness" of KDE3, now ? "Oh, there was none, sorry. But THIS time, it IS futureproof, and the foundation are sane, and bla, bla, bla". Until next time.
All that is pure rationalization. The truth is that they broke things and alienated users. It is not due to any technical reason, only due to the way they manage the project. Same people. Same process. Same mistake.
"And they probably redesigned the whole sickbay too! I know engineers. They love to change things."
Friendly advice ?
> "I am so sick of hearing this bloody retarded English being peddled by so-called English speakers."
If that is what you call "friendly", I'm glad not to be one of your friends.
> "At least I know how definite articles work in French, which is more than I can say for you and indefinite articles in English"
At least, I know how to communicate in English, which seems to be more than I can say for you in French.
Cuts both ways, you see...
While your answer is indeed somewhat funny, I don't exactly understand what is your problem with my post. I just reported a bad experience that was due to a badly designed USB hub from Belkin. It may be uninteresting, but I fail to see how it can be irrelevant to the discussion at hand. Care to enlighten me ?
> So you are supposed to say "a USB port" and "an AMD motherboard".
Good. So, I should that you are *an* ignorant and pompous a-hole that is correcting the spelling of someone that isn't even a native English speaker ?
> I am so sick of hearing this bloody retarded English being peddled by so-called English speakers.
Ceci dit, on peux comparer mon Anglais avec ton Français. J'ai comme à l'idée que tu ne vas pas pouvoir aligner plus de deux mots sans te ridiculiser. Mais bon, ca n'a pas l'air de te faire peur, alors allons-y...