Right. But as a root process it can scribble over any memory it wants. That's the problem. Linux is not a good platform for games because everybody has to implement graphics in userspace.
If unreal server crashes your linux box... that's the fault of linux. I don't think it needs to run as root. As for unreal client, that's another story. That's the combined bugs of your graphics accelerator, their closed-source libraries (in most cases), MESA and X working against you. So don't be too quick to blame them... instead blame the kernel that dosn't talk to the video card, but instead hands full control to userspace programs... and forces each one to implement their own hardware level drivers. (in userspace, no less!)
Is there anything similar to DBI for C? Obviously you can't do it exactly the way DBI does, but you can present a uniform interface to all databases.
With that in mind, all the C database code I write is through wrappers designed to work the same way as DBI, but I only write it for the databases I actually work with. (Postgresql and Oracle)
Am I re-inventing the wheel or is this something that's only been done for perl?
The make test failed loading a shared library, so I wrote a simple c prog just to load all the.so libs in a dir and it would load every.so file in the disk except any one in the oracle tree (Forget the error now, sorry). The kluging that followed was horrific.
PLEASE tell me you added the oracle library dir to/etc/ld.so.conf?? That's generally the problem. (Oracle installer dosn't do that for you)
I'd far rather have someone who says what needs to be said and steps on toes in the process, than someone who says whatever his 'people' want to hear. eg: Bill Clinton, Perens, etc.
Are you really serious? ESR? Saying what needs to be said as opposed to what people want to hear?
The man spews out EXACTLY what you want to hear. Microsoft is evil. Open source is the only way. All bow down and worship the code...
It's all about the coverage, the notice... the ego.
The problem I see with it is that you have to trust the other party... which means they have to trust you. There's a lot of quite probably flawed trust going on.
You trust them to not take a screenshot or other capture of the "unsaveable" message.
They allow the message to be deleted because they trust your judgement that it's not a safe thing to keep around.
... by extension, that means they trust YOU to have not kept a copy around, as a possible "get out of jail free" card.
... wich they could have used themselves. "Look! I was acting under orders!"
Other problems include actually deleting the message: impossible. Deleting the key: Impossible. In fact, I can open the message multiple times without even being a "hacker". Don't install their "plugin" on your machine. Send a copy to yourself at your home email address, then read the copy with their enabled-client. When you get home, do the same. If they use IP connectivity to log to a central server, they fail on two counts. A) their product is worthless in any kind of secure environment (firewalls) and B) I can just save the central server's response and patch that into the message.
It's a gimmik (as has been pointed out before). Without control over the recipiant's recieving hardware, it's impossible to ensure that it actually happens.
There's ways of doing it, of course. Some even require some decent skills to defeat... or a digital camera. Even Jim could have used that.
Takes just as long to truncate a file as it does to delete it, since the same operations remain in effect (moving all the blocks to the freelist). All that happens is your > filename takes a long time and rm is instant.
As someone who often works with VERY large files (burning webserver logs to CD and deleteing them once verified) I can state that it takes way too long to delete 500+ meg files.
Linux lacks a commercial quality Journaling File System. This means that in the event of a system failure, such as a power outage, data loss or corruption is possible. In any event, the system must check the integrity of the file system during system restart, a process that will likely consume an extended amount of time, especially on large volumes and may require manual intervention to reconstruct the file system.
8gig is about the performance limit of ext2. After that it starts to take significant time to simply mount the disk read/write. (Lots of structure to copy into memory). 25+gig arrays take an indecent amount of time. And don't even talk about e2fsck.
When linux crashes (not if, this is a high load box) recovery time is astronomical.
They missed one point, though. No good way to backup large volumes. Dumpe2fs craps at about 8gig... lots of 64bit problems. Tarballing a 30 gig array as a "backup" solution just dosn't cut it... no good way to do incrementals.
Hopefully we'll see some decent large-filesystem support by 3.0, since theres a few different projects working on it already.
Actually, the problem is not that the KKK or porn sites have content relating to Amber Bleys or norse mythology. The problem he is talking about is that porn sites (and other morally questionable sites) use keywords completely unrelated to the site in question in order to confuse search engines. This is an approximate equivilant to turning on the Disney channel and finding someone is sneaking in hardcore porn.
What is needed is a technical solution to this, and the furor over "protect the children" will die down. If people quit getting porn shoved in their faces, they're happy. They don't care if it's a technical solution or an international treaty enforced by men in black helicopters, they just want it.
... and of course the religeous reich would have a shit-fit about legal "kiddie" porn. Gotta love the clash of science and religon.
This will be suppressed, of course. Just like "miracle" drugs, you'll have to go out of the country for it.
Most likely, it'll happen in some third world country where the people don't matter anyway. You may end up with a different body color, but that dosn't really matter if you're rich. (Ex. Michael Jackson)
If I know how to crack something, and I'm completely sure that nobody else does, would it not be to my advantage to say "Look, I'm using it!" to get others to as well?
... and, of course, you have to PROVE your card got stolen/abused. You can't just buy that brand-new-spiffy-keen cryotech athalon and say "my card was stolen!" and get it for 50 bucks.
and, unlike criminal court, you are guilty until you prove yourself innocent. The $50 limit is for when your wallet is physically stolen and you report it. It does not apply to someone copying your information and abusing it.
* The 2GB of RAM needed to diagonalize the giant matrix just isn't quite as frightening and impressive as it was a couple of years ago...
* This algorithm is massively parallelizable...
In fact, it's not. At all. The sieving is done in parallel, but the actual matrix caclulations require one massive machine to do. Crays are nice because that's what they were designed to do. Simply tossing 2gig of ram into a dec alpha won't help you process the sparse matrix.
it forces you to because coments like yours are completely worthles. We're all aware that there are java ads. Thats why we turn off java/j{,ava}script in our browsers. It's also what ad-filtering proxies are for. (gotta love 'em)
actually, you can do a simple redundant and shared gateway... it's just going to take some work.
implement your standard NAT (including database of active connections) but round-robin your packets between them. Have a seperate connection between NAT servers where they announce their changes to the subnet pool.
That way, the only problem you have is a temporary 50% packetloss until your round-robin system drops the dead router.
Interesting... I had some inital trouble making proxy ARP work properly in linux but after upgrading my net utils to match my kernel it worked properly. I think I can proxy arp for networks as well, but I havn't tried that. If I misunderstood your problem, let me know. but yes, linux does handle ARP properly.
John dosn't even take the time to verify his spell-checker is outputting the right words! Somehow I doubt a call for proper punctuation will get very far. He's nigh-impossible to read because of it, too.
How about your right to make an honest living? Are you willing to give that up? It's not as far-fetched as it seems.
Lets take another "good idea" law, this one from the "war on drugs". This law gave the police the right to sieze any money from people carrying large amounts, by saying "anyone with lots of cash is buying drugs. Period. That's our probable cause."
Sounds reasonable, right? Only drug dealers would carry that kind of cash, of course, so no honest citizens would be affected...
... or so you would think. So, a cattle rancher takes out a large loan at the start of a new season. He converts it to cash, and drives to an auction. Oops, he was late so he sped a little, and got pulled over. Now he goes to jail.
Mind you, it dosn't end there. He quickly proves that he's not a drug dealer, and is released. No problem, right?
Nope. He forfit the entire loan. All of it. Even though he is an honest citizen, his money now belongs to the state. And they do not HAVE to give it back.
Do we really want to wait for a blatent abuse of this new law before we decide that it too is a bad idea?
Remember, the "War on Drugs" was to "save the children" from drug abuse. A noble goal, for sure, but it has failed and has done nothing but to increase the violence. Sometimes good ideas go astray, which is what the founding fathers understood. I wish we had their guidence today.
No? Try to fsck that and have the system back in a reasonable time.
What, 9 isn't enough pain? Try 25. See what happens when that once-in-a-blue-moon(that strangly happens every other week) kernel 'oops' happens and your primary storage array is down while e2fsck is looking at 100% of a 99.9% clean disk.
Is it possible to stop hyping linux for about another 12 months until there's something to hype about? Sorry to dissapoint the kiddiez, but it's NOT ready for serious applications. Personal servers, development workstations (Ha! if you want to produce code incompatable with everything else.)... possibly samba/dhcp on a small office LAN. Scale beyond that and you start failing.
(Mind you, I said something negative about linux, so I must be a troll. Never mind the fact that I run my entire network on it so know firsthand just how "stable" it really is in production.
You are missing the point of the GPLed output. The idea is to prevent someone from taking GCC, patching it with a flag that dumps the entire source of gcc as a tarball to stdout and saying "It's the output of the program! Now it's under any licence I chose!". Hence, compression, encryption or rot-13ing the strings dosn't change anything. --Dan
Re:He means in C, and it actually does work (usual
on
Linux 2.2 DoS Attack
·
· Score: 1
echo 'main() {exit(0);}// useless program' | sed 's#//\(.*\)$#/*\1 */#'
I'll have to agree with you on this one. Yes, those chairs are wonderful. My friend got one as a present (hint hint) and I've had time to check it out. Much better then the $50-$150 office chairs I usually sit in. As soon as I find the spare cash, this is in my future. --Dan
Slashdot Mirror
--Dan
--Dan
With that in mind, all the C database code I write is through wrappers designed to work the same way as DBI, but I only write it for the databases I actually work with. (Postgresql and Oracle)
Am I re-inventing the wheel or is this something that's only been done for perl?
--Dan
PLEASE tell me you added the oracle library dir to /etc/ld.so.conf?? That's generally the problem. (Oracle installer dosn't do that for you)
--Dan
Are you really serious? ESR? Saying what needs to be said as opposed to what people want to hear?
The man spews out EXACTLY what you want to hear. Microsoft is evil. Open source is the only way. All bow down and worship the code...
It's all about the coverage, the notice... the ego.
--Dan
- You trust them to not take a screenshot or other capture of the "unsaveable" message.
- They allow the message to be deleted because they trust your judgement that it's not a safe thing to keep around.
... by extension, that means they trust YOU to have not kept a copy around, as a possible "get out of jail free" card. ... wich they could have used themselves. "Look! I was acting under orders!"
Other problems include actually deleting the message: impossible. Deleting the key: Impossible. In fact, I can open the message multiple times without even being a "hacker". Don't install their "plugin" on your machine. Send a copy to yourself at your home email address, then read the copy with their enabled-client. When you get home, do the same. If they use IP connectivity to log to a central server, they fail on two counts. A) their product is worthless in any kind of secure environment (firewalls) and B) I can just save the central server's response and patch that into the message.It's a gimmik (as has been pointed out before). Without control over the recipiant's recieving hardware, it's impossible to ensure that it actually happens.
There's ways of doing it, of course. Some even require some decent skills to defeat... or a digital camera. Even Jim could have used that.
--Dan
As someone who often works with VERY large files (burning webserver logs to CD and deleteing them once verified) I can state that it takes way too long to delete 500+ meg files.
--Dan
8gig is about the performance limit of ext2. After that it starts to take significant time to simply mount the disk read/write. (Lots of structure to copy into memory). 25+gig arrays take an indecent amount of time. And don't even talk about e2fsck.
When linux crashes (not if, this is a high load box) recovery time is astronomical.
They missed one point, though. No good way to backup large volumes. Dumpe2fs craps at about 8gig... lots of 64bit problems. Tarballing a 30 gig array as a "backup" solution just dosn't cut it... no good way to do incrementals.
Hopefully we'll see some decent large-filesystem support by 3.0, since theres a few different projects working on it already.
--Dan
What is needed is a technical solution to this, and the furor over "protect the children" will die down. If people quit getting porn shoved in their faces, they're happy. They don't care if it's a technical solution or an international treaty enforced by men in black helicopters, they just want it.
--Dan
This will be suppressed, of course. Just like "miracle" drugs, you'll have to go out of the country for it.
Most likely, it'll happen in some third world country where the people don't matter anyway. You may end up with a different body color, but that dosn't really matter if you're rich. (Ex. Michael Jackson)
--Dan
--Dan
and, unlike criminal court, you are guilty until you prove yourself innocent. The $50 limit is for when your wallet is physically stolen and you report it. It does not apply to someone copying your information and abusing it.
--Dan
In fact, it's not. At all. The sieving is done in parallel, but the actual matrix caclulations require one massive machine to do. Crays are nice because that's what they were designed to do. Simply tossing 2gig of ram into a dec alpha won't help you process the sparse matrix.
--Dan
we havn't been on the gold standard in forever. Someone flunked Social Studies, I take it.
Either have something to contribute, or leave.
--Dan
implement your standard NAT (including database of active connections) but round-robin your packets between them. Have a seperate connection between NAT servers where they announce their changes to the subnet pool.
That way, the only problem you have is a temporary 50% packetloss until your round-robin system drops the dead router.
Easy? No. Possible? Yes. Expensive? Definatly.
Interesting... I had some inital trouble making proxy ARP work properly in linux but after upgrading my net utils to match my kernel it worked properly. I think I can proxy arp for networks as well, but I havn't tried that.
If I misunderstood your problem, let me know. but yes, linux does handle ARP properly.
John dosn't even take the time to verify his spell-checker is outputting the right words! Somehow I doubt a call for proper punctuation will get very far. He's nigh-impossible to read because of it, too.
Lets take another "good idea" law, this one from the "war on drugs". This law gave the police the right to sieze any money from people carrying large amounts, by saying "anyone with lots of cash is buying drugs. Period. That's our probable cause."
Sounds reasonable, right? Only drug dealers would carry that kind of cash, of course, so no honest citizens would be affected...
Mind you, it dosn't end there. He quickly proves that he's not a drug dealer, and is released. No problem, right?
Nope. He forfit the entire loan. All of it. Even though he is an honest citizen, his money now belongs to the state. And they do not HAVE to give it back.
Do we really want to wait for a blatent abuse of this new law before we decide that it too is a bad idea?
Remember, the "War on Drugs" was to "save the children" from drug abuse. A noble goal, for sure, but it has failed and has done nothing but to increase the violence. Sometimes good ideas go astray, which is what the founding fathers understood. I wish we had their guidence today.
--Dan
Useful? 9 gig.
No? Try to fsck that and have the system back in a reasonable time.
What, 9 isn't enough pain? Try 25. See what happens when that once-in-a-blue-moon(that strangly happens every other week) kernel 'oops' happens and your primary storage array is down while e2fsck is looking at 100% of a 99.9% clean disk.
Is it possible to stop hyping linux for about another 12 months until there's something to hype about? Sorry to dissapoint the kiddiez, but it's NOT ready for serious applications. Personal servers, development workstations (Ha! if you want to produce code incompatable with everything else.) ... possibly samba/dhcp on a small office LAN. Scale beyond that and you start failing.
(Mind you, I said something negative about linux, so I must be a troll. Never mind the fact that I run my entire network on it so know firsthand just how "stable" it really is in production.
--Dan
Is also good for backing up a consistant filespace. "This system as it was at exactly midnight."
--Dan
Perl frontend. Note it's gtk-perl. I'd hazard a guess the actual gruntwork is done in C. --Dan
You are missing the point of the GPLed output. The idea is to prevent someone from taking GCC, patching it with a flag that dumps the entire source of gcc as a tarball to stdout and saying "It's the output of the program! Now it's under any licence I chose!".
Hence, compression, encryption or rot-13ing the strings dosn't change anything. --Dan
echo 'main() {exit(0);} // useless program' |
sed 's#//\(.*\)$#/*\1 */#'
I'll have to agree with you on this one. Yes, those chairs are wonderful. My friend got one as a present (hint hint) and I've had time to check it out. Much better then the $50-$150 office chairs I usually sit in. As soon as I find the spare cash, this is in my future. --Dan