There is an obscene amount of POSIX compliant code that breaks without 64-bit specific fixes. I know because I have to fix it at work.
And don't get me started on DLL - possibly the worst design decision in all of Windows (and that's saying a lot!), particularly given how DLLs proved to be a massive attack vector on Windows systems.
Yes, dynamic libraries were such a bad idea.
Increased attack surface? ya, youbetcha. Trade-off? You try running 150 processes on a machine with full software stacks without sharing memory pages, and let me know how that works for you. I've done my share of hackery with linux shared objects too. There was always a trade-off in them, and Microsoft neither invented that tradeoff, or had a worse implementation of it than anyone else. They're simply the most visible.
On Windows and Linux, the file systems don't have this "resource fork" idea
Funny enough, NTFS has supported the idea since inception. But extensions weren't broke, so weren't fixed. And not to mention would have been sucky for any software relying on file extensions.
Also funny, OSX now does it by hiding extensions and interpreting the metadata inside of a folder.
Lastly, stop splitting hairs. C is a high level language
At its core, not really. It could be implemented purely as assembler macros. Now of course, compiler optimization is pretty far into the realm of magical, so language mapping to assembly code can be pretty perplexing... but unoptimized, C is really just an assembly macro language with fancy name mangling.
So in short, I think you should stop splitting hairs.
I've done a significant amount of reverse engineering in my time that (iPhone, Android, Moto Razr flip-phone)
Compiled C is easy to infer from assembler. C++ and above, is a different story. C++ is still pretty low-level really, but the runtime required to make it work puts it somewhere between a low level language and a high level language.
ALTERNATIVELY, provided that this notice is retained in full, this product
may be distributed under the terms of the GNU General Public License (GPL),
in which case the provisions of the GPL apply INSTEAD OF those given above.
Like I said, it's a bsd-like license, or GPL. Your call.
That's obscenely close. Comcast likely has much of its internal networking from the customer to the edge obscured via MPLS or other transport mechanisms. Your 6 hops don't complete in *zero*.246ms.
For the sake of being informative, google is also ever so slightly faster from here, as well.
[x@x ~]$ traceroute 8.8.8.8
traceroute to 8.8.8.8 (8.8.8.8), 30 hops max, 60 byte packets
1 x (x.x.x.x) 0.290 ms 0.334 ms 0.405 ms
2 x (x.x.x.x) 0.314 ms 0.385 ms 0.468 ms
3 x (x.x.x.x) 0.419 ms 0.506 ms 0.584 ms
4 six.sea01.google.com (206.81.80.17) 0.315 ms 0.339 ms 0.357 ms
5 108.170.245.113 (108.170.245.113) 0.262 ms
- 108.170.245.97 (108.170.245.97) 1.307 ms
- 108.170.245.113 (108.170.245.113) 0.275 ms
6 108.170.237.189 (108.170.237.189) 0.230 ms
- 209.85.250.19 (209.85.250.19) 0.667 ms
- 209.85.246.219 (209.85.246.219) 0.668 ms
7 google-public-dns-a.google.com (8.8.8.8) 0.172 ms 0.188 ms 0.192 ms
Realistically, I think their anycast node just responds quicker. I doubt there's any difference in latency across the pipe.
If you're getting 40% packet loss, the ping times would be higher
That's dependent on the amount of buffer bloat you have. Ideally, no, the ping times won't be different.
or intermittent
Absolutely- like missing 40% of the time....................
It's still a better metric for the end user for DNS than bandwidth.
End user? Yes. Though again, you're going to notice a saturated link long before you notice an extra 40ms of latency in DNS RTT.
Sure, typically fast link means something else - but we have context here.
I'd argue incorrect, or at best highly unorthodox usage, even given the context. Full disclosure, I am a network engineer. I do this for a living. My DNS infrastructure hosts 12285 domains, and I'm the head engineer for an AS with approximately 8000 customers. I'm not talking out of my ass.
I'd also like to add, that if we fixed the ability for money to buy actual governmental power, we may see a fundamental shift in the economy anyway.
Ideas like trickle-down may actually work as planned, and our obscene wealth disparity may begin to shift back to a more reasonable distribution once there's less reason to accumulate infinite capital, and a real fear over the power of your employees over the government.
It's not important at all for DNS queries. It's important for DNS servers. I run 8 authoritative nameservers. You know what's worse than 100ms of latency? 40% packet loss because you don't have the bandwidth to handle the queries.
Typically, when someone says, That's a fast link. they're referring to bandwidth. I see that you were not;)
So, you can't say that governments are corruptible and also say that isn't an indictment against government.
Yes, I can. Because governments are governed by laws and constitutions. The kind of corruption our government suffers from is completely legal and sanctioned, and it's the fault of the voters.
We always have to be wary of the fact that when we put big groups of people in charge of stuff, we're going to have actions that are not necessarily in our collective best interests.
Yes, agreed, however, that's a lot further down the line. We can discuss further fears once the corruption is actually fucking illegal.
So where is capitalism implicated in this? Because that is where the money comes from? So what?
So what? I'm not sure how you keep coming to that. Like capitalism is innocent because all it did was order the hit, not pull the trigger.
You don't think there would be a different power concentration if the US government controlled the economy directly?
I have no idea, and I'm in no way interested in finding out... I'm no advocate for government controlled economy. I'm an advocate against plutocracy and corporatocracy. I'm an advocate for strictly limiting the power our privately owned economy can actually wield.
The government is the problem, along with it's corporate extensions.
Yes, along with its corporate extensions.
Excise them. Burn them, and then vaccinate it against them. We control the governments, we just refuse to exercise that control. We don't even have theoretical control of the corporations until we assert or control over the government.
Sanders amendment IS a good place to start. The right place to start. But still- attacking "government" in general because government has been infected, and can be infected is ridiculous. Inoculate it. Don't kill it.
Governments, and bodies of humans are corruptible, period. Everyone has a price.
With governments, at least, you have the theoretical ability to vote in change until you make attempting that corruption punishable by fucking death.
The fact that the government is currently corrupt isn't an indictment against government, it's an indictment against THIS government. Replace it with one that is serious about fighting corruption. This government is corrupt because we are sheep, too stupid to care as long as we have cud to chew on.
Also- it works fine for DNS. We've been playing with it for a little bit.
It's in the global BGP tables, so you're going to be able to access it basically anywhere, except possibly a few networks operated by morons, or behind equipment that made the unfortunate choice of using 1.1.1.0 a management prefix internally.
Oh it's absolutely attracting it.
Prior to 1.1.1.0/24 becoming a global routed prefix again, that traffic was blackholed in every individual AS.
Now that cloudflare is announcing that block to me, we are routing that traffic to them. There really isn't any more accurate way of putting it other than that they are attracting it.
Link-local addresses exist for this reason. 169.254.0.0/16.
It's used for IPv4 zeroconf communications, but that's just an application of it. It's purpose is for non-routed link-local communications.
Na. It's anycast. Your ping is dependent upon how close you are to the closest node. Being I peer with cloudflare at the SIX, i'm very close to my closest node.
[x@x ~]$ traceroute 1.1.1.1
traceroute to 1.1.1.1 (1.1.1.1), 30 hops max, 60 byte packets
1 x (x.x.x.x) 0.232 ms 0.313 ms 0.371 ms
2 x (x.x.x.x) 0.295 ms 0.381 ms 0.466 ms
3 x (x.x.x.x) 27.807 ms 27.894 ms 28.005 ms
4 six.as13335.com (206.81.81.10) 0.293 ms 0.292 ms 0.292 ms
5 1dot1dot1dot1.cloudflare-dns.com (1.1.1.1) 0.212 ms 0.213 ms 0.246 ms
I'm a network engineer, so I am not remotely justifying what I'm about to describe. I'm the chief engineer on several large residential fiber to the home deployments, and as such get to play around a lot in not-off-the-shelf CPE equipment. You'd be amazed how much I see 1.1.1.1 used. It confused me for a while, but now I get it. If you need an RFC1918 address that you're basically guaranteed no user or ISP back end configuration will overlap with- guess what.
The current equipment I'm working on (and have just discovered a *massive* vulnerability in) use 1.1.1.0/30 for communications between the main Broadcom SOC and a quantenna 5ghz wlan SOC.
No, it doesn't sound like that at all.
It sounds like I'm arguing that not taking a hard line against immigration is not diametrically opposed to cleaning up organized crime.
The diametric opposition you claim is in fact a farce. An example of bad logic, or an attempt at being manipulative. Not sure which.
You realize that those two things are diametrically opposed? I mean, you admit that one of the problems is the government. Government interference in a market is a socialism thing.
Christ that logic hurts my brain.
Let me get this straight-
If the mob pays an illegal immigrant to shake you up for protection money, illegal immigration is the problem, not organized crime?
I'm pretty agnostic in the operating system department. I use all 3 on a regular basis, though I do primarily use Linux. I've come to realize they all are good in some ways, all suck in some ways, and they all have this in common- some things on them don't work real great out of the box. You can descend into the realms of janky hackery to make them work, (compiling kernels, modules, modifying.INF files, plists) but you'll pay for it in the overall polished experience of the OS. Which sucks, but that's life. Operating system bickering is partisanship these days. People rabidly defend their favorite sports team with zero grasp on reality. They'll pretend things are true about their favorite sports team even if they aren't, and they'll make up things that aren't true about their rival, even if they are. All the while, people like me are just sitting here keeping mental lists of what we can do in each OS easiest, and watching people like you call people morons for daring to criticize your precious Red Sox. As ridiculous as OP may have been, you're still attacking him by perpetuating a bad argument. Maybe to him, having to recompile his kernel (though I can't remember the last time I've had to do that) is less bothersome than having his stupid external display stop working, or whatever. Maybe the vendor *should* have tried harder to make sure they were involved in Mac development so that they wouldn't have been blind sided by this... though given the Mac market share, it's understandable that Mac support is a second class citizen to them. Don't be offended by that. But one thing is for sure- you can spot the rabid fan with their constant use of whataboutisms and double standards. Apple should do more QC before a release. If this device was big enough to make the news, then it should have been big enough for someone in Apple to say "hey, let's not break this."
That goes for Microsoft as well, though I have no opinion on who does it better. They all feel about the same to me. End of the day- updates suck, and every operating system is an idiosyncratic pain in the ass.
I will say this, though. My boss just got a brand new Mac Pro (I think? huge ass computer-and-monitor-in-one, 10 core monster with like 128GB of ram) and that is one sexy ass machine. I'm still 4x more productive than him on my Linux machine, but hey- he's my boss. More power to him.
Slashdot Mirror
There is an obscene amount of POSIX compliant code that breaks without 64-bit specific fixes. I know because I have to fix it at work.
And don't get me started on DLL - possibly the worst design decision in all of Windows (and that's saying a lot!), particularly given how DLLs proved to be a massive attack vector on Windows systems.
Yes, dynamic libraries were such a bad idea.
Increased attack surface? ya, youbetcha. Trade-off? You try running 150 processes on a machine with full software stacks without sharing memory pages, and let me know how that works for you. I've done my share of hackery with linux shared objects too. There was always a trade-off in them, and Microsoft neither invented that tradeoff, or had a worse implementation of it than anyone else. They're simply the most visible.
On Windows and Linux, the file systems don't have this "resource fork" idea
Funny enough, NTFS has supported the idea since inception. But extensions weren't broke, so weren't fixed. And not to mention would have been sucky for any software relying on file extensions.
Also funny, OSX now does it by hiding extensions and interpreting the metadata inside of a folder.
Lastly, stop splitting hairs. C is a high level language
At its core, not really. It could be implemented purely as assembler macros. Now of course, compiler optimization is pretty far into the realm of magical, so language mapping to assembly code can be pretty perplexing... but unoptimized, C is really just an assembly macro language with fancy name mangling.
So in short, I think you should stop splitting hairs.
I've done a significant amount of reverse engineering in my time that (iPhone, Android, Moto Razr flip-phone)
Compiled C is easy to infer from assembler. C++ and above, is a different story. C++ is still pretty low-level really, but the runtime required to make it work puts it somewhere between a low level language and a high level language.
ALTERNATIVELY, provided that this notice is retained in full, this product may be distributed under the terms of the GNU General Public License (GPL), in which case the provisions of the GPL apply INSTEAD OF those given above.
Like I said, it's a bsd-like license, or GPL. Your call.
That's obscenely close. Comcast likely has much of its internal networking from the customer to the edge obscured via MPLS or other transport mechanisms. Your 6 hops don't complete in *zero*.246ms.
Am I missing anything?
You are. unrar's source is licensed under either a BSD-like license, or the GPL, your choice.
There are plenty [symantec.com] of [schalley.eu] examples [experts-exchange.com] of people suggesting ping to 1.1.1.1 as a delay in batch scripting.
That is literally one of the dumbest fucking things I've ever heard. And from symantec, no less. Terrible.
For the sake of being informative, google is also ever so slightly faster from here, as well.
[x@x ~]$ traceroute 8.8.8.8
traceroute to 8.8.8.8 (8.8.8.8), 30 hops max, 60 byte packets
1 x (x.x.x.x) 0.290 ms 0.334 ms 0.405 ms
2 x (x.x.x.x) 0.314 ms 0.385 ms 0.468 ms
3 x (x.x.x.x) 0.419 ms 0.506 ms 0.584 ms
4 six.sea01.google.com (206.81.80.17) 0.315 ms 0.339 ms 0.357 ms
5 108.170.245.113 (108.170.245.113) 0.262 ms
- 108.170.245.97 (108.170.245.97) 1.307 ms
- 108.170.245.113 (108.170.245.113) 0.275 ms
6 108.170.237.189 (108.170.237.189) 0.230 ms
- 209.85.250.19 (209.85.250.19) 0.667 ms
- 209.85.246.219 (209.85.246.219) 0.668 ms
7 google-public-dns-a.google.com (8.8.8.8) 0.172 ms 0.188 ms 0.192 ms
Realistically, I think their anycast node just responds quicker. I doubt there's any difference in latency across the pipe.
If you're getting 40% packet loss, the ping times would be higher
That's dependent on the amount of buffer bloat you have. Ideally, no, the ping times won't be different.
or intermittent
Absolutely- like missing 40% of the time....................
It's still a better metric for the end user for DNS than bandwidth.
End user? Yes. Though again, you're going to notice a saturated link long before you notice an extra 40ms of latency in DNS RTT.
Sure, typically fast link means something else - but we have context here.
I'd argue incorrect, or at best highly unorthodox usage, even given the context. Full disclosure, I am a network engineer. I do this for a living. My DNS infrastructure hosts 12285 domains, and I'm the head engineer for an AS with approximately 8000 customers. I'm not talking out of my ass.
I'd also like to add, that if we fixed the ability for money to buy actual governmental power, we may see a fundamental shift in the economy anyway.
Ideas like trickle-down may actually work as planned, and our obscene wealth disparity may begin to shift back to a more reasonable distribution once there's less reason to accumulate infinite capital, and a real fear over the power of your employees over the government.
I was going to say the same thing. Cheers, lol.
It's not important at all for DNS queries. It's important for DNS servers. I run 8 authoritative nameservers. You know what's worse than 100ms of latency? 40% packet loss because you don't have the bandwidth to handle the queries.
;)
Typically, when someone says, That's a fast link. they're referring to bandwidth. I see that you were not
So, you can't say that governments are corruptible and also say that isn't an indictment against government.
Yes, I can. Because governments are governed by laws and constitutions. The kind of corruption our government suffers from is completely legal and sanctioned, and it's the fault of the voters.
We always have to be wary of the fact that when we put big groups of people in charge of stuff, we're going to have actions that are not necessarily in our collective best interests.
Yes, agreed, however, that's a lot further down the line. We can discuss further fears once the corruption is actually fucking illegal.
So where is capitalism implicated in this? Because that is where the money comes from? So what?
So what? I'm not sure how you keep coming to that. Like capitalism is innocent because all it did was order the hit, not pull the trigger.
You don't think there would be a different power concentration if the US government controlled the economy directly?
I have no idea, and I'm in no way interested in finding out... I'm no advocate for government controlled economy. I'm an advocate against plutocracy and corporatocracy. I'm an advocate for strictly limiting the power our privately owned economy can actually wield.
The government is the problem, along with it's corporate extensions.
Yes, along with its corporate extensions.
Excise them. Burn them, and then vaccinate it against them. We control the governments, we just refuse to exercise that control. We don't even have theoretical control of the corporations until we assert or control over the government.
Sanders amendment IS a good place to start. The right place to start. But still- attacking "government" in general because government has been infected, and can be infected is ridiculous. Inoculate it. Don't kill it.
Governments, and bodies of humans are corruptible, period. Everyone has a price.
With governments, at least, you have the theoretical ability to vote in change until you make attempting that corruption punishable by fucking death.
The fact that the government is currently corrupt isn't an indictment against government, it's an indictment against THIS government. Replace it with one that is serious about fighting corruption. This government is corrupt because we are sheep, too stupid to care as long as we have cud to chew on.
Also- it works fine for DNS. We've been playing with it for a little bit.
It's in the global BGP tables, so you're going to be able to access it basically anywhere, except possibly a few networks operated by morons, or behind equipment that made the unfortunate choice of using 1.1.1.0 a management prefix internally.
Oh, you meant fast as in RTT fast... My bad. I thought you meant "throughput" fast. (wide pipe vs. short pipe)
That's my bad.
Oh it's absolutely attracting it.
Prior to 1.1.1.0/24 becoming a global routed prefix again, that traffic was blackholed in every individual AS.
Now that cloudflare is announcing that block to me, we are routing that traffic to them. There really isn't any more accurate way of putting it other than that they are attracting it.
Heh. In the network engineering industry, dropping the host address zeros is common practice when talking about prefixes.
Link-local addresses exist for this reason. 169.254.0.0/16.
It's used for IPv4 zeroconf communications, but that's just an application of it. It's purpose is for non-routed link-local communications.
Does their frugivorism preclude their having a desire for vengeance?
That's a fast link.
Na. It's anycast. Your ping is dependent upon how close you are to the closest node. Being I peer with cloudflare at the SIX, i'm very close to my closest node.
[x@x ~]$ traceroute 1.1.1.1
traceroute to 1.1.1.1 (1.1.1.1), 30 hops max, 60 byte packets
1 x (x.x.x.x) 0.232 ms 0.313 ms 0.371 ms
2 x (x.x.x.x) 0.295 ms 0.381 ms 0.466 ms
3 x (x.x.x.x) 27.807 ms 27.894 ms 28.005 ms
4 six.as13335.com (206.81.81.10) 0.293 ms 0.292 ms 0.292 ms
5 1dot1dot1dot1.cloudflare-dns.com (1.1.1.1) 0.212 ms 0.213 ms 0.246 ms
I'm a network engineer, so I am not remotely justifying what I'm about to describe. I'm the chief engineer on several large residential fiber to the home deployments, and as such get to play around a lot in not-off-the-shelf CPE equipment. You'd be amazed how much I see 1.1.1.1 used. It confused me for a while, but now I get it. If you need an RFC1918 address that you're basically guaranteed no user or ISP back end configuration will overlap with- guess what.
The current equipment I'm working on (and have just discovered a *massive* vulnerability in) use 1.1.1.0/30 for communications between the main Broadcom SOC and a quantenna 5ghz wlan SOC.
No, it doesn't sound like that at all.
It sounds like I'm arguing that not taking a hard line against immigration is not diametrically opposed to cleaning up organized crime.
The diametric opposition you claim is in fact a farce. An example of bad logic, or an attempt at being manipulative. Not sure which.
You realize that those two things are diametrically opposed? I mean, you admit that one of the problems is the government. Government interference in a market is a socialism thing.
Christ that logic hurts my brain.
Let me get this straight-
If the mob pays an illegal immigrant to shake you up for protection money, illegal immigration is the problem, not organized crime?
I'm pretty agnostic in the operating system department. I use all 3 on a regular basis, though I do primarily use Linux. I've come to realize they all are good in some ways, all suck in some ways, and they all have this in common- some things on them don't work real great out of the box. You can descend into the realms of janky hackery to make them work, (compiling kernels, modules, modifying .INF files, plists) but you'll pay for it in the overall polished experience of the OS. Which sucks, but that's life. Operating system bickering is partisanship these days. People rabidly defend their favorite sports team with zero grasp on reality. They'll pretend things are true about their favorite sports team even if they aren't, and they'll make up things that aren't true about their rival, even if they are. All the while, people like me are just sitting here keeping mental lists of what we can do in each OS easiest, and watching people like you call people morons for daring to criticize your precious Red Sox. As ridiculous as OP may have been, you're still attacking him by perpetuating a bad argument. Maybe to him, having to recompile his kernel (though I can't remember the last time I've had to do that) is less bothersome than having his stupid external display stop working, or whatever. Maybe the vendor *should* have tried harder to make sure they were involved in Mac development so that they wouldn't have been blind sided by this... though given the Mac market share, it's understandable that Mac support is a second class citizen to them. Don't be offended by that. But one thing is for sure- you can spot the rabid fan with their constant use of whataboutisms and double standards. Apple should do more QC before a release. If this device was big enough to make the news, then it should have been big enough for someone in Apple to say "hey, let's not break this."
That goes for Microsoft as well, though I have no opinion on who does it better. They all feel about the same to me. End of the day- updates suck, and every operating system is an idiosyncratic pain in the ass.
I will say this, though. My boss just got a brand new Mac Pro (I think? huge ass computer-and-monitor-in-one, 10 core monster with like 128GB of ram) and that is one sexy ass machine. I'm still 4x more productive than him on my Linux machine, but hey- he's my boss. More power to him.