Who uses Microsoft Antispyware?

[KrazyHorse]

So even if I just simply ssh, I get the same problem

I just mentioned the other two commands because usually I'm Xed in (since I'm working with a GUI programming editor)

[KrazyHorse]

And before you ask, yes, I do have write privileges on xyz, and write to files remotely all the time using other utilities.

[Asher]

Originally posted by Urban Ranger
Asher makes yet another assertion that he fails to substantiate, what else is new?

Oh, that's rich. I'm still waiting for you to post anything but rhetoric or op-ed from open source advocates.

This is the sixteenth time I've asked you explicitly for you to back up your claim that Windows is fundamentally flawed, and you've danced around it and ignored it sixteen times now.

It amuses me that you can misquote somebody's post even though that post is in plain sight. Just a little, though.

It seems your memory is failing. Here is a direct quote from you:

files under GNU/Linux cannot execute themselves the way they can under Windows. Thus, you will never get anything like e-mail viruses, slimeware that intall themselves behind your back, etc.

Seems to me like you're saying that "you will never get anything like e-mail viruses, slimeware that install themselves behind your back"?

Seems to me like you're avoiding looking like you made a foolish comment again. Quick, go edit it or delete the thread...

Double clicking on an e-mail attachment in Linux doesn't execute it. That's basic knowledge. So much for your Linux expertise.

Err... since when was Linux an email program?

Indeed, you are right. What does that have to do with what I wrote?

You said that "you will never get anything like e-mail viruses, slimeware that install behind your back". Consider this: you install an rpm called, say, "Kazaa". Let's hypothetically say Kazaa comes with adware called Gator, which is bundled in the rpm as well so it gets installed when you install Kazaa.

Explain to me how this is impossible to happen on Linux, because I package rpms and tarballs regularly with install scripts that can basically do whatever-the-hell we want them to do. If we were malicious, it may as well be spyware...

Mr MS Fanboy start throwing stuff out from the standard MS FUD manual.

Been there, seen that.

It's not FUD, it's common sense. Linux has less available software because it has less users. Viruses are software...so is spyware.

Oooh, another strawman. How many of them can you throw in one post?

You just told me you can never get viruses or spyware installed "behind your back" (ie, without explicit permission).

You have yet to substantiate this, and now you're dismissing my rebuttals saying that your opinion was outrageous as strawmen. Perhaps you should attend some remedial classes and brush up on your English skills, because clearly the words you are coherently typing are not your illusive arguments.

Laying on the ad hominems thick isn't going to help your case, Mr Milford. In fact, it is a sign of weakness on your part, because you don't have arguments and fact on your side.

Anyone with some semblance of intelligence and has seen more than a couple of these threads can see you up to your usual tricks. I've been throwing sensible arguments at you from every direction, and all you do is squirm, slide, and ignore without ever addressing them. You'll handily dismiss it as "FUD", throw out some rhetoric like "windows is fundamentally flawed", then disappear only to resurface with other claims like Linux can never have a virus...

You think way too highly of yourself. I just can't be arsed to deal with your lame fallacies most of the time. But today is Saturday, the office is quiet, and mild amusement is better than none.

How dumb do you think we are?

You make replies, but I don't think I've ever seen you contribute something concrete or even sensible in these debates. Every time, you have the same routine of proclaiming "FUD", saying all counterarguments are strawmen, that you don't have time to deal with the likes of me, then you disappear only to repeat again in a new thread.

It's getting old, and people can see you for what you really are. It helped quite a bit when you keep making glaring public gaffes.

[Asher]

Originally posted by KrazyHorse
Maybe one of you two compsci geniuses can help me with this:

Locked up in a room I don't have free access to is a Linux machine called xyz (it's on the pha.jhu.edu network, but to log on to it you need to have a separate account, which I have; incidentally, it's also not called xyz, but no matter)

Now, if I run a certain command-line interactive program (it's a Fortran binary) directly from xyz everything works out perfectly.

If, however, I'm logged into xyz remotely, from another Linux machine on the general pha.jhu.edu network, the program crashes and says something like "core dumped" (after I have input all my parameters in the interactive mode, and as soon as it begins calculating what I told it to).

Other programs work perfectly (even ones with GUIs etc). It's just any of the binaries provided in the HEALPix software package by Gorski

http://www.eso.org/science/healpix/content/HEALPix_Documentation_html/facilities.htm

Specifically, I've used smoothing and anafast, and both have the same problem.

To log on to xyz I do:

xhost +xyz
ssh xyz
setenv DISPLAY localhost:0.0 (this last and the xhost have no bearing on the problem, as omitting them leads to the same error)

Does it mention what kind of signal it generates when it crashes? ie, SIGSEV or "Segmentation violation" or "7" or something?

[Asher]

Off the top of my head, they may be restricting the stack size or maximum memory for remote users, but not for local users.

Do you know how to debug a core file? For example, "gdb core"?

Type "ulimit -a" and paste the results here please.

[KrazyHorse]

Can't do it right now. Am at home. Will try when I'm at school.

[KrazyHorse]

I think it just says segmentation fault when it ****s up.

But that seems to be a catchall term for every error ever.

[Asher]

Originally posted by KrazyHorse
I think it just says segmentation fault when it ****s up.

But that seems to be a catchall term for every error ever.

Probably stack/memory issue then.

Do you use the exact same account locally and remotely? Or is it a different account for remote?

[KrazyHorse]

Same account.

I'm supposed to have all the same privileges remotely, and I run much more computationally intensive stuff than this remotely (these routines take ~1 minute to run, and I sometimes run routines that take 20+ hours).

[Asher]

Depends how memory intensive the code is...if the 20+ hours use less memory you may not run into the same error.

Also, if the ~1 minute run program has a LOT of recursion, it can hit a stack wall pretty quickly. (though from the link you gave I don't think it's recursive, but iterative)

[Mr. Harley]

Asher - you are throwing in a confusor. The market penetration, and therefor the attractiveness, of a product is germane to it's security. If you have the most popular program, with a de facto monopoly (please note - de facto) you will have to create a more secure product that is harder to hack due to the fact you will have a larger number of people trying to hack it.

Thus if Microsoft has forty times as much market penetration as Linux, for example (I am just pulling a figure out of my butt, FYI) but spends ten times as much on security - it will probably be less secure. Note when I say spends, you have to note the amount of programming time, Linux has the advantage of all the free programming out there by devoted users. So UR has a legitimate point, without even discussing the problem of ActiveX controls and their integration directly into the OS.

[Asher]

UR's point about Linux being more secure is not really a point.

In fact, Linux leads Windows in terms of electronic break & enters for businesses. In part because their marketshare is more comparable in the small business server world, and Linux isn't as secure as UR wants to believe it is.

Independent third-party firm "mi2g" did a study late last year on it: http://www.windowsitpro.com/Article/...4398.html?Ad=1

According to a study the British security firm mi2g, Linux is the world's "most breached" OS and is exploited more frequently than Windows. The company recently analyzed more than 235,000 successful attacks against computers that were permanently connected to the Internet during the past year and concluded that Linux was responsible for most of the successful exploits.

"For how long can the truth remain hidden, that the great emperors of the software industry are wearing no clothes fit for the fluid environment in which computing takes place, where new threats manifest every hour of every day?" DK Matai, mi2g's executive chairman, said in a statement. "Busy professionals ... don't have the time to cope with umpteen flavors of Linux or to wait for Microsoft's Longhorn when Windows XP has proved to be a stumbling block in some well-chronicled instances."

According to mi2g, Linux-based computers accounted for more than 65 percent of all successful electronic attacks during the past year, whereas Windows-based systems were responsible for only 25 percent. Attacks against Berkeley Software Distribution (BSD)-based systems were successful less than 5 percent of the time. However, it's worth noting--although mi2g didn't--that BSD-based machines make up a small percentage of the installed base of permanently connected machines. In all probability, those machines weren't attacked simply because there was little incentive to do so, not because of any inherent superiority over Linux- or Windows-based systems.

The mi2g study also analyzed the impact of malware during the same time period and found that most malware attacks--about 60 percent--successfully targeted small businesses, whereas about 33 percent successfully targeted home users. Only 6 percent of malware attacks successfully targeted midsized businesses, whereas 2.5 percent successfully targeted enterprises, government agencies, and similar firms. According to the company, 459 successful malware attacks occurred during the past year, most of which targeted Windows-based systems. Malware rarely targeted BSD-based and Linux systems.

These electronic attacks are taking an economic toll. The firm says that electronic attacks such as Distributed Denial of Service (DDoS) attacks caused as much as $123 billion in damages during the past year. Malware attacks were responsible for $202 billion in damages during the same time period.

UR is spewing unfounded rhetoric. It's nothing new with him.

It's not a "no-brainer" like he wants it to be, Linux isn't some miracle that he makes it out to be, and he's not the expert he thinks he is...

[Ari Rahikkala]

Originally posted by Asher
Depends how memory intensive the code is...if the 20+ hours use less memory you may not run into the same error.

Also, if the ~1 minute run program has a LOT of recursion, it can hit a stack wall pretty quickly. (though from the link you gave I don't think it's recursive, but iterative)

Fortran 77 didn't even have explicit support for recursion, it seems unlikely to me that Fortran coders would use it much even these days...

KH: Aside differences in ulimits, are there differences in the environment between when you log in locally and when you log in remotely? I can't think of any specific environment variable that could cause your problem, but it should be easy to just diff the outputs of set from the different login shells and see if there's anything interesting there.

[KrazyHorse]

Hmm.

I don't know. All these computers were set up by the same people, so I can't see why one would have a different default environment than the other.

[Ari Rahikkala]

That's true, but if you're going to try to debug this anyway, comparing the environments is cheap and easy and you never know if it will reveal something. It probably won't, but at least then you will be happy knowing that you've eliminated one possible cause.