Instead of using/etc/hosts, get junkbuster and put the hostnames in the blockfile (you can also put URL patterns, and other stuff in there). Turn logging on, and then grep for how many lines in the logfile end in ' crunch\!$' vs. how many don't.
That'll probably be next - they'll start suing people coding proxies (JunkBuster, Squid, Proximitron,...) saying it's violating whatever intellectual property right they care to make up to cover shoving ads down your throat.
This is exactly what they don't seem to understand. Does anyone seriously expect me to purchase something after an ad for it has been rudely shoved in my face, interrupting whatever I was trying to do? If anything, the produce just got added to the list of things I would never consider buying.
On the other hand, having written your own tripwire (and don't call it tripwire) - the hacker will not know this, not be familiar with your mechanism, and thus, will be unable to circumvent it.
I wouldn't be this confident. Don't you think that a skilled hacker (not some s'kiddie with a pre-fab rootkit) would check the running process list, check the root cron, check your log directory, etc., etc., for anything that looks suspicious? Especially if your program is just a script, once he locates it (it's path is going to be listed somewhere, either in ps or crontab, right?) he could just read the source and figure out exactly how to beat it.
I would think the most obvious attack would be to add ignore-patterns to the configuration file; or, since you're likely to catch a modification to the that (I assume you mod it yourself occasionally), just modify the script to ignore whatever he's done to your filesystem. Or he could update the checksums in your DB by hand. And so on.
Looks like Altavista has a problem understanding the chunked HTTP transfer-encoding method. I went to look at my own results, and scattered throughout the description of my page are 3-digit hexadecimal numbers.
In a page downloaded via "chunked" transfer-encoding, the response body is broken down into small chunks, each chunk being prefixed by their length, represented as a three-digit hexadecimal number. The three chunk-lengths AV is showing as text are c45, 138, and fe4. Telnetting to my site by hand confirmed that these are the first three chunk sizes Apache is using.
Also, I suppose it is just possible that Win2K has a _theoretical_ unlimited interrupt latency, but I don't think it is _effectively_ unlimited, because otherwise I wouldn't be able to do anything on my computer, would I?
What that means is that one process can take complete control and prevent any interrupts from being handled. Thus, the latency for the next interrupt to arrive could be unlimited.
The law would effectively prevent you from connecting them to the (now DRM-enabled) Internet, because the old computers cannot speak the right protocols.
It looks like a kernel coder was intentionally parodying an older actual error message that would say lp1 on fire. On some older systems, printers would report to the computer when they become overheated, and this was the (rather alarmist) error that the system would display.
This is most likely the result of someone calling an error-handling routine (On Unix, it's strerror(errno), not sure about BeOS), without making sure that, after an error supposedly occured, errno actually contained a proper error code. Every OS has their own variants of this, it seems. On Windows, it's "The operation completed successfully," on most Unices it's usually "No error" or "Success." In some versions, it actually says "Error 0" - that might confuse a user, but at least a programmer can infer he accidentally passed 0 to strerror().
Yep, and this same bug manifests itself on Unix-based OSes in the form of "Error: No error" or "Error: Success" error messages. If you call strerror(errno) while errno is 0, you'll get something stupid like this.
Some versions of cat just print a generic "cannot open" message when a file cannot be opened for whatever reason. Other versions will use the standard C errno codes so you'll get a meaningful error message, such as "No such file or directory", "Is a directory", or "Permission denied."
The application "unknown" has unexpectedly quit because an error of type 11 occurred.
These errors would pop up immediately after an application suicided. For some reason Apple was never able to make the correct application's name appear in the dialog box, hence the "unknown".
Every once in a while I see even worse errors, like a dialog that only says "-43" or "-139" in it. C'mon, don't they have strerror() in Cupertino?
"Without copyright, I could legally take GPL'ed code, modify it, compile it, and release the executable without releasing the new source code."
Yeah, my original statement wasn't entirely correct. But, here's another point to consider: without copyright, if I were able to acquire your protected source, somehow, with or without your permission, you'd have no legal way of preventing me from distributing it.
Without copyright there wouldn't need to be a GPL -- the GPL essentially says that a piece of software is completely, freely available, except to those who wish to wrap it in their own proprietary works.
An easy countermeasure to this is a proxy that makes the GET request for the banner, but just directs it to the bitbucket instead of your browser window. If you use a proxy to save bandwidth, this would not be very useful, but if you just use a proxy to protect your sanity from the latest incarnation of Punch the Monkey, this would work fine.
Slashdot Mirror
Instead of using /etc/hosts, get junkbuster and put the hostnames in the blockfile (you can also put URL patterns, and other stuff in there). Turn logging on, and then grep for how many lines in the logfile end in ' crunch\!$' vs. how many don't.
telnet... sounds like some kind of DMCA-violating circumvention device to me...
That'll probably be next - they'll start suing people coding proxies (JunkBuster, Squid, Proximitron, ...) saying it's violating whatever intellectual property right they care to make up to cover shoving ads down your throat.
This is exactly what they don't seem to understand. Does anyone seriously expect me to purchase something after an ad for it has been rudely shoved in my face, interrupting whatever I was trying to do? If anything, the produce just got added to the list of things I would never consider buying.
Google is your friend. 460,000 times.
I would think the most obvious attack would be to add ignore-patterns to the configuration file; or, since you're likely to catch a modification to the that (I assume you mod it yourself occasionally), just modify the script to ignore whatever he's done to your filesystem. Or he could update the checksums in your DB by hand. And so on.
images2.slashdot.org
/*.*/banner/
/*.*/.*\.swf
(The last line kills all Flash files, so be careful.)doubleclick.net
Looks like Altavista has a problem understanding the chunked HTTP transfer-encoding method. I went to look at my own results, and scattered throughout the description of my page are 3-digit hexadecimal numbers.
In a page downloaded via "chunked" transfer-encoding, the response body is broken down into small chunks, each chunk being prefixed by their length, represented as a three-digit hexadecimal number. The three chunk-lengths AV is showing as text are c45, 138, and fe4. Telnetting to my site by hand confirmed that these are the first three chunk sizes Apache is using.
SGML? How about just straight-up UTF-8?
The law would effectively prevent you from connecting them to the (now DRM-enabled) Internet, because the old computers cannot speak the right protocols.
Caesar's assassination led to a civil war, which led to the victor of that war, Octavian, becoming an emperor.
Emulator?
...)
(WINE, SoftPC,
It looks like a kernel coder was intentionally parodying an older actual error message that would say lp1 on fire. On some older systems, printers would report to the computer when they become overheated, and this was the (rather alarmist) error that the system would display.
This is most likely the result of someone calling an error-handling routine (On Unix, it's strerror(errno), not sure about BeOS), without making sure that, after an error supposedly occured, errno actually contained a proper error code. Every OS has their own variants of this, it seems. On Windows, it's "The operation completed successfully," on most Unices it's usually "No error" or "Success." In some versions, it actually says "Error 0" - that might confuse a user, but at least a programmer can infer he accidentally passed 0 to strerror().
Yep, and this same bug manifests itself on Unix-based OSes in the form of "Error: No error" or "Error: Success" error messages. If you call strerror(errno) while errno is 0, you'll get something stupid like this.
Some versions of cat just print a generic "cannot open" message when a file cannot be opened for whatever reason. Other versions will use the standard C errno codes so you'll get a meaningful error message, such as "No such file or directory", "Is a directory", or "Permission denied."
Are you thinking of:
The application "unknown" has unexpectedly quit because an error of type 11 occurred.
These errors would pop up immediately after an application suicided. For some reason Apple was never able to make the correct application's name appear in the dialog box, hence the "unknown".
Every once in a while I see even worse errors, like a dialog that only says "-43" or "-139" in it. C'mon, don't they have strerror() in Cupertino?
Try Pennsylvania. The state government is forcing WorldCom to block content from outside the U.S. because it finds it objectionable.
"Without copyright, I could legally take GPL'ed code, modify it, compile it, and release the executable without releasing the new source code."
Yeah, my original statement wasn't entirely correct. But, here's another point to consider: without copyright, if I were able to acquire your protected source, somehow, with or without your permission, you'd have no legal way of preventing me from distributing it.
Copyrights are life + 75 right now. Under the current system, people who live shorter lives already have shorter copyright terms.
Without copyright there wouldn't need to be a GPL -- the GPL essentially says that a piece of software is completely, freely available, except to those who wish to wrap it in their own proprietary works.
No.
No, wait, I mean yes. Yes. Er...
An easy countermeasure to this is a proxy that makes the GET request for the banner, but just directs it to the bitbucket instead of your browser window. If you use a proxy to save bandwidth, this would not be very useful, but if you just use a proxy to protect your sanity from the latest incarnation of Punch the Monkey, this would work fine.