Their firewall seems like more of an automatic proxy than a simple port filter. The http requests are verified, and it does something to the telnet connections because it always sends "Eagle Secure Gateway" to telnet client before connecting
This DNS tunnel is neat, especially since I have always been a fan for HTTP tunnels. The only reason to build a tunnel is to circumvent security. But one must remember the hacker ethic:
It's ok to get around a little security in order to get more work done
One great example of this is in a paranoid school or company that firewalls outgoing traffic. They allow telnet, but not SSH. My home machine only accepts SSH. (for obvious reasons) A little tunneling, and presto, I can secure shell to my home computer.
Especially nifty is using http tunnel to establish a secure shell then using the secure shell to tunnel other protocols with encryption
Well, now we know that the whole subliminal message and reverse psychology scene is doomed. The new craze for fooling moderators is posting DeCSS! Soon one will merely need to put a DeCSS link in their.sig to achieve instant +5, Redundant status!
Well I know DeCSS has been posted more than a few times here on slashdot, and a few users (including me) have it in their user profiles. It's also in an everything2 node. BSI is proud to host DeCSS!
We should instead be blaming the variable-length non-typesafe argument lists of C. This falls into the same category as pointers- something that is very useful and powerful, but can't be trusted to the outside.
The arg lists could be made safer with some modification to the compiler and libc, but could it be done portably?
This could all too quickly turn into some strange C and C++ hybrid, or a C/Java holy war. Yipes!
So, a question to all... how to you write your code so that it's flexible enough for translation, but not open to attack?
How about a function, sort of like printf, in which you specify the number and type of args seperately from their position?
Something like this:
untrustedfmt = "Can't open file %";
secureprintf(untrustedfmt,"%s",filename);
Basically, substituting each % in the input with a format specifier from the hard-coded string. If there are too many "%" in the input, it doesn't crash. ("Error printing error message")
Some variation on this could allow the order of the format specifiers to be rearranged. Heck, maybe something more fancy like
"Error number $errnum: Can't open file $file"
Already easy to do in perl, but I've never seen this done in C.
I installed Mandrake 7.1 a few days ago after
finally getting an install disk (ironic, upgrades always seem to come out after I've just downloaded/borrowed the current version:)
The installer definitely surprised me, but it was pretty nifty. The configuration tools seem pretty
good for anyone who doesn't want to poke in config files, but I'll probably dual boot with Debian. Only complaint I had was the default Gnome installation. I don't know what the problem was, but the panel sucked up about 60MB of memory. Worked just fine after upgrading to the Helix gnome RPMs though. I was pleasently surprised to find that it came with XFree 4.0 (which I need for my TNT2 card) but it choked on the NVidia drivers. I had to install X from the official distribution to get the NVidia working.
In short, Mandrake's a great distro for ease-of-configuration, but like all distros needs some incantations for the strange hardware
I imagine that HP is using Windows GDI to accomplish the same thing. That way the printer is nothing more than a mechanism for moving paper in front of a print head, it
doesn't require a whole lot (any?) programmable logic or the cost involved in developing it.
This is pretty much what HP's PPA printers (700 series and some others) do. And, they do have a linux driver that in addition to ghostscript can render postscript to them.
I'm kind of doubtful how well this technology works...
But, if it does work as well as they say, what about displaying the song's fingerprint next to each song on an OpenNAP server? This would make it easy to verify that it's the right one before you download.
I don't know what I'd do without a keyboard. How
the heck do you dictate Perl?
As for the Windows keys, I'm using a keyboard that's old enough not to have them. If they really do break as often as you say, maybe you're just mistreating it. Only keyboards I have that don't work are cheap ones I bought just to steal parts from.
Yeah, I figured it out a few minutes after I posted...:)
Still, many people were confused about the boot time, and "It will be useless until it boots faster" posts modded up didn't help much.
Just trying to clarify things a bit.
Constrast that with the fine, fine output of those 100 engineers (although I suspect that their Linux driver "team" numbers in the ones). The NVidia binary-only driver runs
only on particular kernel revisions, and does not allow the user to switch between SMP and UP operation. Worse, NVidia could choose to stop supporting Linux at any
time, and there's nothing we could do about. Since we have the specs for the G400, we can support that card forever.
The XFree module and the GL libs are binary only, but the source to the small kernel module is distributed. Admittedly, it has problems on many kernels, but at least they let you try to compile it on which ever kernel you happen to be using.
Which linksys card are you referring to? Yes, they put a copy of tulip on the disks. But, I own three LNE100TX cards (the 10/100 PCI thingies) and even though they work on linux, under a high load they become unusable until a hard-reboot. NFS and most other UDP protocols won't work under them at all. Sure it's probably the driver's problem, but my point is that just because a company puts forth some good-will gestures to linux doesn't mean they actively support it.
I don't think this applies to nVidia at all though. I have 2 TNT2 cards (go q3a!) and the nVidia drivers are great. yes, they're binary only, and yes they crash sometimes. Hopefully this will be fixed soon, though, and at least they're fast. I'm especially impressed that a hardware company can get their act together and maintain a driver source code base that compiles for multiple OSs.
Re:So what does linux on a microcontroller offer?
on
Microcontroller Linux
·
· Score: 1
Running traditional linux apps seems to be an unlikely goal - I'm not paying to run a webserver from my cell phone and X applications are not ideal for a 128x72 LCD
screen.
Quite possibly, the need for a real filesystem might be a good dividing line. Where you need such a thing, linux gives you lots of goodies for providing and using the
filesystem. In a system which has no such need, linux looks too much like an overwight dinosaur.
While linux isn't the smallest kernel available,
it can be pretty small. The kernel used in ucLinux usually compiles to around 300k. One of
the largest aspects of linux on desktops, though,
is glibc. uclinux uses a stripped-down libc instead. This means that while it doesn't take much work to port most apps, not much will simply
recompile for ucLinux. I have found busybox, the
util suite used on the debian rescue disks, to
work well in ucLinux.
ucLinux is definitely not designed to run desktop apps on small CPUs. It's more like a happy medium
between the ultra-tiny kernels and the compatibility of the standard linux kernel.
Ah, I've been hoping this would come up on slashdot sometime...
To fill in some of the gaps in the/. story: ucLinux is a port of the 2.0 linux kernel to MMU-less architectures. It's mainly run on the Motorola 68000 series, but it is also used quite a bit on the (much faster) motorola coldfire chips.
The ucSimm is based on the motorola 68EZ328 Dragonball chip, with onboard memory controller
and LCD controller. It's the same chip used in
most Palm computers.
I haven't used the ucSimm, but I have first-hand uclinux experience with my own hardware. (kiwi.sourceforge.net) It runs pretty nicely, and it definitely doesn't take 30 minutes to boot. Don't know where that came from. Because most of the time-consuming things like loadable modules usually aren't included in the boot process, it only takes seconds to boot. Most of the time is spent decompressing the initial filesystem into DRAM.
ucLinux has been used for many things, including industrial control (microcontroller replacement) and robotics. I have been working, slowly, to
make some type of usable PDA based on ucLinux. The Kiwi has been my crude but mostly operational attempt at hardware, and I've been working on a small and efficient GUI system, PicoGUI (sorry i haven't updated these pages in a while. Most of the fun stuff is taking place in the sourceforge CVS)
If the first one is bigger than the second one, skip to page two. Otherwise, continue with the instructions below.
Don't forget about Hypercard!
put the value of the object named 'field 1' into the variable 'x' if it is less than 50
vs if (*field1<50) x = *field1;
And the worst thing about hypercard was that even though the commands were supposed to be more english-like, they still had a firm syntax. It was much more inconsistant than C so i always had to look it up...
Slashdot Mirror
Maybe if people start to realize that there should be an alternative to PalmOS, it will be good for the ucLinux and PicoGUI projects?
Their firewall seems like more of an automatic proxy than a simple port filter. The http requests are verified, and it does something to the telnet connections because it always sends "Eagle Secure Gateway" to telnet client before connecting
Holy !@#$, what about TCP/IP!
It's ok to get around a little security in order to get more work done
One great example of this is in a paranoid school or company that firewalls outgoing traffic. They allow telnet, but not SSH. My home machine only accepts SSH. (for obvious reasons) A little tunneling, and presto, I can secure shell to my home computer.
Especially nifty is using http tunnel to establish a secure shell then using the secure shell to tunnel other protocols with encryption
Well, now we know that the whole subliminal message and reverse psychology scene is doomed. The new craze for fooling moderators is posting DeCSS! Soon one will merely need to put a DeCSS link in their .sig to achieve instant +5, Redundant status!
Post DeCSS on MSN! Microsoft usually buys stuff they perceive as a threat, so you might get some money out of it.
I have a base64-encoded-gzipped-tarball with a few of the source files in my user information page.
Well I know DeCSS has been posted more than a few times here on slashdot, and a few users (including me) have it in their user profiles. It's also in an everything2 node. BSI is proud to host DeCSS!
The arg lists could be made safer with some modification to the compiler and libc, but could it be done portably?
This could all too quickly turn into some strange C and C++ hybrid, or a C/Java holy war. Yipes!
How about a function, sort of like printf, in which you specify the number and type of args seperately from their position?
Basically, substituting each % in the input with a format specifier from the hard-coded string. If there are too many "%" in the input, it doesn't crash. ("Error printing error message")Something like this:
Some variation on this could allow the order of the format specifiers to be rearranged. Heck, maybe something more fancy like Already easy to do in perl, but I've never seen this done in C.
The installer definitely surprised me, but it was pretty nifty. The configuration tools seem pretty good for anyone who doesn't want to poke in config files, but I'll probably dual boot with Debian. Only complaint I had was the default Gnome installation. I don't know what the problem was, but the panel sucked up about 60MB of memory. Worked just fine after upgrading to the Helix gnome RPMs though. I was pleasently surprised to find that it came with XFree 4.0 (which I need for my TNT2 card) but it choked on the NVidia drivers. I had to install X from the official distribution to get the NVidia working.
In short, Mandrake's a great distro for ease-of-configuration, but like all distros needs some incantations for the strange hardware
New concept of an alarm clock:
# crontab -e
6 * * * * killall -9 sleep
Uh oh.
Does US citizenship have a click-through license?
This is pretty much what HP's PPA printers (700 series and some others) do. And, they do have a linux driver that in addition to ghostscript can render postscript to them.
But, if it does work as well as they say, what about displaying the song's fingerprint next to each song on an OpenNAP server? This would make it easy to verify that it's the right one before you download.
And on the Everything2 node for DeCSS there's a mime-encoded tarball pasted right in.
As for the Windows keys, I'm using a keyboard that's old enough not to have them. If they really do break as often as you say, maybe you're just mistreating it. Only keyboards I have that don't work are cheap ones I bought just to steal parts from.
And although not very deep, some music is just fun to listen to.
Yeah, I figured it out a few minutes after I posted... :)
Still, many people were confused about the boot time, and "It will be useless until it boots faster" posts modded up didn't help much.
Just trying to clarify things a bit.
The XFree module and the GL libs are binary only, but the source to the small kernel module is distributed. Admittedly, it has problems on many kernels, but at least they let you try to compile it on which ever kernel you happen to be using.
I don't think this applies to nVidia at all though. I have 2 TNT2 cards (go q3a!) and the nVidia drivers are great. yes, they're binary only, and yes they crash sometimes. Hopefully this will be fixed soon, though, and at least they're fast. I'm especially impressed that a hardware company can get their act together and maintain a driver source code base that compiles for multiple OSs.
Quite possibly, the need for a real filesystem might be a good dividing line. Where you need such a thing, linux gives you lots of goodies for providing and using the filesystem. In a system which has no such need, linux looks too much like an overwight dinosaur.
While linux isn't the smallest kernel available, it can be pretty small. The kernel used in ucLinux usually compiles to around 300k. One of the largest aspects of linux on desktops, though, is glibc. uclinux uses a stripped-down libc instead. This means that while it doesn't take much work to port most apps, not much will simply recompile for ucLinux. I have found busybox, the util suite used on the debian rescue disks, to work well in ucLinux.
ucLinux is definitely not designed to run desktop apps on small CPUs. It's more like a happy medium between the ultra-tiny kernels and the compatibility of the standard linux kernel.
There's a .org for the OS and a .com for the hardware.
To fill in some of the gaps in the /. story: ucLinux is a port of the 2.0 linux kernel to MMU-less architectures. It's mainly run on the Motorola 68000 series, but it is also used quite a bit on the (much faster) motorola coldfire chips.
The ucSimm is based on the motorola 68EZ328 Dragonball chip, with onboard memory controller
and LCD controller. It's the same chip used in
most Palm computers.
I haven't used the ucSimm, but I have first-hand uclinux experience with my own hardware. (kiwi.sourceforge.net) It runs pretty nicely, and it definitely doesn't take 30 minutes to boot. Don't know where that came from. Because most of the time-consuming things like loadable modules usually aren't included in the boot process, it only takes seconds to boot. Most of the time is spent decompressing the initial filesystem into DRAM.
ucLinux has been used for many things, including industrial control (microcontroller replacement) and robotics. I have been working, slowly, to make some type of usable PDA based on ucLinux. The Kiwi has been my crude but mostly operational attempt at hardware, and I've been working on a small and efficient GUI system, PicoGUI
(sorry i haven't updated these pages in a while. Most of the fun stuff is taking place in the sourceforge CVS)
Don't forget about Hypercard!
put the value of the object named 'field 1' into the variable 'x' if it is less than 50
vs
if (*field1<50) x = *field1;
And the worst thing about hypercard was that even though the commands were supposed to be more english-like, they still had a firm syntax. It was much more inconsistant than C so i always had to look it up...