I saw this the other day in Geelong (dinky little burg about 80KM down the M1 west of Melbourne).
Cop was sitting in a pale, undercover-cop coloured (car companies in Australia seem to make two colours, an insipid green and an even more insipid metallic blue, which are only used by undercover cops and mobile speed traps) Ford XR6T. Oncoming traffic was flashing its lights at me in warning. I could see a blue arm resting on the window, appearing to cradle something.
As I drove past (at 40KM/h, this was a 40KM/h zone,) I noticed that the cop was doing nothing more than cradle his baton whilst resting his right arm (we drive on the left in.au) on the window ledge.
The aim of speeding fines is not as a revenue source, but to improve public safety, i.e. reducing the number of road deaths.
That's a nice ideal. Unfortunately, in the People's Democratic Republic of Victoria, Chairman Steve Bracks (about -4, 8 on the Political Compass, in my opinion) has decided to do great things such as
Set the tolerance for automatic speed cameras (fixed AND mobile) to 2 or 3KM/h, when the ADR (australian design regulations) which governs roadworthiness of vehicles (and to which your car must be tested in order to be registered in the PDR of Victoria) states that a speedometer must be accurate to within 10%
Pork-barrelled a lovely set of fixed speed cameras on Melbourne's major western orbital road (the aptly named Western Ring Road, state M80) from a company which went out of business not long afterwards, which DID NOT maintain the induction coils under the gantries (yes, Veronica, induction coils) within their engineered specifications; as a result, a lady was charged with doing 170KM/h in the 100 zone in a car which was later proven (in court, by a mechanical engineer) to have trouble exceeding 120KM/h.
On the same road (and others), installing 'variable speed limits' which serve ABSOLUTELY no purpose other than to confuse motorists, slow down traffic (a vast percentage of people were so scared of the fixed cameras that they would habitually do 10KM/h UNDER the limit), and tie into the fixed speed camera network (fortunately, now defunct) in order to fine people who were used to having a fixed 100KM/h limit (note, there are several roads in Melbourne where a variable speed limit is apropos, such as the Westgate Bridge; these roads have variable limits which predate Chairman Bracks by many years)
Instituting a mandatory 50KM/h policy in urban streets 'unless otherwise signed', then conveniently 'forgetting' to install 60KM/h speed signs in some streets (Mackie Road, Bentleigh seems appropriate to cite in this instance; there are others.) until they'd raised a few hundreds of thousdands of dollars.
Instituting, and aggressively policing, mandatory 40KM/h zones outside any building which could conceivably be used for (K12) educational purposes. I have no problem with a 40KM/h zone outside a primary school, grade three kids have little to no road sense; outside a senior school, with youth who are expected to KNOW to STAY OFF THE ROAD is obvious revenue raising.
Please note two things: One, I agree that speeding, in the majority of instances is indefensible. When I was a younger man, I got my share of idiotic speeding fines, which I copped on the chin; I deserved them.
Two, I appreciate that Bracks has to make a buck after Jeff Kennett's reign (Jeff Kennett would be, in my opinion, about 7, 10 on the same Political Compass), given that Kennett sold EVERY public utility in the state, leaving Victorians with poor privatised service and no ongoing revenue generation OTHER than stamp duty (has increased under Bracks' administration,) land tax (ditto,) public transport infringements (ditto,) and speeding fines. However, I find it INDEFENSIBLE that innocent people need to face a court (quite a harrowing experience for John Q. Motorist) in order to defend their ability to drive.
The LKP intercepts various calls, and passes them up to the 'host' kernel.
...
I see a big distinction between a discrete module execution (autonomous and distinct as a NIC driver, and equally as loadable and dischargeable) and grafting Linux source code into their kernel or gcc library source, or even header files, etc into a singularity.;
A kernel module is not 'discrete'. It runs in the same address space as the kernel itself. Think of it in the same way that you would think of a.so or a.dll file. LKP is definitely implemented as either a loadable kernel module or as part of the kernel itself, as it implements a trap for the int0x80 call gate (which linux and others use to jump from user to supervisor mode upon the commencement of a system call.)
What you have repeatedly ignored is the GPL's policy re: linking. From gnu.org:
This General Public License does not permit incorporating your program into proprietary programs. If your program is a subroutine library, you may consider it more useful to permit linking proprietary applications with the library. If this is what you want to do, use the GNU Lesser General Public License instead of this License.
Ergo, linking a proprietary piece of code to a piece of GPL'd code is not permitted (exceptions from the author(s) of the GPL'd code notwithstanding.)
This/is/ a grey area, there's no point pretending that it isn't.
However, when you 'load' an LKM or KLD type entity, you're not simply 'running a program'. You're linking the LKM/KLD with the kernel. Big distinction.
This is the fundamental difference between the GPL and the LGPL. The LGPL permits binary works linking against LGPL libraries, as long as there's a clean way for the recipient of the end work to upgrade to a newer (obviously, ABI compatible) version of the LGPL code. In the bad old days, this meant that programs were released in both linked and unlinked form, and the end recipient had the option of relinking the proprietary code against a new version of the LGPL library(ies) in question. These days, this is done with DSOs, which are sort of a user-space equivalent of LKM/KLDs.
The GPL does not permit this. If you LINK, in ANY WAY, GPLd code to ANY OTHER WORK, that other work then must fall under the auspices of the GPL or a compatible license.
Although, I get the feeling that you understand this, and are merely failing to grasp the fact that loading a kernel module is no different to linking it into the kernel at build time. The symbol table for the post-linked entity contains symbols from both of its progenitors.
Regardless of whether the LKP is a loadable module or not, it is still being linked with the UnixWare kernel at runtime, and as such, the UnixWare kernel must either be licensed in a way which is compatible with the GPL, or the distribution is illegal.
Your confusion stems from the fact that Linus & co granted the OPPOSITE exemption (in part) for the linux kernel; third party kernel modules may be under a GPL-incompatible license, provided that they declare so with a preprocessor directive in one of their source files (MODULE_LICENSE or something.) These non-GPL modules are restricted to linking to limited set of exported symbols, and cannot substantially alter the kernel's behavior; they're really only useful for hardware device drivers (which is about 99% of the market for binary device drivers in any event.)
They didn't cover their posteriors at all. Since they're distributing GPL'd code (which doesn't have an exception similar to the above) which links to non-GPL'd, binary only code, they've lost their rights to distribute the GPL'd code. Pretty simple, really.
Of course, this all assumes that the LKP actually DOES have linux kernel code in it. That's by no means a certainty at this point.
In order for the system to crash during the failure of a disk in a software RAID-x system (at least in Linux), it must be sharing an IDE channel with one of the remaining good disks (in my experience.)
I have several software RAID 5 setups here, two of which use SCSI (and have had multiple disk failures but never actually crashed), and one of which uses IDE with one disk per channel (ditto.) In the case of IDE, the kernel DOES get a bit annoyed with the failed disk, and blocks somewhere within the RAID driver for a few seconds before declaring the disk dead and carrying on as usual.
I don't know how those cheapo 'RAID hardware' cards (which aren't) work, I assume it would depend on the cheapo RAID drivers which they use.
Reminds me of an anecdote I heard a few years back. It's off-the-wall enough to be true, but I don't vouch for its accuracy. It was a pub conversation, after all.
Co-worker at a previous job had an acquaintance who was working for a defense contractor (RLM, i think it was), on some crazy uber-classified Over-the Horizon Radar project. They used an absolute stackload of data in Compaq (ex DEC) SANs, I'm told.
Due to the fact that all this data was classified at some level, and they were a good customer, Compaq gave them an unconditional replacement guarantee on the disks in their RAID arrays. If one failed, Compaq didn't want it back.
So, this friend of a friend started sending in bogus RMA requests and taking the disks home. When this came to light, Compaq, obviously, were rather aggrieved. Since they couldn't do him for theft (the contract being rather ambiguous, and they HAD issued him with the RMAs,) they had the Australian Fed. Police arrest him for Treason.
Um... (quick conversion to metric units, very rough)
3000lbs ~= 1.5 metric tons. 20 miles ~= 32 metric kilometres. 1 US gallon ~= 3.8 litres.
My 1995 Ford Fairlane weighs 1.7 metric tons (tare) plus payload plus fuel. I average 10.5 litres of fuel used per 100KM in mixed highway/city driving.
Some quick maths: 10.5L/100 divided by 100, multiplied by 32 gives me 3.36 litres per 32 kilometres, which is roughly.89 gallons per 20 miles. AND it weighs 400 imperial pounds more.
Which purchase agreement? The one I enter into when I exchange money for tangible goods of value? The same agreement which has NO pre-sale conditions attached other than those imposed by the doctrine of first sale? The hypothetical 'reasonable man' would assume that this was a standard purchase, no different from buying a hammer. When was the last time that the Best Buy or K-Mart clerk told you 'Sir, before you buy that X-Box, you need to understand a few things....'?
It's fundamentally flawed, defeatist, subservient attitudes like yours which will lead us down the path alluded to in this article. Pay attention, or before you know it, your bonnet will be welded shut and your C compiler will require proof of licensing and a 3 day cooling off period.
Not quite. You're thinking of EGCS, which was a project to update GCC 2.7.2 because the EGCS devs didn't like the direction that the FSF were going in with GCC 2.8.
PGCC was a fork of EGCS which was able to emit code optimised for i586-class CPUs. There were versions based on EGCS 1.0.2, 1.0.3 and 1.1. Eventually, the PGCC optimisations got folded into a version of EGCS, and EGCS begat GCC 2.95, which eventually became GCC 3.0.
DNS server A records will only be provided in the 'additional' section of the response (so-called 'glue' records) if the zone in which they live is delegated to themselves.
That is to say, if i query a.gtld-servers.net for www.bob.com IN A, and bob.com is delegated to ns1.bob.com and ns2.bob.com, the registry will know the IP addresses of bob.com's two nameservers and return them in the 'additional' section (otherwise nobody would ever be able to find anything under bob.com, including ns1 and ns2.bob.com).
If bob.com is delegated to ns1.foo.com and ns2.foo.com, but foo.com is delegated somewhere else, then ns1/ns2.foo.com's A records WILL NOT be returned as glue, and your resolver will have to recursively query for those, as well.
But... Goatse is already legitimate art. Free Goatses for all is codified by Amnesty International and the United Nations as a fundamental human right.
IPv6 has an autoconfiguration mechanism whereby an IPv6 autoconfiguration server will spit out a 64-bit prefix (all local networks are/64s in IPv6), and a host will create an EUI-64 address to postpend to it, as a deterministic function of the interface's layer 2 address.
I'd find the RFC but i'm too lazy. Search for 'IPv6 autoconfiguration' on rfc-editor.org or google.
Slashdot Mirror
Who ISN'T a fan of Bob Goatse?!? Heathens....
I saw this the other day in Geelong (dinky little burg about 80KM down the M1 west of Melbourne).
.au) on the window ledge.
Cop was sitting in a pale, undercover-cop coloured (car companies in Australia seem to make two colours, an insipid green and an even more insipid metallic blue, which are only used by undercover cops and mobile speed traps) Ford XR6T. Oncoming traffic was flashing its lights at me in warning. I could see a blue arm resting on the window, appearing to cradle something.
As I drove past (at 40KM/h, this was a 40KM/h zone,) I noticed that the cop was doing nothing more than cradle his baton whilst resting his right arm (we drive on the left in
Sneaky, these coppers.
That's a nice ideal. Unfortunately, in the People's Democratic Republic of Victoria, Chairman Steve Bracks (about -4, 8 on the Political Compass, in my opinion) has decided to do great things such as
Please note two things: One, I agree that speeding, in the majority of instances is indefensible. When I was a younger man, I got my share of idiotic speeding fines, which I copped on the chin; I deserved them.
Two, I appreciate that Bracks has to make a buck after Jeff Kennett's reign (Jeff Kennett would be, in my opinion, about 7, 10 on the same Political Compass), given that Kennett sold EVERY public utility in the state, leaving Victorians with poor privatised service and no ongoing revenue generation OTHER than stamp duty (has increased under Bracks' administration,) land tax (ditto,) public transport infringements (ditto,) and speeding fines. However, I find it INDEFENSIBLE that innocent people need to face a court (quite a harrowing experience for John Q. Motorist) in order to defend their ability to drive.
You are repeatedly ignoring my point.
The LKP intercepts various calls, and passes them up to the 'host' kernel.
...
I see a big distinction between a discrete module execution (autonomous and distinct as a NIC driver, and equally as loadable and dischargeable) and grafting Linux source code into their kernel or gcc library source, or even header files, etc into a singularity.;
A kernel module is not 'discrete'. It runs in the same address space as the kernel itself. Think of it in the same way that you would think of a .so or a .dll file. LKP is definitely implemented as either a loadable kernel module or as part of the kernel itself, as it implements a trap for the int0x80 call gate (which linux and others use to jump from user to supervisor mode upon the commencement of a system call.)
Reference: http://uw714doc.sco.com/en/LX_uw/LKP_system_call_i nterface.html
What you have repeatedly ignored is the GPL's policy re: linking. From gnu.org:
This General Public License does not permit incorporating your program into proprietary programs. If your program is a subroutine library, you may consider it more useful to permit linking proprietary applications with the library. If this is what you want to do, use the GNU Lesser General Public License instead of this License.
Ergo, linking a proprietary piece of code to a piece of GPL'd code is not permitted (exceptions from the author(s) of the GPL'd code notwithstanding.)
This /is/ a grey area, there's no point pretending that it isn't.
However, when you 'load' an LKM or KLD type entity, you're not simply 'running a program'. You're linking the LKM/KLD with the kernel. Big distinction.
This is the fundamental difference between the GPL and the LGPL. The LGPL permits binary works linking against LGPL libraries, as long as there's a clean way for the recipient of the end work to upgrade to a newer (obviously, ABI compatible) version of the LGPL code. In the bad old days, this meant that programs were released in both linked and unlinked form, and the end recipient had the option of relinking the proprietary code against a new version of the LGPL library(ies) in question. These days, this is done with DSOs, which are sort of a user-space equivalent of LKM/KLDs.
The GPL does not permit this. If you LINK, in ANY WAY, GPLd code to ANY OTHER WORK, that other work then must fall under the auspices of the GPL or a compatible license.
Although, I get the feeling that you understand this, and are merely failing to grasp the fact that loading a kernel module is no different to linking it into the kernel at build time. The symbol table for the post-linked entity contains symbols from both of its progenitors.
Not exactly, no.
Regardless of whether the LKP is a loadable module or not, it is still being linked with the UnixWare kernel at runtime, and as such, the UnixWare kernel must either be licensed in a way which is compatible with the GPL, or the distribution is illegal.
Your confusion stems from the fact that Linus & co granted the OPPOSITE exemption (in part) for the linux kernel; third party kernel modules may be under a GPL-incompatible license, provided that they declare so with a preprocessor directive in one of their source files (MODULE_LICENSE or something.) These non-GPL modules are restricted to linking to limited set of exported symbols, and cannot substantially alter the kernel's behavior; they're really only useful for hardware device drivers (which is about 99% of the market for binary device drivers in any event.)
They didn't cover their posteriors at all. Since they're distributing GPL'd code (which doesn't have an exception similar to the above) which links to non-GPL'd, binary only code, they've lost their rights to distribute the GPL'd code. Pretty simple, really.
Of course, this all assumes that the LKP actually DOES have linux kernel code in it. That's by no means a certainty at this point.
In order for the system to crash during the failure of a disk in a software RAID-x system (at least in Linux), it must be sharing an IDE channel with one of the remaining good disks (in my experience.)
I have several software RAID 5 setups here, two of which use SCSI (and have had multiple disk failures but never actually crashed), and one of which uses IDE with one disk per channel (ditto.) In the case of IDE, the kernel DOES get a bit annoyed with the failed disk, and blocks somewhere within the RAID driver for a few seconds before declaring the disk dead and carrying on as usual.
I don't know how those cheapo 'RAID hardware' cards (which aren't) work, I assume it would depend on the cheapo RAID drivers which they use.
Reminds me of an anecdote I heard a few years back. It's off-the-wall enough to be true, but I don't vouch for its accuracy. It was a pub conversation, after all.
Co-worker at a previous job had an acquaintance who was working for a defense contractor (RLM, i think it was), on some crazy uber-classified Over-the Horizon Radar project. They used an absolute stackload of data in Compaq (ex DEC) SANs, I'm told.
Due to the fact that all this data was classified at some level, and they were a good customer, Compaq gave them an unconditional replacement guarantee on the disks in their RAID arrays. If one failed, Compaq didn't want it back.
So, this friend of a friend started sending in bogus RMA requests and taking the disks home. When this came to light, Compaq, obviously, were rather aggrieved. Since they couldn't do him for theft (the contract being rather ambiguous, and they HAD issued him with the RMAs,) they had the Australian Fed. Police arrest him for Treason.
He got 5 to 10 years.
Um...
.89 gallons per 20 miles. AND it weighs 400 imperial pounds more.
(quick conversion to metric units, very rough)
3000lbs ~= 1.5 metric tons.
20 miles ~= 32 metric kilometres.
1 US gallon ~= 3.8 litres.
My 1995 Ford Fairlane weighs 1.7 metric tons (tare) plus payload plus fuel. I average 10.5 litres of fuel used per 100KM in mixed highway/city driving.
Some quick maths: 10.5L/100 divided by 100, multiplied by 32 gives me 3.36 litres per 32 kilometres, which is roughly
Which purchase agreement? The one I enter into when I exchange money for tangible goods of value? The same agreement which has NO pre-sale conditions attached other than those imposed by the doctrine of first sale? The hypothetical 'reasonable man' would assume that this was a standard purchase, no different from buying a hammer. When was the last time that the Best Buy or K-Mart clerk told you 'Sir, before you buy that X-Box, you need to understand a few things....'?
It's fundamentally flawed, defeatist, subservient attitudes like yours which will lead us down the path alluded to in this article. Pay attention, or before you know it, your bonnet will be welded shut and your C compiler will require proof of licensing and a 3 day cooling off period.
now Sun's announces it will revile part of its code
Solaris has been reviled for some time.
Jesus Christ! Couldn't you have just linked directly to the goatse.cx image? At least we're somewhat prepared for that...
Here you go
I generally replace the battery in my car (650 cold crank amps or thereabouts) every 3 years for the princely sum of ~$A200.
It takes me five minutes. From what I've heard, you can't even get the back plate off an iPod in five minutes.
Dumb analogy.
I will lash out dancing like a madman when you're gone! I will spit the blue flame and hurl my glass against the wall!
Mr Bitterness is arguably more apropos a soul coughing reference for a story about SCO.
'factoring prime numbers'?
You mean deriving the factors of products of primes, right?
Congratulations.
You just described kuro5hin.
Carl hasn't left AOL... He's no longer their postmaster god, but he's moved sideways into a role known as 'Director, Host Mail Development.'
I'd assume he's still doing good things at AOL as far as anti-spam goes, given his new title.
Not quite. You're thinking of EGCS, which was a project to update GCC 2.7.2 because the EGCS devs didn't like the direction that the FSF were going in with GCC 2.8.
PGCC was a fork of EGCS which was able to emit code optimised for i586-class CPUs. There were versions based on EGCS 1.0.2, 1.0.3 and 1.1. Eventually, the PGCC optimisations got folded into a version of EGCS, and EGCS begat GCC 2.95, which eventually became GCC 3.0.
DNS server A records will only be provided in the 'additional' section of the response (so-called 'glue' records) if the zone in which they live is delegated to themselves.
That is to say, if i query a.gtld-servers.net for www.bob.com IN A, and bob.com is delegated to ns1.bob.com and ns2.bob.com, the registry will know the IP addresses of bob.com's two nameservers and return them in the 'additional' section (otherwise nobody would ever be able to find anything under bob.com, including ns1 and ns2.bob.com).
If bob.com is delegated to ns1.foo.com and ns2.foo.com, but foo.com is delegated somewhere else, then ns1/ns2.foo.com's A records WILL NOT be returned as glue, and your resolver will have to recursively query for those, as well.
It's amazing that DNS works at all.
Your dad's an idiot.
But... Goatse is already legitimate art.
Free Goatses for all is codified by Amnesty International and the United Nations as a fundamental human right.
No state 'income' tax. I bet they tax property. I know some of those states have the highest sales tax, gas tax, beef tax, etc...
Do any states in America have a goatse tax?
I stand corrected. Thankyou.
Nope. Read the IPv6 specs.
/64s in IPv6), and a host will create an EUI-64 address to postpend to it, as a deterministic function of the interface's layer 2 address.
IPv6 has an autoconfiguration mechanism whereby an IPv6 autoconfiguration server will spit out a 64-bit prefix (all local networks are
I'd find the RFC but i'm too lazy. Search for 'IPv6 autoconfiguration' on rfc-editor.org or google.
Have a nice day.
MTA Authorization Records In DNS. It was an IETF working group.
Google for it.