Well, this one is not only "64-bit" only, it is specifically amd64 only. No arm64, no siree. And I'm typing these words on an armhf laptop.
Not that anyone sane would run a closed blob of known spyware anywhere near their client machine, which either has important private keys on it, has ssh keys (most likely currently unlocked) to log to machines with such private keys on, or has a key stick put into it from time to time.
Intel's ME uses ARC not ARM, and its requirements change per-submodel. You also need it to be signed by Intel's key, also specific to that particular submodel. It is heavily encrypted as well, with some very serious precautions against someone reviewing that code.
No idea about AMD, but I expect the NSLs they received to specifically ban open sourcing.
Thus, ironically, it's cheap Chinese makers who are trustworthy here, as adding a separate core just to undetectably backdoor you would cut too much into their razor-thin margins.
Please tell me when it meant 1000 bytes instead of 1024. It's been in use for several decades before a marketing department of a drive maker decided otherwise.
Both Intel and AMD look like a lost cause. Intel ME is certain to contains a network-accessible backdoor, AMD's version is slightly less vicious but not good either.
There's hope in the ARM world: there's the TrustZone but some computers with non-locked bootloaders allow you to load your own code there. You obviously don't want to write such code yourself thus you probably want to use ATF but it's open for modification and, more importantly here, review.
For example Pinebook (an incoming $89 laptop) allows you to do that, and if you don't need accelerated Mali400 proprietary drivers, you can use free software. Well, [near-]mainline support is not yet mature: simplefb LCD display is like four days old and the DRM driver is not yet working, but as shipment dates have been delayed again (currently for March 20ish), the kernel+uboot should be usable by then.
Obviously, a computer where you not only can but need to tinker with to get working is not for everyone, but it is up to people like us to make it so. It is trustable, which is what you and me are looking for. And once Icenowy and the rest make the kernel usable, we can take that kernel, give it better userland than "dd this image of dubious provenience" and release to regular users.
Documenting an incorrect usage doesn't mean it is valid to use that.
It'd be like if, after several decades of use, some committee sponsored by drive makers rightly sued for false advertising, declared that 1KB instead of 1024 bytes suddenly means 1000.
A word with two distinct meanings in the same domain is worthless.
update issues like the one you describe are happening to less 1 in 10,000 systems
In my testing, that's 2 in 2. On both of these, not every update fails, just somewhere around 2 in 3, but that's per version -- if a given update fails, it fails every time.
And from a glance at Slashdot, I'm not alone.
I pity those who actually have to do something important on Windows...
My layout for N900 is pretty convenient for typing regexps and coding in general, especially compared to Nokia's anti-genious default where to type most symbols you need to press a key combination to pull up an on-screen menu, shift your fingers, navigate the menu (as terrible as the typical Android/iPhone on-screen "keyboard") and select something. All while there's a crapload of unbound shift/Fn combinations.
A physical keyboard is so massively superior to an on-screen one that I wonder why anyone would use the latter for anything but sparse status updates on this week's MySpace's remake (I don't know what's popular nowadays, now that Fecesbook is passe and only orange clowns are Twits?). But it's status updates on social media what's where advertising money comes from...
Just go virtual. Linux and FreeBSD will support pretty much everything pretty much forever. Run Win7 in VirtualBox forever.
This is a trivial and good solution for us Slashdot crowd, but it exceeds the technical skill of a "normal" person.
Another downside is for those who play AAA (ie, containing DRM malware) games which are notoriously unportable and run like shit (if even) under Wine.
Virus or malware? Revert to latest known-good snapshot. Reboot in 5-10 seconds and you are back up and running.
Hell yeah. Compartmentalizing stuff is another reason: if you run dubious windows program A in a VM on its own, it has no chances to break program B in another VM.
Kaby Lake (and newer) processors are only officially supported on Windows 10.
This is actually only FUD and an excuse to force certain business users, win7 does work fine (well, fine for Windows). It's not the case for new network/graphics/etc cards, though.
-rwsr-xr-x 1 root root 18768 Feb 19 21:17/usr/lib/chromium/chrome-sandbox
Bit wrong on this one. The binary is "owned" by root.
And thus, via the setuid bit, a process that execs this file gets full root privileges.
You as a user only have read/execute rights the second set of permissions "-xr" and then "everyone" "-x".
Eh? What "-xr", what "-x"? The permissions are: group "r-x" which doesn't matter as you don't belong to group root and the file isn't setgid, and others "r-x" so you can execute it.
This doesn't mean the file runs as root.
That's exactly what setuid means. The process can then shed its privileges, but a browser shouldn't need them in the first place.
There's WSL which can't be locked down if it's expected to do its job, and can run wine. Just have people migrate to that. Once we're there, shed the pointless outer layer and you don't need to worry about Microsoft lockdown anymore.
As bad as their random-read performance is, their random-WRITE performance is usually much, MUCH worse.
You get a massive speed increase if you switch to a better filesystem: btrfs or f2fs.
git reset --hard: 3m45s btrfs, 3m55s f2fs, 12m30s ext4, 16-18m xfs (huge variance) "./configure && make -j4 && make test" of a shit package with only ~2MB of persistent writes: f2fs 95s, btrfs 97s, xfs 120s, ext4 122s
(class-4 card in a Pine64)
And almost none do their own wear-leveling, so a Linux swapfile can literally max out the lifetime writes of a microSD card in 1-3 months
Not SD but eMMC: Samsung's fancy-schmancy eMMC cards are apparently made by someone no one told about the write endurance problem: I've been running Debian archive rebuilds and other I/O-heavy loads since early 2013 on a 64GB card ($89) in an Odroid-U2, and despite me heavily overcommitting memory (ie, heavy swapping a lot of the time), the card is still going strong.
Make all the rationalizations that you want, SCOTUS has already decided. This is not a matter of opinion or rightness, but fact. Non-citizens do not have constitutional rights. They do have human rights and any rights granted by treaty or specific laws, but constitutional rights are only guaranteed for citizens.
Fact: the consitution, in literal and clear words, say A. Fact: a branch of the government, because it's more convenient for them, says B.
I demand rights I have, not rights a government wants me to have. That I cannot exercise those rights at present is why I'm complaining. And I'm among people who can, if we got off our asses, fight back: while we can't fight the way the congresscritters prefer, by the Golden Rule, as we don't have the gold, we can research ways and educate people how to get your data unmolested in face of unlawful searches and forced password exposure.
Americans should not be subject to unreasonable searches and seizures.
Note the wording: The right of the people to be secure in their persons, houses, papers, and effects, [...]. It doesn't say "Americans" anywhere. So while I can't run for US president, if I visit, I am supposed to have thugs keep the [expletive] out of my "papers and effects". Which does include my phone.
Slashdot Mirror
Unless your output goes to a text-to-speech reader, elinks is worlds better. Lynx doesn't even have tabs.
But indeed, both win over graphical bloatware.
Well, this one is not only "64-bit" only, it is specifically amd64 only. No arm64, no siree. And I'm typing these words on an armhf laptop.
Not that anyone sane would run a closed blob of known spyware anywhere near their client machine, which either has important private keys on it, has ssh keys (most likely currently unlocked) to log to machines with such private keys on, or has a key stick put into it from time to time.
But, but, but those Romans used lead pipes! It surely must be better to die of thirst than to swallow homeopathical amounts of lead or radiation!
Or, simpler, move out of California, as nowhere else cellphones produce such deadly amounts of radiation.
Intel's ME uses ARC not ARM, and its requirements change per-submodel. You also need it to be signed by Intel's key, also specific to that particular submodel. It is heavily encrypted as well, with some very serious precautions against someone reviewing that code.
No idea about AMD, but I expect the NSLs they received to specifically ban open sourcing.
Thus, ironically, it's cheap Chinese makers who are trustworthy here, as adding a separate core just to undetectably backdoor you would cut too much into their razor-thin margins.
Please tell me when it meant 1000 bytes instead of 1024. It's been in use for several decades before a marketing department of a drive maker decided otherwise.
Both Intel and AMD look like a lost cause. Intel ME is certain to contains a network-accessible backdoor, AMD's version is slightly less vicious but not good either.
There's hope in the ARM world: there's the TrustZone but some computers with non-locked bootloaders allow you to load your own code there. You obviously don't want to write such code yourself thus you probably want to use ATF but it's open for modification and, more importantly here, review.
For example Pinebook (an incoming $89 laptop) allows you to do that, and if you don't need accelerated Mali400 proprietary drivers, you can use free software. Well, [near-]mainline support is not yet mature: simplefb LCD display is like four days old and the DRM driver is not yet working, but as shipment dates have been delayed again (currently for March 20ish), the kernel+uboot should be usable by then.
Obviously, a computer where you not only can but need to tinker with to get working is not for everyone, but it is up to people like us to make it so. It is trustable, which is what you and me are looking for. And once Icenowy and the rest make the kernel usable, we can take that kernel, give it better userland than "dd this image of dubious provenience" and release to regular users.
Documenting an incorrect usage doesn't mean it is valid to use that.
It'd be like if, after several decades of use, some committee sponsored by drive makers rightly sued for false advertising, declared that 1KB instead of 1024 bytes suddenly means 1000.
A word with two distinct meanings in the same domain is worthless.
There's a simple solution for this plague on the mankind, proposed in Henry VI, Part 2, Act IV, Scene 2.
This sounds more like Flayed Ones than Terminator.
Try natural water, rather than sanitized chlorinated pools. Fish SHIT in it!
It's not Windows, you can uninstall that xapian crap, it's about as useful as nepomuk.
update issues like the one you describe are happening to less 1 in 10,000 systems
In my testing, that's 2 in 2. On both of these, not every update fails, just somewhere around 2 in 3, but that's per version -- if a given update fails, it fails every time.
And from a glance at Slashdot, I'm not alone.
I pity those who actually have to do something important on Windows...
My layout for N900 is pretty convenient for typing regexps and coding in general, especially compared to Nokia's anti-genious default where to type most symbols you need to press a key combination to pull up an on-screen menu, shift your fingers, navigate the menu (as terrible as the typical Android/iPhone on-screen "keyboard") and select something. All while there's a crapload of unbound shift/Fn combinations.
A physical keyboard is so massively superior to an on-screen one that I wonder why anyone would use the latter for anything but sparse status updates on this week's MySpace's remake (I don't know what's popular nowadays, now that Fecesbook is passe and only orange clowns are Twits?). But it's status updates on social media what's where advertising money comes from...
Just go virtual. Linux and FreeBSD will support pretty much everything pretty much forever. Run Win7 in VirtualBox forever.
This is a trivial and good solution for us Slashdot crowd, but it exceeds the technical skill of a "normal" person.
Another downside is for those who play AAA (ie, containing DRM malware) games which are notoriously unportable and run like shit (if even) under Wine.
Virus or malware? Revert to latest known-good snapshot. Reboot in 5-10 seconds and you are back up and running.
Hell yeah. Compartmentalizing stuff is another reason: if you run dubious windows program A in a VM on its own, it has no chances to break program B in another VM.
Kaby Lake (and newer) processors are only officially supported on Windows 10.
This is actually only FUD and an excuse to force certain business users, win7 does work fine (well, fine for Windows). It's not the case for new network/graphics/etc cards, though.
-rwsr-xr-x 1 root root 18768 Feb 19 21:17 /usr/lib/chromium/chrome-sandbox
Bit wrong on this one. The binary is "owned" by root.
And thus, via the setuid bit, a process that execs this file gets full root privileges.
You as a user only have read/execute rights the second set of permissions "-xr" and then "everyone" "-x".
Eh? What "-xr", what "-x"? The permissions are: group "r-x" which doesn't matter as you don't belong to group root and the file isn't setgid, and others "r-x" so you can execute it.
This doesn't mean the file runs as root.
That's exactly what setuid means. The process can then shed its privileges, but a browser shouldn't need them in the first place.
Win7 will be killed the same way they got people to downgrade from XP to 7 -- no drivers for new hardware.
Well, Microsoft is a master of bribing^Wlobbying. See Munich.
There's WSL which can't be locked down if it's expected to do its job, and can run wine. Just have people migrate to that. Once we're there, shed the pointless outer layer and you don't need to worry about Microsoft lockdown anymore.
Hell yeah. Especially browsers have never, ever a reason to run as root. /usr/lib/chromium/chrome-sandbox
-rwsr-xr-x 1 root root 18768 Feb 19 21:17
As bad as their random-read performance is, their random-WRITE performance is usually much, MUCH worse.
You get a massive speed increase if you switch to a better filesystem: btrfs or f2fs.
git reset --hard: 3m45s btrfs, 3m55s f2fs, 12m30s ext4, 16-18m xfs (huge variance)
"./configure && make -j4 && make test" of a shit package with only ~2MB of persistent writes: f2fs 95s, btrfs 97s, xfs 120s, ext4 122s
(class-4 card in a Pine64)
And almost none do their own wear-leveling, so a Linux swapfile can literally max out the lifetime writes of a microSD card in 1-3 months
Not SD but eMMC: Samsung's fancy-schmancy eMMC cards are apparently made by someone no one told about the write endurance problem: I've been running Debian archive rebuilds and other I/O-heavy loads since early 2013 on a 64GB card ($89) in an Odroid-U2, and despite me heavily overcommitting memory (ie, heavy swapping a lot of the time), the card is still going strong.
By "native American", do you mean the first, second or third wave of pre-Eric-the-Red immigration?
Make all the rationalizations that you want, SCOTUS has already decided. This is not a matter of opinion or rightness, but fact. Non-citizens do not have constitutional rights. They do have human rights and any rights granted by treaty or specific laws, but constitutional rights are only guaranteed for citizens.
Fact: the consitution, in literal and clear words, say A.
Fact: a branch of the government, because it's more convenient for them, says B.
I demand rights I have, not rights a government wants me to have. That I cannot exercise those rights at present is why I'm complaining. And I'm among people who can, if we got off our asses, fight back: while we can't fight the way the congresscritters prefer, by the Golden Rule, as we don't have the gold, we can research ways and educate people how to get your data unmolested in face of unlawful searches and forced password exposure.
Americans should not be subject to unreasonable searches and seizures.
Note the wording: The right of the people to be secure in their persons, houses, papers, and effects, [...]. It doesn't say "Americans" anywhere. So while I can't run for US president, if I visit, I am supposed to have thugs keep the [expletive] out of my "papers and effects". Which does include my phone.