There was an incident a few years ago (that led to at least one subsystem maintainer resigning) where RdRand was used as the EXCLUSIVE entropy source for some items if it were present. http://cryptome.org/2013/07/intel-bed-nsa.htm - Matt Mackall resigned over it.
This is BAD.
If it is now merely feeding the pool as one of multiple sources, then it's OK. If anything is directly exposed to raw rdrand output, something is very wrong.
Yup. The Sony Smartwatch is actually pretty dumb - it basically acts as a secondary touchscreen and vibration notifier for your phone.
As a result it requires a very low-power CPU (I think it may be an AVR...) and gets decent battery life. (Not as good as the greyscale-LCD-only smartwatches like MetaWatch though.)
It's also cheaper, I don't know where the SW2 is going to wind up, but the original SmartWatch runs for around $100.
This is not possible, unless you just reflashed CM over an existing install that had gapps - CM has a built in backup system that lets addon packages create a backup script which will be used by the CM installer. The most well known user of this feature is gapps - so it gets backed up and reinstalled any time you flash a CM update.
Right now, as much as I dislike Samsung, I have to give them credit with respect to bootloader locking. Verizon devices were the ONLY bootloader-locked Samsungs until the Galaxy S4, when AT&T was added to the list. Note that pretty much all HTCs on these carriers have been locked down too. (The One on AT&T is strange, I think HTC intentionally "made a mistake" by whitelisting these devices.)
"which is why the Galaxy S is on version 4" - bullshit. The original Aries family deadended at Gingerbread, despite the fact that they were nearly identical spec-wise to the Nexus S which made it as far as 4.0 or 4.1 (I can't remember if crespo got 4.1 or if it EOLed when 4.2 was released...)
This is despite the fact that the community had 4.0 running within a few weeks, has released 4.1 and 4.2 for that device, and even 4.3 seems to be coming along OK so far.
To be clear, I meant that it's only viable on a major carrier with T-Mo.
MVNOs are a whole other story, although the primaries are now trying to crack down.
For example, you can no longer get Straight Talk SIMs on AT&T. Net10 SIMs are limited to 1.5GB/month if on AT&T, and it's EXTREMELY difficult to get one on AT&T. If T-Mobile has ANY coverage in the area code you input (even if only EDGE), you only are offered T-Mobile SIMs.
Red Pocket is more expensive ($60 for 3GB/mo) but at least it's not difficult to get AT&T SIMs with them.
I have a Sony Smartwatch, and it is specifically designed to be a dumb second screen for your phone.
Its battery life is not as good as the Pebble or Metawatch due to the full color screen. The MW seems to be a good compromise - reasonably standalone, but still primarily a display for your phone. (The Sony SW is a little "too" dumb when disconnected.)
Yup. It's been possible to decompile the code for ages.
Reinjecting it easily and without high risk of detection has been difficult since around the end of the first year of the game.
And yes: Back in the early days, I was starting to get bored with the game, and wrote what I believe was one of the first hacked autopilots. I didn't distribute it though so it wasn't well known. This was before warp-to-zero, so SuperPilot would actually go through all bookmarks in the system to attempt to find one that would allow it to drop within jump range of a target gate.
Reprocessing, low in volume and concentrated so relatively easy to manage.
As compared to gas drilling, which is resulting in widespread distributed environmental damage. It just doesn't attract as much attention because a single gas incident is just a drop in the bucket - the problem is that regulation and enforcement are so lax in the gas drilling industry that all the drops in the bucket amount to a downpour.
Except when your SDK is in beta and explicitly states that apps are not to be distributed without your approval until the SDK comes out of beta.
In this case, I'm not sure if it's even an SDK change - it's a removal of an exploit that was used to allow a non-whitelisted unofficial app to behave like a whitelisted one.
Thing is, one needs to keep in mind that AirCast was abusing a loophole in Google's whitelisting process (which is in place while the SDK is in beta) - It was behaving as if it were an instance of the Chromecast extension for Chrome.
One thing Koush never indicated was whether this change affected SDK-based apps like his original (unreleased due to the current licensing of the beta SDK) local content playback app.
Hmm... I wonder what the HSIC interface on Samsung Exynos4 is.
HSIC is a stripped-down form of USB that's frequently used as the interface for onboard modems on mobile phones. Samsung Exynos4 devices have been plagued by frequent HSIC disconnects for nearly a year.
At the software level, HSIC uses one of the existing HCI interfaces - the main differences between HSIC and USB are physical layer things, not software.
S3 Mini is NovaThor based, and I am not sure if a single NovaThor device is receiving nightlies.
Some NovaThor-based Sonys got unofficial weekly builds from FreeXperia, but the platform was a nightmare to work with and I think jerpelea gave up on NovaThor after either CM9 or CM10.0
And yes - there are platforms even worse for a CM maintainer than Exynos4. NovaThor and MTK65xx being two of them.
I'm not sure if they do yet, but one of the team members has been working on it and I think found the culprit. They're just trying to determine the ramifications of reverting the change that broke things right now. (They were about to leave on vacation, and I think may have band-aided it before leaving though...)
Play Store is part of gapps, which has always been a separate install because it's Google's stuff and not open source. It's a Play Store problem, not a base OS problem (which is why it started happening suddenly on devices that did not receive any form of OS update).
Also, some tasks are simply not parallelizable. Image processing is one of the exceptions, but for most image processing operations out there (such as JPEG compression and HDR compositing), there is dedicated hardware on most ARM devices.
Hotplugging overhead on most current ARM chips is murder, and also, many of the nicest power management features of ARM chips don't work well when more than one core is active (I think it's related somewhat to all of Linus' rants about braindead cache architectures). Interestingly, Qualcomm is the only SoC manufacturer that has any decent mitigation for this limitation (asynchronous clocking of each core, which mostly makes up for the fact that cpuidle goes to shit when more than one core is lit.)
Most of the tasks in this category that you might see running on a phone are executing on dedicated hardware within the SoC.
Qualcomm's Hexagon DSPs are pretty neat, and a typical Qcom chip has a few - they just never market them as extra cores, but they ARE there.
Remember, Qualcomm's core market are phones and tablets, and that is the context in which their comments regarding MTK's octa-A9 should be taken.
Also keep in mind that Qualcomm is coming off of a nice rosy year where their dual Krait SoCs were routinely smoking quad-A9s on typical smartphone/tablet workloads.
Another thing to keep in mind is that, so far, nearly all multicore ARM systems suffer in terms of power management when more than one CPU is lit up - in many cases, many of the deeper idle modes become unavailable if more than one core is active. This is even true on Qualcomm's chips, but at least they can clock each core asynchronously. All of MTK's chips so far are synchronously clocked, which means that if additional cores are lit up, they run at the same speed as the others, often with crippled cpuidle.
Part of that is that unlike his predecessor (Stringer) - Hirai realizes that treating your customers like shit is bad for business.
Sony under Hirai is very different from Sony under Stringer - this is most obvious if you look at Sony Mobile, who are one of the largest contributors to the Android Open Source Project (AOSP) and the only manufacturer that actively maintains AOSP builds for some of their devices. There are numerous signs that, rather than squash the anomalous behavior of the former Sony-Ericsson, the rest of Sony is starting to adopt Mobile's ways.
The thing is, the Real Money -> PLEX -> ISK thing is NOT the same as Real Money ISK
That intermediary item there changes things significantly, such that when you lose a big ship in EVE, unless you obtained it using PLEX, you didn't actually lose anything that was worth any real world money. Lots of media outlets made a big deal over "$9000 ship lost" when that Revenant was downed. The thing is - that guy didn't lose $9000. At worst he lost an opportunity to purchase $9000 worth of game subscription time at the worst pricing tier possible. (PLEX are around $20 for one month of time, a one month subscription is $15-16 or so, and if you prepay for multiple months it gets down to as low as $10-11/month.)
That's the real key here - in all of these transactions, only CCP can receive real world money. Yes, there are illicit real money trades that sometimes occur, but the exchange rate for these in ISK per dollar is shit (from the ISK seller's perspective) since it's inherently risky compared to PLEX. Someone selling ISK directly has to offer MUCH more ISK per dollar due to the massively increased risk compared to what a customer can do by buying a PLEX from CCP and selling it for ISK.
Yeah. My one encounter with a Titan (which occurred 2-3 weeks before quitting the game for 5 years) was: Fleet commander initiates a fleetwarp to the gate I go to warp Client hangs as grid loads Hey look I'm at my clone station
FYI, CCP did not slow anything down. TiDi kicks in automatically when server load goes up.
However, in the event of an anticipated big fight, CCP will move a system to a "reinforced node" (e.g. an extra beefy server). Of interest, though, is that many fights this year have not been entirely anticipated, resulting in TiDi being the only mechanism for load handling in play. For example, Asakai was the result of an unplanned misclick - it was supposed to be a small skirmish, but turned into a massive battle when a titan pilot jumped into the middle of the enemy fleet instead of bridging his fleetmates in. TiDi + unreinforced node = interesting mechanics, in which the fight is slowed down to 1/10 time, but most of the rest of the universe is running at normal speed. In the past, fleet battles were usually over in minutes, often with half of one fleet dying before they could even load the grid. Now they can last hours, long enough for reinforcements to arrive and massive escalation.
I've played EVE off and on (more off than on) since launch... I just came back from a five year break, I think I'm going to wind up quitting again in not too long. That said, reading the news of what's going on around the edges of the galaxy is pretty neat. I just don't have anywhere near enough time to participate in shit like that, and the game can easily get very boring if you're not in on the nullsec action.
Slashdot Mirror
Wrong. That's actually a good motivation to cover up fuckups.
Like reporting to your superiors that your reactor is intact and there is no problem even though there are chunks of burning graphite moderator on the ground all around the plant - http://en.wikipedia.org/wiki/Individual_involvement_in_the_Chernobyl_disaster#Dyatlov
There was an incident a few years ago (that led to at least one subsystem maintainer resigning) where RdRand was used as the EXCLUSIVE entropy source for some items if it were present. http://cryptome.org/2013/07/intel-bed-nsa.htm - Matt Mackall resigned over it.
This is BAD.
If it is now merely feeding the pool as one of multiple sources, then it's OK. If anything is directly exposed to raw rdrand output, something is very wrong.
Yup. The Sony Smartwatch is actually pretty dumb - it basically acts as a secondary touchscreen and vibration notifier for your phone.
As a result it requires a very low-power CPU (I think it may be an AVR...) and gets decent battery life. (Not as good as the greyscale-LCD-only smartwatches like MetaWatch though.)
It's also cheaper, I don't know where the SW2 is going to wind up, but the original SmartWatch runs for around $100.
This is not possible, unless you just reflashed CM over an existing install that had gapps - CM has a built in backup system that lets addon packages create a backup script which will be used by the CM installer. The most well known user of this feature is gapps - so it gets backed up and reinstalled any time you flash a CM update.
Right now, as much as I dislike Samsung, I have to give them credit with respect to bootloader locking. Verizon devices were the ONLY bootloader-locked Samsungs until the Galaxy S4, when AT&T was added to the list. Note that pretty much all HTCs on these carriers have been locked down too. (The One on AT&T is strange, I think HTC intentionally "made a mistake" by whitelisting these devices.)
"which is why the Galaxy S is on version 4" - bullshit. The original Aries family deadended at Gingerbread, despite the fact that they were nearly identical spec-wise to the Nexus S which made it as far as 4.0 or 4.1 (I can't remember if crespo got 4.1 or if it EOLed when 4.2 was released...)
This is despite the fact that the community had 4.0 running within a few weeks, has released 4.1 and 4.2 for that device, and even 4.3 seems to be coming along OK so far.
To be clear, I meant that it's only viable on a major carrier with T-Mo.
MVNOs are a whole other story, although the primaries are now trying to crack down.
For example, you can no longer get Straight Talk SIMs on AT&T. Net10 SIMs are limited to 1.5GB/month if on AT&T, and it's EXTREMELY difficult to get one on AT&T. If T-Mobile has ANY coverage in the area code you input (even if only EDGE), you only are offered T-Mobile SIMs.
Red Pocket is more expensive ($60 for 3GB/mo) but at least it's not difficult to get AT&T SIMs with them.
This is only viable on T-Mobile.
With AT&T, Sprint, and Verizon, you pay that fee whether or not you paid for your phone outright.
With the Nexus 4 at $199 now, paying outright isn't exactly a bad thing.
That's why I'm avoiding the Omate TrueSmart.
I have a Sony Smartwatch, and it is specifically designed to be a dumb second screen for your phone.
Its battery life is not as good as the Pebble or Metawatch due to the full color screen. The MW seems to be a good compromise - reasonably standalone, but still primarily a display for your phone. (The Sony SW is a little "too" dumb when disconnected.)
Yup. It's been possible to decompile the code for ages.
Reinjecting it easily and without high risk of detection has been difficult since around the end of the first year of the game.
And yes: Back in the early days, I was starting to get bored with the game, and wrote what I believe was one of the first hacked autopilots. I didn't distribute it though so it wasn't well known. This was before warp-to-zero, so SuperPilot would actually go through all bookmarks in the system to attempt to find one that would allow it to drop within jump range of a target gate.
Once WTZ went in I didn't bother.
Reprocessing, low in volume and concentrated so relatively easy to manage.
As compared to gas drilling, which is resulting in widespread distributed environmental damage. It just doesn't attract as much attention because a single gas incident is just a drop in the bucket - the problem is that regulation and enforcement are so lax in the gas drilling industry that all the drops in the bucket amount to a downpour.
Except when your SDK is in beta and explicitly states that apps are not to be distributed without your approval until the SDK comes out of beta.
In this case, I'm not sure if it's even an SDK change - it's a removal of an exploit that was used to allow a non-whitelisted unofficial app to behave like a whitelisted one.
Thing is, one needs to keep in mind that AirCast was abusing a loophole in Google's whitelisting process (which is in place while the SDK is in beta) - It was behaving as if it were an instance of the Chromecast extension for Chrome.
One thing Koush never indicated was whether this change affected SDK-based apps like his original (unreleased due to the current licensing of the beta SDK) local content playback app.
Hmm... I wonder what the HSIC interface on Samsung Exynos4 is.
HSIC is a stripped-down form of USB that's frequently used as the interface for onboard modems on mobile phones. Samsung Exynos4 devices have been plagued by frequent HSIC disconnects for nearly a year.
At the software level, HSIC uses one of the existing HCI interfaces - the main differences between HSIC and USB are physical layer things, not software.
S3 Mini is NovaThor based, and I am not sure if a single NovaThor device is receiving nightlies.
Some NovaThor-based Sonys got unofficial weekly builds from FreeXperia, but the platform was a nightmare to work with and I think jerpelea gave up on NovaThor after either CM9 or CM10.0
And yes - there are platforms even worse for a CM maintainer than Exynos4. NovaThor and MTK65xx being two of them.
Except then, they just won't have a single SoC vendor willing to deal with them.
I'm not sure if they do yet, but one of the team members has been working on it and I think found the culprit. They're just trying to determine the ramifications of reverting the change that broke things right now. (They were about to leave on vacation, and I think may have band-aided it before leaving though...)
How can we fix a bug in shit that isn't ours?
Play Store is part of gapps, which has always been a separate install because it's Google's stuff and not open source. It's a Play Store problem, not a base OS problem (which is why it started happening suddenly on devices that did not receive any form of OS update).
Also, some tasks are simply not parallelizable. Image processing is one of the exceptions, but for most image processing operations out there (such as JPEG compression and HDR compositing), there is dedicated hardware on most ARM devices.
Hotplugging overhead on most current ARM chips is murder, and also, many of the nicest power management features of ARM chips don't work well when more than one core is active (I think it's related somewhat to all of Linus' rants about braindead cache architectures). Interestingly, Qualcomm is the only SoC manufacturer that has any decent mitigation for this limitation (asynchronous clocking of each core, which mostly makes up for the fact that cpuidle goes to shit when more than one core is lit.)
Most of the tasks in this category that you might see running on a phone are executing on dedicated hardware within the SoC.
Qualcomm's Hexagon DSPs are pretty neat, and a typical Qcom chip has a few - they just never market them as extra cores, but they ARE there.
Remember, Qualcomm's core market are phones and tablets, and that is the context in which their comments regarding MTK's octa-A9 should be taken.
Also keep in mind that Qualcomm is coming off of a nice rosy year where their dual Krait SoCs were routinely smoking quad-A9s on typical smartphone/tablet workloads.
Another thing to keep in mind is that, so far, nearly all multicore ARM systems suffer in terms of power management when more than one CPU is lit up - in many cases, many of the deeper idle modes become unavailable if more than one core is active. This is even true on Qualcomm's chips, but at least they can clock each core asynchronously. All of MTK's chips so far are synchronously clocked, which means that if additional cores are lit up, they run at the same speed as the others, often with crippled cpuidle.
Part of that is that unlike his predecessor (Stringer) - Hirai realizes that treating your customers like shit is bad for business.
Sony under Hirai is very different from Sony under Stringer - this is most obvious if you look at Sony Mobile, who are one of the largest contributors to the Android Open Source Project (AOSP) and the only manufacturer that actively maintains AOSP builds for some of their devices. There are numerous signs that, rather than squash the anomalous behavior of the former Sony-Ericsson, the rest of Sony is starting to adopt Mobile's ways.
The thing is, the Real Money -> PLEX -> ISK thing is NOT the same as Real Money ISK
That intermediary item there changes things significantly, such that when you lose a big ship in EVE, unless you obtained it using PLEX, you didn't actually lose anything that was worth any real world money. Lots of media outlets made a big deal over "$9000 ship lost" when that Revenant was downed. The thing is - that guy didn't lose $9000. At worst he lost an opportunity to purchase $9000 worth of game subscription time at the worst pricing tier possible. (PLEX are around $20 for one month of time, a one month subscription is $15-16 or so, and if you prepay for multiple months it gets down to as low as $10-11/month.)
That's the real key here - in all of these transactions, only CCP can receive real world money. Yes, there are illicit real money trades that sometimes occur, but the exchange rate for these in ISK per dollar is shit (from the ISK seller's perspective) since it's inherently risky compared to PLEX. Someone selling ISK directly has to offer MUCH more ISK per dollar due to the massively increased risk compared to what a customer can do by buying a PLEX from CCP and selling it for ISK.
Yeah. My one encounter with a Titan (which occurred 2-3 weeks before quitting the game for 5 years) was:
Fleet commander initiates a fleetwarp to the gate
I go to warp
Client hangs as grid loads
Hey look I'm at my clone station
FYI, CCP did not slow anything down. TiDi kicks in automatically when server load goes up.
However, in the event of an anticipated big fight, CCP will move a system to a "reinforced node" (e.g. an extra beefy server). Of interest, though, is that many fights this year have not been entirely anticipated, resulting in TiDi being the only mechanism for load handling in play. For example, Asakai was the result of an unplanned misclick - it was supposed to be a small skirmish, but turned into a massive battle when a titan pilot jumped into the middle of the enemy fleet instead of bridging his fleetmates in. TiDi + unreinforced node = interesting mechanics, in which the fight is slowed down to 1/10 time, but most of the rest of the universe is running at normal speed. In the past, fleet battles were usually over in minutes, often with half of one fleet dying before they could even load the grid. Now they can last hours, long enough for reinforcements to arrive and massive escalation.
I've played EVE off and on (more off than on) since launch... I just came back from a five year break, I think I'm going to wind up quitting again in not too long. That said, reading the news of what's going on around the edges of the galaxy is pretty neat. I just don't have anywhere near enough time to participate in shit like that, and the game can easily get very boring if you're not in on the nullsec action.