Yes. I know. Let's try that again: so you're just assuming the existence of backdoors in the Blackphone specifically which make the mic accessible to *.gov?
Unless I missed something, we don't know for sure one way or the other whether the techniques the FBI have used to do that to other phones, will work on the Blackphone.
You can run all the apps you want but your still walking around with a beacon that has a mic and other neat features to offer any interested city/state/federal/mil/.com
So you're just assuming the existence of backdoors which make the mic accessible to *.gov?
It seems that leaving out a lot of that stuff would leave us with a leaner and nicer kernel. A lot of the compression stuff there is just because Linux has to have "everything".
Look like those in charge disagree, and presumably they have their reasons. There are downsides to being lean-and-mean, of course.
I doubt if you even know what kind of things it is used for.
I sincerely doubt it. Show me good numbers demonstrating that more people are killed by Muslims than Christians/atheists in the UK. If you cannot, we may assume you're just talking crap, and have no better basis for your views than religious prejudice.
Actually, I'd settle for a demonstration of the proportional equivalent of the claim.
You may not be aware, but thus offsetting the risks is not a subjective claim. It is a factual one. And I'm pretty damn sure it's bullshit.
On the otherhand, I do have a valid fear of that from muslims.
Err, in which country, pray?
Here in England I'm pretty sure I'm more likely to be harmed by a Christian or atheist than by a Muslim, just by virtue of the fact that there are more Christians and atheists than Muslims. My 'valid fear' doesn't really run across religious lines.
That is a more restricted definition. You need a stricter definition to safely elide code.
You're right. Now that I look at the 'concession to practicality' section, it's clear that purity and non-escaping arguments aren't enough to guarantee safe elision.
It seems to check if the function can modify it's arguments (for example, by zeroing them).
I don't get you. A pure function is allowed to perform mutation via reference-type (or pointer-type) arguments.
So what you are saying is that you are better than the thousands of folks working on gcc?
If compiler warnings were a 100% reliable way of indicating that something's awry, they'd be errors. They're necessarily broad generalisations. That's not to say they're useless, but saying no compiler-warning should ever be permitted is hardly self-evident.
I've worked on a large codebase where we were required to build with no compiler-warnings, so it's certainly possible. (Well, I guess that depends on how pedantic your compiler is, really.)
Right, which is special treatment. The compiler 'knows' that memset is what D calls a 'pure function'. If memset were just an arbitrary function, about which nothing else was known, it would not be possible to elide the call, as this might eliminate side-effects.
Perhaps though memset is an inline function and the compiler is able to figure out its 'purity' rather than having it listed specially.
GCC is giving memset special treatment, right? If GCC didn't know whether memset would side-effect, it wouldn't elide the call. One could presumably work-around this problem by creating a static library defining memset_secure, which just called memset. When calling memset_secure, GCC wouldn't know whether it side-effects, and so wouldn't elide the call.
(Unless you're allowing link-time optimisation. You'd have to be sure your memset_secure static-library was a real machine-code static library.)
Aside: in the D programming language, it is possible to express 'purity' of a function, and so enable this kind of function elision for arbitrary functions.
Slashdot Mirror
This is good enough to suspect that phone is backdoored
Disgree. It's not enough.
Nice touch about putting blackphone homepage in .ch domain
Any idea why they did that? Just to try to distance themselves from America? Weird.
Yes. I know. Let's try that again: so you're just assuming the existence of backdoors in the Blackphone specifically which make the mic accessible to *.gov?
Unless I missed something, we don't know for sure one way or the other whether the techniques the FBI have used to do that to other phones, will work on the Blackphone.
You can run all the apps you want but your still walking around with a beacon that has a mic and other neat features to offer any interested city/state/federal/mil/.com
So you're just assuming the existence of backdoors which make the mic accessible to *.gov?
It seems that leaving out a lot of that stuff would leave us with a leaner and nicer kernel. A lot of the compression stuff there is just because Linux has to have "everything".
Look like those in charge disagree, and presumably they have their reasons. There are downsides to being lean-and-mean, of course.
I doubt if you even know what kind of things it is used for.
You're right.
Because it can.
Serious. That's why.
Do you actually know about this specific feature of the kernel, or are you just trying to appear cynical and wisened?
One suspects it's there because it's actually used by something. One also suspects it's used for a reason.
I sincerely doubt it. Show me good numbers demonstrating that more people are killed by Muslims than Christians/atheists in the UK. If you cannot, we may assume you're just talking crap, and have no better basis for your views than religious prejudice.
Actually, I'd settle for a demonstration of the proportional equivalent of the claim.
You may not be aware, but thus offsetting the risks is not a subjective claim. It is a factual one. And I'm pretty damn sure it's bullshit.
On the otherhand, I do have a valid fear of that from muslims.
Err, in which country, pray?
Here in England I'm pretty sure I'm more likely to be harmed by a Christian or atheist than by a Muslim, just by virtue of the fact that there are more Christians and atheists than Muslims. My 'valid fear' doesn't really run across religious lines.
I presume the compiler makes those checks whenever it can. The 'purity' scheme is really about helping the programmer, not the compiler, of course.
That is a more restricted definition. You need a stricter definition to safely elide code.
You're right. Now that I look at the 'concession to practicality' section, it's clear that purity and non-escaping arguments aren't enough to guarantee safe elision.
It seems to check if the function can modify it's arguments (for example, by zeroing them).
I don't get you. A pure function is allowed to perform mutation via reference-type (or pointer-type) arguments.
Sure: finances.
Riiight... nothing to do with the fact that programming is hard.
And what about a bug in the sandboxing?
Seems simple enough: the sandbox itself is critical.
+4 Insightful. Good start. Now get +5 Brilliance and you're set.
Not the first time this has happened. PayPal are clowns.
In 2009 (ctrl-f for "bypass the use").
In 2011 (ctrl-f for "don't have your football"), where they allowed use of common-knowledge as a fallback if you didn't have your 'football'.
So what you are saying is that you are better than the thousands of folks working on gcc?
If compiler warnings were a 100% reliable way of indicating that something's awry, they'd be errors. They're necessarily broad generalisations. That's not to say they're useless, but saying no compiler-warning should ever be permitted is hardly self-evident.
I've worked on a large codebase where we were required to build with no compiler-warnings, so it's certainly possible. (Well, I guess that depends on how pedantic your compiler is, really.)
You could say that of any cross-platform GUI toolkit, but it doesn't mean they're useless.
In D's definition of a pure function, that's still pure.
Pure functions are functions which cannot access global or static, mutable state save through their arguments.
Which was my point. No-one is being put out of business.
Also, the quote in your sig goes back way further than some slashdotter.
Right, which is special treatment. The compiler 'knows' that memset is what D calls a 'pure function'. If memset were just an arbitrary function, about which nothing else was known, it would not be possible to elide the call, as this might eliminate side-effects.
Perhaps though memset is an inline function and the compiler is able to figure out its 'purity' rather than having it listed specially.
GCC is giving memset special treatment, right? If GCC didn't know whether memset would side-effect, it wouldn't elide the call. One could presumably work-around this problem by creating a static library defining memset_secure, which just called memset. When calling memset_secure, GCC wouldn't know whether it side-effects, and so wouldn't elide the call.
(Unless you're allowing link-time optimisation. You'd have to be sure your memset_secure static-library was a real machine-code static library.)
Aside: in the D programming language, it is possible to express 'purity' of a function, and so enable this kind of function elision for arbitrary functions.
Well yeah. Wasteful consumerist societies.
They are "dumping", as in selling below cost! What this does is undercut the new product business, and someone was losing money.
That would be a problem, if they were putting African TV manufacturers out of business. I don't think they are, though...
Nothing better than to have competition thrown in jail.
Dumping has nothing to do with jailing the competition.
Help Brendan pay off his student loans [brendansstudentloans.com]
You're asking for free money from strangers, to pay off your student loans?
Classy. I tend to donate to more urgent causes, personally.
No. The forces at play would be massively diluted, so the issue would be lessened enormously.
To higher government, the money raised in these auctions is just spare change.
Incidentally, this is precisely how things are done in the UK. The proceeds of traffic tickets are handled similarly.
Got anything to back up that cynicism?
I'm not convinced Mozilla are sellouts.