Umm, that paper barely mentions memory management, except to say that 1: C++'s destructors make things possible that other languages don't, and 2: garbage collection can't be guaranteed in C++. It has little to do with this discussion on memory leaks.
Valgrind doesn't require you to recompile/relink your code. Also, Valgrind allows you to do extra runtime checking on specified areas of memory, which I don't think checkergcc provides.
Re:Advocacy, we never knew thee.
on
Qt vs MFC
·
· Score: 1
Now for the explanation of the "almost": Swapping the contents of two general-purpose registers is completely stupid.
Ack. That's what I get for getting caught up in defending a neat hack; I overlooked the fact that the only useful situation I could think of wasn't sane. However, in a vain attempt to regain my dignity, I'll mention that x86 "general-purpose" registers aren't completely general purpose. You might want to exchange esi or edi with other registers, for use with the REP instructions you mentioned.
Re:Advocacy, we never knew thee.
on
Qt vs MFC
·
· Score: 2
Please pardon my sloppy HTML.
Well, that depends on your compiler. Testing with gcc 3.1 with -O3, I found that gcc has a special optimization for the standard swap construct (I also tried using the Intel compiler, but it produced nasty code that I had no desire to dissect). It assembles to something like this, where the bold variables are registers, and the others live in memory:
ax = a;//read bx = b;//read a = bx;//write b = ax;//write
This code has 2 memory reads and 2 writes, whereas the xor construct assembles to a monstrosity with far more memory access than necessary. So yes, on gcc the normal swap is much faster, but a well-assembled xor construct, with just the 4 required memory accesses, would probably only be a few cycles slower, rather than a few hundred. The memory accesses contribute the great bulk of time consumed.
You're not quite right here. Most assembly operations are like operation= in C. There is no temporary register used to store the result of a^b, because the operation is done in hardware as a^=b. You're right in this case that they both use the same amount of storage, but that's only because gcc recognized a common construct and optimized it.
No comment.
It's almost intuitive to people who have learned assembly, but probably not to anyone else. In fact, in light of gcc's particular quirks, hand-written assembly, where the variables are already in registers, is probably the only place xor-swapping should be used. It may be an advantage there, because it only uses two registers instead of three for the normal swap, and on x86, registers are at a premium.
Problem is, all the innovative ideas have already been innovated. All these new-fangled patents are just rehashes. So you wouldn't make much money on licensing fees.
I believe he meant that philosophy, referring to the "manifest destiny." The philosophies you mention are of course major and excellent reasons to use Linux.
Because the ext2 code is more mature than the ext3 code. I also read that the ext2 code is currently much better suited to SMP, but ext3 hasn't been worked over to work well with multiple processes/processors.
I can't imagine what 20 formats you might have, but as for command-line players, may I suggest mpg321 for mp3s, ogg123 for oggs, and mplayer for anything it can manage. mpg321 now has "remote control" mode, so you can load different files and pause, play etc.
I'm guessing you have some files in real format; mplayer seems to be making progress in playing real format files, at least the older codecs.
They're all open source, so if you were feeling really ambitious, you could even glom them all together into one program.
I sometimes manage 3.2 Mbit with Adelphia (west PA). I suspect I could get even better speeds if I wasn't at the end of the line (rural area). The catch is that I'm limited to a paltry 2.5GB per month. I don't know if they enforce it, but it does annoy me every month.
Vendors should make a setup like mine then: My power button is linked to the motherboard's ACPI functionality (this is pretty standard now I believe), and I'm running acpid, which monitors acpi events. Whenever I press my power button, "init 0" automatically gets run, and powers off my machine safely.
Just for usefulness, I set it up so that pressing the power button twice reboots safely instead (since my reset button apparently doesn't hook in to ACPI, grr). This could be especially useful for stupid Dells that don't have a reset button.
This is exactly the reason I will never use Linux. Talk about WinXP's bloat all you want, check out the latest Linux distributions...
That's just... silly. You can get exactly as much bloat as you like, just choose the appropriate distro. Use linux on a floppy, it's usable at least for specialized purposes. I can't speak for other distros, but Debian's base installation is less than 100MB IIRC. Not terribly useful at that point, but usable.
Speaking of bloat, Debian recently passed the 10,000 packages mark. Many of them are silly or superfluous. However, the redeeming quality is that you don't have to install much unwanted software. When a package depends on oodles of unnecessary crap (like kde and gnome libraries), there's usually a lighter alternative.
As an example, I have about 250 packages installed on my firewall (including many I don't really need, such as complete X and gui programs), and about 550 on my desktop.
The only area I think is uncomfortably bloating is that each new version of a package is slightly bigger than the last. This is probably feature creep, but some of it is also probably new documentation (which is good!).
Anyway, my point is, you have a ridiculous number of distributions to choose from. Don't judge Linux as a whole based on the mainstream distros, which are trying to gain marketshare by making Linux more Windows-y. Linux is trying to be "everything to everyone," but each distro caters to a different subset of "everyone."
Why on earth is the whole of Slashdot arguing for more internet legislation?
They're not. Several posts have recognized that the peculiar accusations of these "cyberskeptics" (oh how I hate "cyber-" words) against "cyberbuffs" don't accurately describe the "technophiles'" beliefs.
I get the impression from the article that these skeptics think technophiles are a bunch of whiny, arrogant, short-sited buffoons (well, this is slashdot). But it's not the technophiles who want more legislation, it's the technophobes, like the RIAA, censorship advocates, and other business interests.
These skeptics don't even understand who they should be accusing of short-sitedness (besides themselves). They seem too confused to be considered authorities on the subject, yet they're publishing articles about their views, which are widely read by lawyerly people.
I for one would be thrilled if new Internet legislation was halted, or even mostly repealed. I think many slashdotters would be with me on this.
I like your point about communities. Not only can the "community [feel] that cyberspace is something different," but the community can basically exist in cyberspace, because physically, they're spread over a huge area. Furthermore, Internet communities make their own rules, in the form of "'netiqutte," almost like a set of laws.
It's also interesting to note that while you say they take a radical position, they would probably call you the radical.
(Nearly?) all bootable CDs allow you to do this. They have a 1.44MB floppy image on the CD that you can write to a floppy using dd in Linux, or with the supplied rawrite2 in Windows. Then you can boot the floppy and it will start Linux from the CD. I know at least the Debian and RedHat boot cds, Knoppix, and DemoLinux all have this. I recommend Knoppix, BTW; it was only a month out of date last time I checked.
Re:In all seriousness, random libs *suck*
on
Pet Bugs?
·
· Score: 1
You should have just hooked up a radio to the sound card or something:)
I don't think he's jumping to conclusions. He doesn't like being forced to upgrade without explanation. I don't either. Part of his annoyance is probably derived from the Debian security announcement he quoted. The patch he was referring to was the new version of ssh, not the theoretical vulnerability fix. Before accusing others of misinterpretation, please make sure you are interpreting their words accurately.
So read it and test it - you have the code. What more do you want?
The new version doesn't fix the vulnerability. It merely makes it impossible to exploit due to it now being run chroot and not as root. So looking at the new version wouldn't show you what the problem is, only that it's in the chroot'd portion of the new code.
And because everyone carries all the equipment (UV lights with generator, etc.) to check bills everywhere, you never have to be swindled.
You donh't need a UV light, just look at the watermark. Or the embedded strip. Or the colored threads. Or the color-shifting ink. Or the tiny writing. Or the...
It would make more sense if they had security devices that are both extremely hard to copy AND easy to spot at a glance... [emphasis mine]
Anything is easy to counterfeit if you only glance at it. The checks I mentioned are just as easy as putting it under a light or marking it with a pen. No need to get ripped off.
Slashdot Mirror
Umm, that paper barely mentions memory management, except to say that 1: C++'s destructors make things possible that other languages don't, and 2: garbage collection can't be guaranteed in C++. It has little to do with this discussion on memory leaks.
Valgrind doesn't require you to recompile/relink your code. Also, Valgrind allows you to do extra runtime checking on specified areas of memory, which I don't think checkergcc provides.
This code has 2 memory reads and 2 writes, whereas the xor construct assembles to a monstrosity with far more memory access than necessary. So yes, on gcc the normal swap is much faster, but a well-assembled xor construct, with just the 4 required memory accesses, would probably only be a few cycles slower, rather than a few hundred. The memory accesses contribute the great bulk of time consumed.
Problem is, all the innovative ideas have already been innovated. All these new-fangled patents are just rehashes. So you wouldn't make much money on licensing fees.
I believe he meant that philosophy, referring to the "manifest destiny." The philosophies you mention are of course major and excellent reasons to use Linux.
Shows you how many people read the article.
Yes, my uni's math department uses/knows about UNIX more than the CS department. It's sad, really.
Because the ext2 code is more mature than the ext3 code. I also read that the ext2 code is currently much better suited to SMP, but ext3 hasn't been worked over to work well with multiple processes/processors.
I'm guessing you have some files in real format; mplayer seems to be making progress in playing real format files, at least the older codecs.
They're all open source, so if you were feeling really ambitious, you could even glom them all together into one program.
Humor? What's that, the stuff in your eye? (vitreous humor, for the people I'm referring to)
I sometimes manage 3.2 Mbit with Adelphia (west PA). I suspect I could get even better speeds if I wasn't at the end of the line (rural area). The catch is that I'm limited to a paltry 2.5GB per month. I don't know if they enforce it, but it does annoy me every month.
Just for usefulness, I set it up so that pressing the power button twice reboots safely instead (since my reset button apparently doesn't hook in to ACPI, grr). This could be especially useful for stupid Dells that don't have a reset button.
Anyway, I'm happily running sid on one computer, and woody on another, so no Debian problems here :)
Speaking of bloat, Debian recently passed the 10,000 packages mark. Many of them are silly or superfluous. However, the redeeming quality is that you don't have to install much unwanted software. When a package depends on oodles of unnecessary crap (like kde and gnome libraries), there's usually a lighter alternative.
As an example, I have about 250 packages installed on my firewall (including many I don't really need, such as complete X and gui programs), and about 550 on my desktop.
The only area I think is uncomfortably bloating is that each new version of a package is slightly bigger than the last. This is probably feature creep, but some of it is also probably new documentation (which is good!).
Anyway, my point is, you have a ridiculous number of distributions to choose from. Don't judge Linux as a whole based on the mainstream distros, which are trying to gain marketshare by making Linux more Windows-y. Linux is trying to be "everything to everyone," but each distro caters to a different subset of "everyone."
I get the impression from the article that these skeptics think technophiles are a bunch of whiny, arrogant, short-sited buffoons (well, this is slashdot). But it's not the technophiles who want more legislation, it's the technophobes, like the RIAA, censorship advocates, and other business interests.
These skeptics don't even understand who they should be accusing of short-sitedness (besides themselves). They seem too confused to be considered authorities on the subject, yet they're publishing articles about their views, which are widely read by lawyerly people.
I for one would be thrilled if new Internet legislation was halted, or even mostly repealed. I think many slashdotters would be with me on this.
It's also interesting to note that while you say they take a radical position, they would probably call you the radical.
(Nearly?) all bootable CDs allow you to do this. They have a 1.44MB floppy image on the CD that you can write to a floppy using dd in Linux, or with the supplied rawrite2 in Windows. Then you can boot the floppy and it will start Linux from the CD. I know at least the Debian and RedHat boot cds, Knoppix, and DemoLinux all have this. I recommend Knoppix, BTW; it was only a month out of date last time I checked.
You should have just hooked up a radio to the sound card or something :)
I don't think he's jumping to conclusions. He doesn't like being forced to upgrade without explanation. I don't either. Part of his annoyance is probably derived from the Debian security announcement he quoted. The patch he was referring to was the new version of ssh, not the theoretical vulnerability fix. Before accusing others of misinterpretation, please make sure you are interpreting their words accurately.
The new version doesn't fix the vulnerability. It merely makes it impossible to exploit due to it now being run chroot and not as root. So looking at the new version wouldn't show you what the problem is, only that it's in the chroot'd portion of the new code.
Gives a new meaning to the term "Monopoly money."
You donh't need a UV light, just look at the watermark. Or the embedded strip. Or the colored threads. Or the color-shifting ink. Or the tiny writing. Or the
Anything is easy to counterfeit if you only glance at it. The checks I mentioned are just as easy as putting it under a light or marking it with a pen. No need to get ripped off.
/etc/pam.d/passwd: password required pam_unix.so nullok obscure min=4 max=20 md5
'nuff said.