bodyfour

37 ms·

271.
by bodyfour 13y ago
The short answer: very few state legislators have auto manufacturers in their districts, but every one of them has car dealers.
272.
by bodyfour 13y ago
The two-write operation you propose is simpler, but you're losing a lot of generality. Specifically, think about: * I want to notify a network client that its write is done * When this write is done, I want to delete a temporary
273.
by bodyfour 13y ago
True, and it's good of you to clarify. However, what if the "other pending writes" are to the same file descriptor? Just because I'm eager to know when one write happens doesn't mean I care about other writes I ma
274.
by bodyfour 13y ago
> or you have allocated blocks or otherwise performanced a write which required updating file system metadata What if you're appending to a file, and want to checkpoint every so often. I guess you can be clever with fallocate(FALLO
275.
by bodyfour 13y ago
> So a little bit of data loss is okay, but a lot isn't? No data loss is OK. However, delaying acknowledgement to the network client requesting the db update often is OK. Especially if your protocol allows many simultaneous requ
276.
by bodyfour 13y ago
> you could literally be waiting forever Technically that's true, but it's not an issue in real life. Any sensible operating system will flush dirty buffers to disk eventually. Imagine if a sudden power outage lost writes you
277.
by bodyfour 13y ago
I think part of the problem is that fsync() is that it's an insufficient interface. Most of the time, you want two things: * write ordering ("write B must hit disk after write A") * notification ("let me know when w
278.
by bodyfour 13y ago
The problem isn't that they don't have enough time to get from point A to point B. The delays happen all over the line. Since the route has dozens of stops, and a multi-hour delay could happen between any of them, where do you pu
279.
by bodyfour 13y ago
While I don't share the cartoonists political views, I can't blame him for the frustration. From http://en.wikipedia.org/wiki/Coast_Starlight > In recent years the train acquired the nickname "Coast
280.
by bodyfour 13y ago
I'm usually not much of a Java fan, but making bool a stricter type was a good change. I always write my C code as if it had the same rules. For example, if you have: bool b; int i; unsigned u; void *p; You could
281.
by bodyfour 13y ago
How is it worse than: $ ./configure && make I guarantee you that it'd be easier to hide something nefarious in 3000 lines of autoconf boilerplate.
282.
by bodyfour 13y ago
It's a war story from my father, so I have to admit that the details could be corrupted. He worked with a variety of machines in the 60s, so it was believable that one had a permissive enough FORTRAN to allow that. It's also pos
283.
by bodyfour 13y ago
Certain ancient FORTRAN compilers would let you do that as well. If the target CPU didn't have a "move immediate" instruction, the compiler would just stash the number into memory. The quickest way to implement it is to jus
284.
by bodyfour 13y ago
> if we had to wait for a 60GB mailbox to sync every time we loaded our webmail, well, obviously that's not going to work First, the Merkle tree would just be of the IDs not the message contents. I.e. the problem it is trying to s
285.
by bodyfour 13y ago
> the criticism is usually that IMAP does not require maintaining UIDs Yes, exactly. UIDs should be sticky, but they should also be allocatable by widely distributed nodes without the need for central coordination. If your UIDs are 32
286.
by bodyfour 13y ago
I think the LSB bits of arrival time is actually a really good distribution long-term (remember: we want the distribution to be uneven in the short term, but even in the long term) Maybe you could improve it a little bit by making each bin
287.
by bodyfour 13y ago
Do those IMAP servers allow separate control of the envelope recipients, or does it just take them directly from the message headers? To be fair, POP3's "XTND XMIT" had that flaw as well (IIRC qpopper just passed the message
288.
by bodyfour 13y ago
The reinjection on clash (CSMA/CD for numbers :-) works in close-knit clusters but has issues with network splits. Imagine one server is in SF and the other is in Amsterdam, and a network problem makes them unable to talk for a few ho
289.
by bodyfour 13y ago
Specifically, you want a Merkle tree where the final bucket is based on the lowest N bits of the message arrival time, probably at a granularity of a several seconds. In the typical sync case all of the unseen messages will then be cluster
290.
by bodyfour 13y ago
Unfortunately it includes my biggest gripe about IMAP -- the requirement that messages are given a server-side message number. Maybe breaking with IMAP on this would make interop too hard though. This requirement is convenient for the clie
291.
by bodyfour 13y ago
If you choose a random nonce then it follows that a nonce could be randomly reused. If there are N possible nonces this will happen on average after approximately sqrt(N) packets. If you use a counter as a nonce it will only repeat after
292.
by bodyfour 13y ago
I'm not the commenter you're replying to, but I'll take a crack. Saying it was a matter of "simplicity" glosses over too many of the details to be useful. Also consider that many things that early UNIX didn't
293.
by bodyfour 13y ago
Shame this is getting downvoted, probably by people who don't understand the reference. See the second section ("You are not expected to understand this") in: http://cm.bell-labs.com/cm/cs/who/
294.
by bodyfour 13y ago
I wasn't aware classic MacOS worked that way. The "sockets-emulated-over-STREAMS" was the standard way of doing things in SVR4-based UNIXes. In the early-to-mid 90s this gave them a poor reputation among early web admins si
295.
by bodyfour 13y ago
More or less, yes. In the original TCP implementations from the early 80s, performance was (understandably) not the main priority.. they just wanted to get it working first. Also, nobody was sure what networking protocols would become popu
296.
by bodyfour 13y ago
Yeah, any modern UNIX supports the same sendmsg/SCM_RIGHTS pattern. You might need to tweak the includes and supply your own CMSG_*() macros on some though. Some older UNIX variants only supported the BSD4.3-style msg_accrights style
297.
by bodyfour 13y ago
Does anyone know approximately how close we are to 0.12? I know 0.11.9 came out a few weeks ago, but I don't follow the community close enough to get a feel for how near the next stable branch is.
298.
by bodyfour 13y ago
> Why isn't laboratory analysis blinded? Exactly my thought. The chemists doing the analysis shouldn't have any way of finding out who the defendant or prosecutor in the case is until they've given their report. Similarl
299.
by bodyfour 13y ago
FWIW, the boost::filesystem C++ package uses the / operator in the same way
300.
by bodyfour 13y ago
I think almost every restaurant in London would be considered "vegetarian friendly", at least by US standards. No matter where I went, there were multiple options clearly marked on the menu. It's the easiest place I've

More ›