Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
crest
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
23 ms
·
361.
▲
by
crest
2y ago
The problem with overcommitting beyond the sum of RAM and swap is that userspace processes can't protect themselves against the kernel's lies. You allocate memory, handle the result, access the address space and someone dies. The
362.
▲
by
crest
2y ago
You can still use ham packet radio protocols over ISM bands not held back by those stupid historical limitations and stay within the law.
363.
▲
by
crest
2y ago
Oh an old “Indy without the go.”
364.
▲
by
crest
2y ago
Do you only offload aes-gcm or do the NICs handle chacha20-poly1305 too?
365.
▲
by
crest
2y ago
Waymo should be punished at least as harsh as a single driver honking away every night.
366.
▲
by
crest
2y ago
Much worse. If you can make the radio controlled servo to throw the chain during sprints this could cause a specific rider to crash.
367.
▲
by
crest
2y ago
Executives behind bars and multiple times the revenue from their illicit deals should work as incentive to guide the industry.
368.
▲
by
crest
2y ago
There is not just the big end of year congress, but also lots of smaller events organised and run by regional CCC (like) groups in Europe e.g. MRMCD, EasterHegg, the Dutch camps changing the name every time (next one is WHY2025).
369.
▲
by
crest
2y ago
The term implies the existence of a contract. By his own clear statements Dmitry did the firmware on his own as a volunteer because he liked the RP2350 and wanted to to contribute to the badge project.
370.
▲
by
crest
2y ago
Not the RP2040. That chip has no boot security from anyone with physical access to the QSPI or SWD pins.
371.
▲
by
crest
2y ago
> no hardware integer division The RP2040 SIO block contains one hardware divider per CPU core.
372.
▲
by
crest
2y ago
Have the companies operating those constellations pay towards orbital observatories and science profits too.
373.
▲
by
crest
2y ago
The funniest to watch was systemd deciding to kill user processes on "last logout" so it would happily kill your elected PostgreSQL cluster leader as soon as the last follower is done catching up the WAL via rsync (over SSH). It w
374.
▲
by
crest
2y ago
The first comes with a lot of trade-offs, but I would love to see two QSPI controllers or at least two chip select lines that can be controlled by an extra address bit. I also want the DMA engine to be able to address the SIO block to acces
375.
▲
by
crest
2y ago
Z-buffers don't just require about an other frame buffer worth of memory, but also lots of read and write bandwidth per pixel. This extra memory bandwidth requirement made them expensive to implement well. High end implementations used
376.
▲
by
crest
2y ago
SMT also allows the cores to hide memory latency by executing the other thread(s). Taken to the extreme end this would give you a barrel CPU.
377.
▲
by
crest
2y ago
I wish FreeBSD and Linux had Solaris style doors extended to support kqueue()/epoll() in addition to threads.
378.
▲
by
crest
2y ago
Files are normally stored hierarchically (e.g. atomically move directories), and updated in place. Objects are normally considered to exist in a flat namespace and are written/replaced atomically. Object storage requires less expensive
379.
▲
by
crest
2y ago
Are you claiming that SeaweedFS requires twice as much RAM as the sum of the sizes of the stored objects?
380.
▲
by
crest
2y ago
I just can't get over Cabo's choice to put 65 (yes 65) Bit fixed sized ints in the wire format and to top it off by making it ones complement (the sign bit is part of the type and the longest fixed sized ints are 64 bit range)...
381.
▲
by
crest
2y ago
I assume it's supposed to work by having the compiler track the lifetime of the inodes. The compiler is expected to help with ephemeral references (the file system still has to store the link count to disk).
382.
▲
by
crest
2y ago
So because an incumbent exists ESA should give up, let their production and development pipeline rot away and "enjoy" the monopoly until it's too late to change something?
383.
▲
by
crest
2y ago
Because the cost is a lot lower. Paying a few percent of overcapacity doesn’t change that.
384.
▲
by
crest
2y ago
It's almost as if you should understand security critical C code before you start patching it to death.
385.
▲
by
crest
2y ago
Because Russian "offerings" are lacking in comparison, but they don't have access to a lot of it.
386.
▲
by
crest
2y ago
One of most important technical details about resources in early MacOS is that it allowed the system to swap resources by using double indirect pointers (aka handles) with the lock bit stuffed into the upper 8 bits of the 32 bit. Stealing t
387.
▲
by
crest
2y ago
Users come to expect the performance that was never promised or even properly requested. Once you efficiently load the system they complain because you overdelivered and they got used to it.
388.
▲
by
crest
2y ago
if only syslog() supported something like libxo...
389.
▲
by
crest
2y ago
When your "compressed database" is both larger and slower to query than the plaintext, has an unstable on disk format, and isn't even crash safe like journald's database I have to assume it's a sick joke and not a s
390.
▲
by
crest
2y ago
Should've been called the "Workshop on defective DRAM".
More ›