Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
dom0
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
43 ms
·
511.
▲
by
dom0
9y ago
KDF != PB-KDF / password hashes. There is no need that a KDF needs to be slow, e.g. you want fast handshakes for protocols (e.g. TLS), not slow handshakes.
512.
▲
by
dom0
9y ago
https://www.youtube.com/watch?v=zrWoG8IckyE
513.
▲
by
dom0
9y ago
> Context switching is done for threads in HW. Uhm, x86-64 can't switch between tasks/contexts in hardware.
514.
▲
by
dom0
9y ago
That's not correct. THD measures some non-linearities and is unable to detect others (e.g. IMD).
515.
▲
by
dom0
9y ago
Frequency and phase response are just linear characterisations, they don't model any non-linear effects. That being said, speakers and headphones all have a sound (unlike any properly designed amplifier); there is hardly any right or
516.
▲
by
dom0
9y ago
> We need objective benchmarks for everything. You might want to read Zen and the Art of Motorcycle Maintenance which digs into this fallacy. You might also be surprised how misleading objective IQ benchmarks are; the post processing i
517.
▲
by
dom0
9y ago
Python 2 has old-style classes, new-style classes, functions, bound methods, unbound methods, slot wrappers, descriptors, method wrappers, built-ins, ... and I'm rather certain I forgot some. All kinda callable types I mentioned behave
518.
▲
by
dom0
9y ago
A good formal reference uses way-over-the-top wording, a bad formal reference basically just says "ok". This is due to German law not allowing for "the employee sucks", nor allows it to use "code-speak" (it was
519.
▲
by
dom0
9y ago
Data protection laws set quite tight boundaries on chatting between former and prospective employers, without the employee's consent. Further, I think the general consensus is that such chats are thought to be subject to the same rules
520.
▲
by
dom0
9y ago
Germany, for example.
521.
▲
by
dom0
9y ago
Yes, but differently; C/C++ explicitly prohibit zero-sized types due to object identity.
522.
▲
by
dom0
9y ago
Ah, I almost got it right. It discarded too much, thus reducing throughput - even more subtle. https://github.com/torvalds/linux/commit/18aba41cbfbcd138e9f...
523.
▲
by
dom0
9y ago
Since I have to squint my eyes to read your comment the principle you postulate seems to work.
524.
▲
by
dom0
9y ago
Linux ~4.7 or so fixed a bug in fadvise, specifically FADV_DONTNEED, that incorrectly rounded page boundaries to the effect of making some calls less effective. Found and fixed by a developer who wondered why his page cache was filling up,
525.
▲
by
dom0
9y ago
Flagged as possible replacement for my Visio-in-a-VM workflow (I don't use most of Visio's vast feature set), since it can do that: https://news.ycombinator.com/item?id=14555490
526.
▲
by
dom0
9y ago
That doesn't really help if the exporting application distributed the figure across multiple pages. E.g. in Visio you have to work around that by using the "export current viewport" option, which is quite annoying and irrepro
527.
▲
by
dom0
9y ago
rr := https://github.com/mozilla/rr rr is a lightweight tool for recording and replaying execution of applications (trees of processes and threads).
528.
▲
by
dom0
9y ago
Though I agree that it was an annoying change. hex/fromhex have helped there, though a 3.4 codebase retains stupid helpers like bin_to_hex.
529.
▲
by
dom0
9y ago
import base64
530.
▲
by
dom0
9y ago
Python 2 is a vastly more complex language compared to Python 3. Just try to name the basic classes of types that exist in Python 2. Hint: Cannot be counted on two hands.
531.
▲
by
dom0
9y ago
And/or because many Python projects have completely insufficient testing which means that any change that might break anything anywhere can go undetected.
532.
▲
by
dom0
9y ago
Nowadays it's more likely that a Python 2 user wants to use a library and finds it doesn't work on Python 2, because the developers never targeted nor tested on 2.
533.
▲
by
dom0
9y ago
So, you see, I know a bit about writing efficient code. I have held a firm belief, and still do, that if you're doing heavy lifting in CPython, you're doing it wrong, and better parallelism will not significantly help you.
534.
▲
by
dom0
9y ago
PDF export: can it create automatically cropped PDFs, e.g. for inclusion as bare figures in LaTeX? Can it render PNGs at high qualities?
535.
▲
by
dom0
9y ago
Databases use journaling to avoid these issues. The difference between RDBMS' journaling and file system's journaling is the RDBMS journals application data, the file system does not (with few exceptions). Thus, a RDBMS gives you
536.
▲
by
dom0
9y ago
Personally I don't think sectioning is that bad, a little annoying when working with other people that don't do it, though... foo ====================== ... the easy and built-in extensibility is a huge plus though, as
537.
▲
by
dom0
9y ago
restructuredtext?
538.
▲
by
dom0
9y ago
Mice for MOBAs and such have up to tweleve extra hotkeys...
539.
▲
by
dom0
9y ago
while(1){write();} is about as far away as it gets from actual application code. This topic may be fun, but remains an utterly pointless micro-benchmark.
540.
▲
by
dom0
9y ago
It's almost funny that you try to refute, or rather, dismiss, criticism of PHP with "but it's fast", when none of what you cite even mentions that. I just want to say: STFU.
More ›