Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
fenollp
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
9 ms
·
61.
▲
by
fenollp
9y ago
In Erlang you cannot run non-BEAM code for more than ~10ms or your VM will crash. GEMM will be hard to use this way...
62.
▲
by
fenollp
9y ago
It uses `array` (somewhat mutable Erlang structure) and NOTP (no idea how it makes code "zippy" and the repo [1] does not explain anything... it seems to be bypassing the normal way modules are loaded?). I am unsure why anyone wou
63.
▲
by
fenollp
9y ago
Here is my own DSL. Interesting that it turns out we have similar aliases... alias ga='git add' alias gb='git branch' alias gbD='gb -D ' alias gbc='gc -b ' alias gbd='gb -d
64.
▲
by
fenollp
9y ago
Here is what it could look like when/if they do restore this area as marais salant: https://goo.gl/maps/d5ATzvYyxs22
65.
▲
by
fenollp
9y ago
Leibniz didn't say happiest did he? I thought he meant the best of all possible timelines, compromise-wise.
66.
▲
by
fenollp
9y ago
I don't know anything about this field but the language seems sound to me. Is there a consensus on this interpretation v. "just adding more dark matter"?
67.
▲
by
fenollp
10y ago
I'm guessing this is doing some kind of tree-diff on the DOM? Now if you would have this generate a graphQL spec file, you could run a graphQL server acting as a proxy to lots of websites. That would be interesting. Not sure how that f
68.
▲
by
fenollp
10y ago
Possible side effects correction: -> if (a) else (b) foo = a+b
69.
▲
by
fenollp
10y ago
Here's the paper: https://pdfs.semanticscholar.org/17fb/c72811582e03bde433a709... As a color blind person (as are 20% of half the population at least), I wonder if we can create glasses that mimic this shape of le
70.
▲
by
fenollp
10y ago
> I have done multiple tests and then took an average, don't know how much these results can be reliable (since to get statistics more plausible we perhaps should have allocated much more memory). Yep. Nothing to see here
71.
▲
by
fenollp
10y ago
http://explainshell.com/explain?cmd=at Or use a CI cron running daily?
72.
▲
by
fenollp
10y ago
Most important to me from this release: webdriver finally supports self-signed certificates.
73.
▲
by
fenollp
10y ago
Less than 700 million years apparently. https://twitter.com/humansareawesme/status/83696170574585856...
74.
▲
by
fenollp
10y ago
That list comprehension though. Could be useful to math-fluent newcomers.
75.
▲
by
fenollp
10y ago
Exactly. I made this for exactly this usage: https://github.com/fenollp/nginx_ssl_compose Just create a folder in ~/www for each host. It's been working great for around a year. Only interruption was due to d
76.
▲
by
fenollp
10y ago
> has released over 500,000 lines of open-source code Interesting metric. I wonder why no "social" VCSs report that metric?
77.
▲
by
fenollp
10y ago
Broke https://chrome.google.com/webstore/detail/deluminate/iebboop...
78.
▲
by
fenollp
10y ago
Dyalizer solves your example already. But finding race conditions is another problem entirely. Have a look at Concuerror and PropEr.
79.
▲
by
fenollp
10y ago
They seem to have done this only for Webpack (splitting JS libs + prefetching). Why use Node at all? It's all static content... Crazy that a top engineering post is about some company discovering "compile-time". Next post wil
80.
▲
by
fenollp
10y ago
Why not plug into LLVM? Just generate an LLVM backend.
81.
▲
by
fenollp
10y ago
You should look for sitemaps. They are often listed in robots.txt. Will save you lots of time! These sites have sitemaps: http://torrentproject.se/robots.txt https://www.torrentdownloads.me/robots.txt http
82.
▲
by
fenollp
10y ago
It'll probably just be a cloud API "look, we can collect more data". Hopefully it will be an efficient implementation of 1bit-weight NNs like XNOR.ai (which has been pushing on the research on 8bit & 3bit nets). ARM SIMD
83.
▲
by
fenollp
10y ago
Alternative facts will push us to find a way to discontinue them. What is left to do to have truly open computation? * somewhat-to completely free computational power?: CircleCI, Travis CI, whatever * open source repositories to run code fr
84.
▲
by
fenollp
10y ago
Easy: add a string to the zipper, like for scuba diver suits.
85.
▲
by
fenollp
10y ago
So the core idea is similar to https://github.com/purifycss/purifycss but for JS code. Then why does Svelte wants to be yet another framework? Do a real good one to the world and be something that you can plug into Gul
86.
▲
by
fenollp
10y ago
Time to implement this idea: Use open source projects' CI file to find which dependencies / commands are needed to compile / run the project. Run all this inside whichever container the CI system uses (circle & travis use
87.
▲
by
fenollp
10y ago
WRT the "timing attack": In most cases, client does not even request bytes from CDN which is then not able to track Client. But then again CDNs can implement tracking based on this lack of requests (which is kind of ironic and sho
88.
▲
by
fenollp
10y ago
> CDN is common enough technique which should be standardized in browsers. HTML should include link to resource hosted by site and its checksum. Now browser can easily use cached resource from any other site with the same checksum or jus
89.
▲
by
fenollp
10y ago
CircleCI gives you 30 minutes of playtime on a 32 cores, 60GB RAM VM. All you have to do is log in with GitHub, start a build of any of your public repo & click "Rebuild with SSH". (No affiliations, I just find this mind blowi
90.
▲
by
fenollp
10y ago
Hey Andreas, I'm unable to find the feature branch bringing Parallella support to OTP https://github.com/margnus1/otp/branches Maybe it was merged upstream already? You came a long way since I saw you in Lond
More ›