vmorgulis

I mean a "Suggest HN" people can vote for with no obligation to Apply (nor for YC to do something, like a Show HN).

A post from Emily Short about the game: https://emshort.wordpress.com/2016/04/04/draculaland-robin-j...

Very cool announce!

Do you think running webtorrent under Duktape could work (with appropriate network APIs) ?

Is it possible to suggest existing projects (in which I do not participate) ?

I have few in mind like Unicorn engine (http://www.unicorn-engine.org/).

Yes and the design of the source code browsing (by column).

Is it the same Kik as in the leftpad npm incident?

http://blog.npmjs.org/post/141577284765/kik-left-pad-and-npm

What I'd like is a language that permits me to trap a dynamic allocations with a knowledge of the callsite.

Instead of having specialized objects for manipulating memory (like deque<>, vector<> ...), this language would provide statically a profile (or signature) of the stack at compiletime. It would allow the user to apply strategies that are used in JITs (like inline caching).

For example, detect that an array is push or shift only:

- if it's push-only, preallocate memory by chunk (like vector.reserve()).

- if it's shift-only, trap the [] operator and change the head of the array by an offset (no resize)

The issues of GC and malloc are around the context.

Do you use a library for the multiprecision numbers?

Very cool!

It (almost) runs on copy.sh V86 emulator: https://copy.sh/v86/

I see the boot process with dots (like linux) and after I have the following message with a red background (if that can help):

    ExDOS64 version 0.01 built 26.03.2016
    =====================================
    Boot error: CPU is not 64-bit capable...

The last lines of the Javascript console:

   [CPU ] mode=real/16 paging=0 vm=0 iopl=0 cpl=0 if=1 cs:eip=0xF000:0x0000FFF0 cs_off=0x000F0000 flgs=0x202 ss:esp=0x0000:0x00000F32 log.js:13:9
   [CPU ] far return eip=0x0000CB01 cs=0xF000 stack_adjust=0x0

SLOCCount use the COCOMO method: http://www.dwheeler.com/sloccount/sloccount.html#cocomo

I'm on Debian and not familiar with Java. I did that:

    aptitude install openjdk-8-jdk

A previous topic about nbench: https://news.ycombinator.com/item?id=11252358

Interesting. C++ and Java are very near on my machine:

    java Prime
    9592 primes between 0 and 100000
    Time taken = 18.484
    9592 primes between 0 and 100000
    Time taken = 18.25
    9592 primes between 0 and 100000
    Time taken = 18.087
    ...

C++: https://news.ycombinator.com/item?id=11412820

Here is my results on a desktop PC from 2010 (i3 CPU 540 @ 3.07GHz × 4):

    g++ -O3 -o prime PrimeChecker.c
    time ./prime

    9592 primes between 0 and 100000
    Time taken = 18.0476

    real	0m18.073s
    user	0m18.048s
    sys 	0m0.000s

It seems the Moore's Law is no longer valid...

Thank you a lot. I'll try Tumbleweed.

That's great!

I found the piece of code I was looking for inside "tinycv" (about OpenCV): https://github.com/os-autoinst/os-autoinst/blob/3c2f5edda09b...

I switched to Debian from Fedora a year ago during the "freeze" in the release cycle. I wonder what is the stability level of Tumbleweed compared to a standard (non-beta) Fedora or a Debian testing.

Is it the good repo for Tumbleweed OpenQA test? https://github.com/os-autoinst/os-autoinst-distri-opensuse

"It uses openCV and a library of reference screenshots (with areas of interest selected to allow openQA to ignore things we don’t care about) which we call ‘needles’ ..."

OpenQA is a fascinating project. The tests are written in perl and I think the OpenCV magical stuff is around the "check_screen()" function.

For example, a test for Firefox:

    x11_start_program("firefox https://html5test.com/index.html", 6, {valid => 1});

    # makes firefox as default browser

    if (check_screen('firefox_default_browser')) {
        assert_and_click 'firefox_default_browser_yes';
    }

https://github.com/os-autoinst/os-autoinst-distri-opensuse/b...

>I think finding a good person is more important than modalities.

Yes it is. I've met many people before finding someone that I can trust a bit.

Carl Rogers was an interesting guy and he had a scientific approach (as I like understand how things work). It's a model that fit well to my mind.

https://www.youtube.com/watch?v=iMi7uY83z-U

I guess all models have nearly the same success rate.

I would recommend Person-centered therapy (Carl Rogers). It's a very human and soft approach.

https://en.wikipedia.org/wiki/Person-centered_therapy

It's a disease. The problem is that there is no unique and universal treatment. It depends on the person.

You should try something like see a doctor and if that doesn't work try something else until it works and it can work and take a bit of time.

    char key[5] = { 0 };

    Simple enough – this is supposed to zero the entire array, but instead it only zeroed the first four bytes.

In C++ and on the stack "{0}" is not required as memory is zeroized. On heap (in a new) "{0,0,0,0,0}" should be used (or memset()). Not sure about that in C.

My understanding is that it's not a bug.

I've seen the name of Gabriel Dos Reis in the committee member list. I wonder if he gave up "the Pivot" (and Stroustrup too)...

Funny.

https://github.com/gcc-mirror/gcc/blob/bd3f0a53c07086e978ea4...

I'm not sure of the behaviour but there is still a loop...

New items for the FQA :-)

Yes, we need that subset. We could call that RC++ (or rC++, like RPython is a reduced Python).

What I would put in that subset:

    - constructors/destructors (RAII)
    - auto (type inference)
    - class with template arguments (no nesting)
    - variadic templates for function parameters

What I would remove:

    - virtual
    - rtti
    - complex templates (nesting)
    - no STL (or something new and/or based on POSIX)
    - no private/protected in classes (public by default)
     - or remove class (only structures)     
    - operator overloading
    - namespaces
    - name mangling (a.b gives a_b)

I imagine a good subset as something that Thorvalds could approve and ideally allow in the kernel.

Thanks too.

Opus testing is impressive:

> Run thousands of hours of audio through the codec with many settings

- Can run the codec 6400x real time

- 7 days of computation is 122 years of audio

Author about content-addressing (on HN): https://bentrask.com/?q=hash://sha256/59fd0cb6d129452290291a...

> a2x should be provided by the asciidoc ...

Yes, thank you (I'm on Debian). I know a bit more after reading the wiki. I will follow the instructions for the sdk.

> I'd merge a pull request adding libsexl support!

Cool! I'm discovering the new world of custom tools for TI (KnightOS, Axe Parser, OS2...).

I will look at the SDL UI to see what I can do.

BTW z80 is good target for superoptimization and I'm very impressed by your toolchain and its history :)

Very cool and even more confidential technology ;-)

> I'm not a big fan of the Bell school of tech, but I definitely understand Rob Pike's source of pride in testifying that he never wrote a cursor-controlled terminal app...

Did he use his own terminal at that time ?

https://en.wikipedia.org/wiki/Blit_%28computer_terminal%29