/g/ - Technology & Programing

Anonymous ## Admin
>>2206559 #
Works fine for me. Which mod on which board was trying to make the ban? Maybe something is more subtly broken.

I changed the JSON API a bit. /json/:board is now /json/boards/:boardand same for thread queries.
>>2184630 #>>2192931 #
GIqSnWaDsEx649 KB1023x1281vb_watcher.png
for >>2201289 ➡ #

1. You must use passive virtual boarding. If you are new user, you don't have to do anything, because it's default. If you don't know about details, click "I'm beginner...." in EasySetting2 menu. See https://github.com/DogMan8/CatChan/wiki/Setup
2. Go to https://meguca.org/all/catalog
3. Set filter, see pic. Open filter console, select your favorite timespan, then adjacent time field will be filled automatically, or you can write it manually, check left "Watch" checkbox.
4. Reload.
5. All boards which have new posts in your timespan will be green in boardlist.

There were some other css fixes that were lost before the 2 week rule was changed. An update to the horizontal posting fix and one thing to make it irc like, plus another theme I think. anyone have them?
>>2114864 #
>>2114196 #
haven't tested them recently, some probably need adjustments to the new post structure, especially the irc thing

#IRC mode

/* grid */
article:not(.reply-form) {
display: grid;
grid-template: "a c d"
"a b d";
grid-template-columns: min-content minmax(144px, auto) auto;
article:not(.reply-form) header {grid-area: a;}
article:not(.reply-form) figure {grid-area: c;}
article:not(.reply-form) .backlinks {grid-area: b;align-self: end;}
article:not(.reply-form) .post-container {display: contents;}
article:not(.reply-form) blockquote {grid-area: d;}

/* stack backlinks*/
article:not(.reply-form) .backlinks {display:flex;}

/* strip non-essentials */
article:not(.reply-form) header time {display:none;}
article:not(.reply-form) header .name {display:none;}
article:not(.reply-form) header .control {display:none;}
article:not(.reply-form) figcaption {display: none;}

/* spacing */
article:not(.reply-form) header {margin-right:5px;}

/* kill reddit-spacing */
article:not(.reply-form) blockquote br { line-height: 0px;display: block;}

## minor tweaks

makes the custom CSS box larger:
#customCSS {min-width:35em;min-height:13em;}

only expand the moderation panel on hover:
#moderation-panel:not(:hover) form {display:none;}
#moderation-panel:not(:hover)::before {content: '🞀';}

space the inputs at the bottom of the reply form to avoid hitting the spoiler link by accident:
#post-controls {display:flex;width:100%;justify-content: space-between;}

## compact horizontal

#thread-container {display:flex;flex-direction: row;flex-wrap: wrap;align-items: stretch;}
GIqSnWaDsEx260 KB1798x1055Screen Shot 2017-06-04 at 3.19.58 PM.png
Ultra-compact horizontal

#thread-container {
flex-direction: row;
flex-wrap: wrap;
align-items: stretch;

article.media {
min-width: calc(100% - 43px);
max-width: calc(100% - 43px);

article.glass:not(.media) {
min-width: calc(20% - 25px);
max-width: calc(20% - 25px);

GIqSnWaDsEx158 KB1020x11214chan_qa_1380723.png
As I posted in 4chan, I can't understand the situation. I'll repost them here. See 1389936 and later.

Overdetection might be caused by a bug of 4chan. I can't understand why this situation was occurred. Pic related. Can anyone explain this?

Pics are contents of catalog.json.
Pic0: thread at page 0.1, time of last post was 1497199058. (threadA)
Pic1: thread at page 0.2, time of last post was 1497198820. (threadB)
Pic2: thread at page 0.3, time of last post was 1497199203. (threadC)
ThreadA and threadB was overestimated clearly.
Pic3: details of detection. Time of last post in a thread at page 0.5 was 1497199210, and later was seems to be ordered correctly. (last 3 letters were 210, 209, 209, 202, 200, 201(probably sage), 200, 199) But threads at page 0.3 and 0.4 were also overestimated.

This thread will be pruned, so I'll wait in meguca also.

Anonymous ## Admin
>>2140643 #
> Are banners coming up?
Yes, but they will take time.
>Also, can you adjust the thread delete algorithm to make trashbinned threads have a shorter lifespan? Like 30 days vs 90?
>Right now the bottom of /pol/'s catalog has a bunch of old deleted threads, most of them accidental duplicates, and they're going to sit there for the full 90 days each time it happens.
I need to think about it. Particularly about the implementation side.
>>2142668 #
That site is down, but what exactly do you have in mind?
>>2143820 #
>>2143640 #
>That site is down, but what exactly do you have in mind?
A warrant canary is a public statement somewhere on the site that you havent been served a subpeona or gag order. If you're given a gag order, you can't say the feds have gotten into your server and are now data mining, but you are allowed to remove a message saying that they haven't. So users know you're compromised if you canary suddenly disappears.


>We have not been served any secret court orders and are not under any gag orders.
it sits at the bottom of the frontpage. You can just add a page to the site like meguca.org/canary.txt that says the same thing.
>>2144353 #>>2147281 #
>>2143820 #
Isn't that for US only though? I doubt there is a low like that here/in France.
>>2143820 #
I have looked into the matter and have not found any laws for such gag orders in either France, where the server is, or here in Latvia. It seems to be a mostly Anglosphere-only kind of cuckery.

If anyone knows better, feel free to link me to the relevant laws.

>>2079192 #
Good stuff! I'll try to look it over tomorrow or Saturday. Before I do, could you add some comments before each block?
I've found that readability/maintainability is the most important thing when I'm coding.

You can also add syntax highlighting in Ghostbin (top-right dropdown).

If you're interested in C++14, here's an exercise:
See if you can use the Algorithms STL library instead of bare `for` loops http://en.cppreference.com/w/cpp/algorithm

I'm a bit rusty with those, but I think they work well with vectors too: http://en.cppreference.com/w/cpp/container/vector
If I'm remembering correctly, vectors are basically the new/improved array for C++11 onwards.
>>2079505 #
>>2079231 #
Yay! I will rewrite it a bit tomorrow during the day!
Thanks for the links anon.
>>2084414 #
I hope the STL algorithms haven't been giving you too much trouble! I forgot how advanced they were, so don't feel too bad if you were hitting roadblocks - I certainly was. Trying for that much at once (and without better learning materials) was probably overambitious of me.

Care to post whatever you've got so far, and we could discuss them, one question and one day at a time?
Question 1.3 was the simplest to implement with `<algorithm>`, followed by 1.2.

For question 1.3, I just used a single algorithm, 2 containers, and iterators for those containers.
Hint: The algorithm I used was is_permutation
Let me know if you'd like any more hints.

I also answered the questions in Python.
Python Hints: You should be able to solve them using `from collections import Counter` and list slicing syntax.
Python solutions: https://ghostbin.com/paste/cqwxd
>>2084414 #
Come back, anon, this was fun. I don't want to do this on my own.

this is now a qt thred


「hydron dev thread」

AnonymousExpandLast 100
GIqSnWaDsEx11 KB191x191proton.jpg
Hydron is a much faster Hydrus alternative backend.
Version 1.0.0 released. Basic API now complete.

README: https://github.com/bakape/hydron/blob/master/README.md
API docs: https://github.com/bakape/hydron/blob/master/docs/client_development.md
Static Windows binary: https://github.com/bakape/hydron/releases/download/v1.0.0/hydron-v1.0.0_windows_x86_64.zip

Anyone interested in developing or co-developing a GUI frontend? Any language is fine.
44 posts and 5 images omitted See All
>>2028596 #
>>2028623 #
Looks like it can only generate bindings for functions, not structs, and does not cast the types in between.
The other problem is, that because of Go garbage collection, I can not return Go pointers from a function, so all my returned types must either be on the stack or allocated from Go with malloc.
Not sure what to do. On one hand using C++ let's me design the actual GUI in Qt Creator, but it's C++ and I keep getting segfaults. On the other hand Go Qt bindings are easier to use, but I can't get them to compile hydron-qt on Windows, I can't use Qt Creator with them and build times are horrible.
>>1926561 #
If you going to start a tag war with diverged features, please don't. Helping is good.
Reimplementing Hydrus is fine, there are more new features for Hydrus on the way.
t. Hydrus user
>>2060679 #
>>2060241 #
I decided to not. Hydrus tag spec are a clusterfuck and very inefficient to scrape. I'd rather just sync tags from various boorus.

I asked about ME and PSP, not for a completely clean processor because that doesn't exist yet without stallman-tier tier dedication. And as far as the ME and PSP are concerned, there are processors available that don't have it. The T500's Core 2 Duo processor can have it disabled, and with AMD there is a cutoff point which I am trying to find.

>The last x86 processors without ME or PSP are the Piledriver Opterons and FX processors, and the Steamroller APUs and Athlons. Thankfully these are readily available new for the time being, and quite cheap with the impending release of Ryzen chips. Get hold of them while you can, I’ve just built myself an Opteron 4332 HE workstation.

To replace my desktop's i7 with a less pozzed CPU.
>>2046581 #>>2046602 #
>>2046506 #
look into me_cleaner to strip down the management engine to the bare minimum which shouldn't have remotely accessible components. if that's not enough then look into coreboot and libreboot.
>>2047064 #
What should I be aware of when buying a motherboard for it? Is UEFI safe? Any poz to watch out for?