/jp/ - Otaku Culture

>>9008035,39
It's a bugnew feature.

I had a feeling it would be like this. For those doing the work, there's nothing better than spending a few days getting paid to "fix" something they put in subtly in the first place (and probably breaking something else in the process.)

>>9027319,6
14GB including or not including thumbnails?

You don't have to store the latter in the database.

The database dumps from last year were 3GB, containing a little over 8M posts. Since then, we have gained 1 million posts. Based on this projection 12GB would be enough for over 37M posts, which is why I made that recommendation --- it seems sufficient for several more years.

>>9008035,28
>>9008035,32
"Pure" regular expressions might be OK since they're linear-time but the extra stuff like backreferences make it NP-complete.

There's no parent selector for CSS, and it was for persformance reasons. As far as I know that is not NP-complete, but it's still resource intensive to process.

tl;dr: DO NOT WANT.

>>9027319,2
Have you considered Oracle XE?

>>5061673,40
That's a backend scalability issue, and has nothing to do with the generated HTML. The solution for that was more powerful DB servers, something that actually solved a problem.

>>5061673,42
>>5061673,43
>>5061673,44
Change a few things, change them back, repeat ad infinitum. It's all useless busy work. The fact that it took so long for this rewrite, along with all these little things that were inexplicably missed, should strongly convince you so.

>That's not how a tech enthusiast/programmer should think.
I don't care what you think someone else should think, but if you were writing a 4chan extension wouldn't it be natural to want a monopoly for it? I've read about all the different ones and it's often the case that they're competing with each other for features. They don't want you to use any other, they want you to use theirs and theirs alone. It's a line of thinking quite common in the software industry.

>For example, the first thing that comes to mind is the stylesheet flicker on page load if you use a Burichan/Futaba theme, which is fixable in 10 lines of code.
I use Burichan and I have never had any of the "flicker" you describe, although I've set it up so that the page stylesheet is forced to that before it even reaches a browser --- any browser --- on my system. And YOU should be "the one to use your brain before posting" because FYI I post on 4chan almost daily. The majority is not on /jp/, and I don't use nonanonymous outside of /jp/ either.

>They have a built-in catalog view and everytime 4chan is mentioned on futaba, there always people bitching about how 4chan is unbrowsable without it
Adding that would've been a much better idea than rewriting everything else, and taken a lot less time. However, if someone is being paid to do it, then rewriting everything would be, economically, a more desirable choice than the few minutes adding a catalog feature would take. Even more so if they put in some bugs (deliberately or not) and have to take more time fixing them, introducing even more bugs, etc. [I am not omitting the possibility that if they were to add a catalog page, they would still try to overcomplicate things to achieve this effect. Something like what 2chan has needn't take more than ~100 lines of code, and likely could be done in considerably less --- My estimate is 50 but I have not done it yet so don't hold me on these nunbers.] Throw in some arguments about how doing this is ostensibly better, and the majority of the people these days wouldn't see past that veil.

>>9008035,12
If your reason is anything other than "because the structure may change" (which is what your second paragraph seems to imply), then you should probably reconsider --- 4chan didn't change its post structure for nearly TEN years, and would've passed that mark if it weren't for this recent jolt.

>All you have to do is start from document.body and dig your way to wherever you want, right?
Notice how I wrote "post structure" --- on 4chan the stuff before that does change semi-regularly and not in predictable ways, so you should use an id lookup to get to that fixed structure.

I don't get this obsession with userscripts anyway. I like(d) 4chan precisely because it isn't full of frivolous crap like most regular forums. I want to read and write posts and that's all they should be catering to here.

I only hope that this won't lead down the slippery slope to eternal minor changes for no real good reason other than to create work.

>>9008035,2
I'm not saying that, I'm saying if they're going to have a data-utc attribute then they should at least use some common sense in deciding how many times it should be present with the exact same content.

The good thing is, at least this update didn't break anything toobadly.

The bad thing is, it's still unnecessarily bloated and awkward. Whose bloody idea was it to include data-utc TWICE in the SAME damn post?!?! Putting the rightarrows in a separate div, when they should be in a span and thus on the same line as the rest of the post header, is also inexplicable.

>>8755714
Obviously I did.

In fact I worked with one of them over a decade ago. And in case you're wondering, I won't say which.

>>8982197,2
A chair of laugh does not seem to be a sturdy thing.

>>8982197
Free accounts have a 1 year retention policy. You were 2 months away from having it purged.

I don't know what you were doing but the point of RWI is precisely to avoid problems like what you're experiencing now, because we keep backups and take that responsibility for you. All you need is an internet connection and a web browser. If you do not update online then it falls on you to follow good backup policies etc. This is not to scold you but that is the reality of it.

Anyway they should be sending you the account info sometime soon.

>>8831411,66
It's easier to parse after this but there has already been a nontrivial amount of effort expended on creating the old version and by everyone who wrote applications to interface with it. No matter how much "better" this newer version is, it's essentially doing the same thing. This is a small insignificant gain in a much larger loss if you consider the whole picture.

>didn't moot mention somewhere that the 4chan Chrome extension parser is 7 times as fast?
Browsers are getting more bloated and inefficient in general, and quoting one specific case being a lot faster doesn't mean anything overall. [I could get you to try using a petaflop supercomputer cluster for a day or two and see if it makes any difference. If all you do is browse websites you will not see any difference despite the underlying system being several orders of magnitude faster than a netbook.]

Your parents probably expended a lot of effort to create YOU too. So you're somewhat indirectly insulting them if you try to insult me. Seems typical of this newer generation, I've seen it too often to care now.

But it's your future after all. You should think carefully about where you want it to go.

>>8831411,61
I have absolutely no intention to change anything on my end.

I can still read the posts and make posts.

It looks a little different but I don't care. It still works.

But deliberately making more work for everyone interfacing with the boards is quite distasteful. The only reason to do it is so that people can get paid for the work.

Meanwhile 2chan remains unchanged and stable as ever.

>>8831411,46
Are you aware that that's been a public tripcode for at least 3 years?

>>8831411,41
I did. XPCOM is not for general web page JS use. Mozilla is a huge "platform" and their browser is just one piece of it.

>>8831411,36
>Managed to finish anything more elaborate than yes(1)?
http://rechan.eu.org/misc/anoncoreutils/STATUS
Anonix is collaborative, and depends on group interest. I think a lot of that has been lost. (Ask HAHAHaruhi, I gave responsibility for it to her.)

>like this, anyone sane instantly discards what you post.
You mean anyone stupid enough to hold biases and prejudice just because they know of someone's identity. I couldn't care less about someone who falls for reductio ad Hitlerum.

>but how do you reason with someone who thinks Unicode is unnecessary and evil?
Only trying to allude to parts of what I said, out-of-context, in an attempt to strengthen your argument? I was expecting more literacy from someone like you. I never said "Unicode is unnecessary and evil". I've already posted a long analysis of it and will not repeat it hear. To summarise: there are certainly situations in which Unicode is extremely useful or even required, but in the majority of situations (1 or 2 languages) it is an extremely overengineered solution.

>>8831411,37
The only effort I am aware of for fansubbing the SICP lectures is the now-defunct sicptitles.org. If by "YouTube community" you mean Google's autotranscriber, that has...
http://i.imgur.com/9W6rQ.jpg
...less-than-optimal results.

>>8831411,38
I'm not a FF developer, but my understanding is that XPCOM stuff is not accessible from a normal web page and thus outside of the realm of normal JS. (A huge part of the browser itself is written in JS too, and that part does have access to XPCOM etc.)

>>8831411,21
http://adblockplus.org/blog/different-ways-to-force-garbage-collection
FF has two different GCs. I was talking about the JS-only one.

>>8831411,24
Works for Japanese text and ASCII, and not much else. Ideal for Japanese-only use.

>>8831411,28
Reference counting should not be considered true garbage collection. It should be considered another form of automatic memory management.

>>8831411,31
In other words you're just not experienced enough with C. You saved 10MB (let's not get into the whole "binary prefixes" thing, in working with computers G is 2^30 M is 2^20 and k is 2^10) --- which is probably 10x more memory than something written in C would use alone. There's nothing memory-intensive about parsing HTML and feeding a database.

>>8831411,33
If I were designing this I would store the configuration in the database and bootstrap it via a text file that only contains enough information for it to connect to the DB and get the rest of the config. Something like

ARCHIVER=(DESCRIPTION=(ADDRESS=(PROTOCOL=TCP)(HOST=10.0.4.15)(PORT=1521))(CONNECT_DATA=(SID=ARCHIVER)))

>>8450209,187
>When an actual em dash is unavailable—as in the ASCII character set—a double ("--") or triple hyphen-minus ("---") is used.

This is the mapping I use:
- = hyphen
-- = en-dash
--- = em-dash

>>87450209,185
And almost everyone who regularly uses TeX.
http://en.wikipedia.org/wiki/Dash#Em_dash

>In TeX, the em dash may normally be input as a triple hyphen-minus (---).

>Corpus studies indicate that em dashes are more commonly used in Russian than in English.
Ok...

>>8831411,8
>You are aware cycle collection is an implementation detail, right?
A bloody important one! You might as well be managing it manually if the GC doesn't work in all cases.
>For example, cycle collection was implemented only in Firefox 3.
[Citation Needed]
A cursory Google reveals only older versions of IE being susceptible to leaking circular references. Implementations that are not simple reference-counting would naturally be immune to it.

>>8831411,3
Your argument would be better based on fact than pure speculation, because you are wrong again.

Even JS and other languages with proper garbage collection work fine with cyclic references.

>>8831315,1
Look for memory leaks. Perl has a particularly weak memory manager that can't handle cyclic references.

http://stackoverflow.com/questions/429254/how-can-i-find-memory-leaks-in-long-running-perl-program

>>8450209,181
>>8450209,182
Let it be known that collusion, deceit, and backstabbing will be dealt with accordingly.

"Do unto others as you would have them do to you --- multiplied."