No.8005[Last50 Posts]
This is the next iteration of Kissu's feedback discussion AKA devblog. Though there's not much in the way of dev there are still bugs to fix. This iteration will hopefully focus on admin-blogging or tool development.
Important Note: If you want a new software feature (or a really any sort of change) provide me with a detailed proposal. It must be at least 3 sentences long and tell me why it's needed.
Updates:
[det]Software
Kissu's features are in a good place. There's no reason to rewrite anything, only improve and fix. On the outside, this is a unique site with an appearance you won't find anywhere else yet still your typical imageboard interface. This puts software at the software state we were sitting in 1 1/4 years ago before I started drafting a new UI except with more features that were more optimally integrated into the package.
[det]Nerd Things
People may not fully realize it, but Vichan is a dead end and their HTML and JavaScript implementations leave developers at a dead end. Not that I really endorse alternatives such as JSChan or Lynxchan which think NodeJS doesn't have similar problems as PHP... at least the UI is more moddern
Administration
Positives: No major raids or attacks on the site yet I've still been promoting Kissu. FAQ and Rules have been clarified to make it more clear to newcomers about what Kissu is about or how to use the new UI.
Negatives: Previously noted that we'd have an IRC channel, that exists(rizon#kissu) for when it needs to be used, but there are other ways to communicate that are better. Finances are what they are.
Future Tasks:
[det]Software
The tl;dr is that I just want to fix bugs at the moment. There are some software that I would like to write, but I can't justify spending the time on it since the gains will be so minimal that it's not effecting anyone except my pride/ego. Chances are the main software that I'll write are tools that support me doing specific tasks.
[det]Though it's in my head that I want to (1) Rewrite/Merge the entire backend in Rust with some other things, (2) Write an IRC server in Rust with some other things, (3) Do major refactors to the UI to make it look pretty and be easy to modify ; Chances are I won't because (1) is a waste of time since the PHP+Golang+NodeJS backend can do everything I want anyways. My spam testing show it can maintain a theoretical PPH of 300 which is faster than any other Japanese themed spinoff. (2) I don't have the means or want or appeal to cater to spammy people and give them a software, Sageru works perfectly fine for everyone and can easily have a bot written in it to do auto-moderation if need be. (3) is more likely to happen hand in hand with optimizations, but refactoring the code to an extent where everything is rewritting would create more bugs and waste time. It's not really worth it.
[/det]
Administration
Further promote the site. Try to expand reach.
Certain organizational issues are present with topics not meshing together. I think this is causing a slight loss in activity. It's a much more blurry question
No.8006
epic, code blocks bugging out.
No.8009
I fucked up in designing the markup. It's such a pain in the ass.
>>>/test/4553
No.8010
Those tab things in your post keep closing on their own. That aside it's a really neat thing to have.
No.8011
Oh, completely broken on vichan. That's kind of odd. I guess I never added that to the old UI
>>8010it must be happening on new post I guess. I'll see if that's the case after I post
No.8012
>>8011It was.
Also, matching tags on vichan is basically impossible. I'm still learning a bit of things about how vichan chose to do it's markup(poorly that is).
No.8013
Would probably be better if SJIS tagged stuff were under an image instead of how it behaves now >>>/jp/21930
No.8014
>>8013Text being put onto newlines is a general issue as well. For instance, >>>/jp/21683 >>>/jp/21681 are both 3 lines, but because the UI squishes smaller posts, the lines get broken and text pushed onto a line below.
No.8016
>>8014it's an annoying bug that I wish people told me bothered them like 5 months ago.
No.8021
never mind, it's fucked as well.
No.8022
thank me, I fixed this trash bug that 4chan hasn't even done. The implementation is ugly and the biggest hack in the software currently, so curse me.
<span aria-hidden="true" class="obscured silly-float-width-hack">
No.8024
>>8023cute, you're welcome
No.8029
Doing some other stuff at the moment so my explanation is limited, but it's been my perception for the past week+ that there are some threads scattered around that are out of place. Related to my thread >>>/poll/1554 3 weeks ago.
>>>/b/3588>>>/qa/57322>>>/qa/65289>>>/qa/70728>>>/qa/62191They'd probably do better to be focused more and placed into the season rotational boards. However, I can't really call them problematic as much as they are distorting the primary function of /b/ and /qa/. Removing them at a time when they're being used might be more disruptive than productive.
The /b/ one though, I'm pretty sure that one has to be moved. /b/ isn't a blogging board.
No.8030
what the heck the blog thread is a /qa/ tradition
also i dont see any issue with any of the other ones except maybe the /b/ one
No.8033
>>8030/b/ one was started to kinda accommodate meta thoughts that didn't deserve a thread
No.8034
>>8029Has anyone shown any distaste towards the threads being where they are? This doesn't seem like a problem and doesn't need a solution, which the poll results are pretty strongly agreeing with.
No.8035
There's a bug preventing me from moving the /b/ one at this time. The /b/ one doesn't need to exist.
The /qa/ ones as I said are a bit stranger. Functionally having them on /qa/ is illogical, but moving threads that are being used in order to help organization requires a good motivation. And I'm not too certain that moving the /qa/ ones would change anything.
>>8034Don't try to distort reality
>I kind of know what you meanimplies there's an issue and I feel it. It's my job to increase the creative output of the site through organizational changes boosting efficiency.
No.8043
>>8042Blogs are not creative endevors. They're one liners that take away from creativity.
Blog threads create friction because some always complain when someone says a word they don't like or is abbrassive and some don't care about words or tone.
A blog is the opposite of what /qa/'s objective is, which is generally more serious and it doesn't fit into /jp/'s generally more random. It's literally just posting about what you ate for dinner.
Rules give clarity about what's allowed and give people a more clear idea of what they can talk about steering conversations in an direction where everyone is onboard.
No.8044
And for the creativity question, the majority of threads get created by a select group of people. More accurately like 3 or 4, so saying idealistic stuff like "it takes away from creativity" doesn't make sense because I know they're not effected by moving blogs away. In fact, maybe people will be more inclined to make threads about their day if they're just outright deleted... but that's too extreme even for me.
No.8045
I'm not just talking about the blog thread. I'm talking in general. I don't know who makes what posts and who makes threads, everyone's just Anonymous to me. It's all just Kissu. I just don't like more rules added. Maybe it's just bad past experiences from suffering through awful moderation on 4chan for a decade, but it really turns me off. I have faith in Kissu's community to know what to post where to post. I'm also quite a laid back person who doesn't really care where a post is made and I have no problem just ignoring something I'm not interested in. But I know there's some people not like that (I don't mean this in a derogatory way), who would prefer a more clear divide between boards.
Either way, when something like this gets announced, it usually happens. I just wanted to give my opinion on it. I'll probably get over it. even if I disagree with the direction. Going back to the blog thread though, I hate to see a thread that's existed since 4/qa/ get kicked off of it because it's not serious enough.
No.8047
>>8046The wideness isn't what's concerning me. It's character alignment.
That font isn't SJIS though so hmm. I wonder if there's a font that's better for this.
No.8049
>>8048because it looks poor on linux
No.8051
k, i looked on my laptop and there's a font missing on the linux library. 'MS PGothic' obviously isn't allowed to be on linux by default.
Maybe some system specific stylesheet options need to be setup.
No.8052
actually, looking at my screenshots it's not even using any of the japanese font fallbacks. Is there a Linux japanese font...
No.8053
>>8052Mona Font seems to be the best font available for cross-platform usage.
http://monafont.sourceforge.n
et/index-e.html
No.8054
>>8053is not.
Have no fear, I'm going to go through all ~190 Ubuntu fonts and find which one works best. If I can't then I'll try and make it so the word-spacing only occurs on Linux
No.8058
>>8057I'm not having people come to my site and explaining to them that they should have installed the font before coming
No.8060
>>8059Which garbage site is that?
No.8061
Why not have SJIS tagged text be mona font and then if the font isn't installed have a notification on-screen notifying people to either install it to view the art properly or display SJIS art with the default font?
No.8062
>>8061Because I am not a lazy admin and I will find a font that works so my users are not forced to do extra work to view the site. Now quiet down let me find that font and resolve a bizare bug in CSS font-family fallbacks
No.8063
'MS PGothic', 'ＭＳ Ｐゴシック', 'Mona', 'MS PGothic AA', 'IPAMonaPGothic', 'IPA モナー Pゴシック', 'Latin Modern Sans Quotation' , 'aakar' , sans-serif
like I don't get it. 'MS PGothic' and 'ＭＳ Ｐゴシック' and denying the fallback to 'Latin Modern Sans Quotation'
No.8064
Actually I can probably just upload MS-PGothic.ttf to the server if it doesn't violate the copyright of that font and set it up in the CSS
No.8065
>>8064That's literally what I was trying to tell you about Mona...... especially considering it's public domain...
No.8066
>>8065You were not. You pointed out examples of sites begging their users to install fonts to use the features on it.
No.8067
>>8043I have no frucking idea what you're talking about.
No.8068
downloadable font: OS/2: Bad version number, setting it to 1: 2 (font-family: "MonaT" style:normal weight:400 stretch:100 src index:0) source: https://kissu.moe/webfonts/monaT.ttf
downloadable font: hmtx: Failed to read side bearing 7225 (font-family: "MonaT" style:normal weight:400 stretch:100 src index:0) source: https://kissu.moe/webfonts/monaT.ttf
downloadable font: hmtx: Failed to parse table (font-family: "MonaT" style:normal weight:400 stretch:100 src index:0) source: https://kissu.moe/webfonts/monaT.ttf
downloadable font: rejected by sanitizer (font-family: "MonaT" style:normal weight:400 stretch:100 src index:0) source: https://kissu.moe/webfonts/monaT.ttf
I love fonts
I love fonts
I love fonts
No.8069
Mona added to the site's files, it better not be a copyrighted font.
>>8059Which site is this. I want to call the admin a piece of shit.
No.8071
>>8066That was someone else.
No.8074
>>8069>it better not be a copyrighted font.The sourceforge project page says it's licensed under public domain, as does the Wikipedia page on CJK fonts. Moreover, the site itself lists acknowledgements for fixes, and credits the font it is based on, Shinonome Font, which is also licensed under public domain. Furthermore, the last update for the project was 7 years ago, uploading old files, and the last update to the font itself was released nearly 17 years ago.
If it's copyrighted, I'll gladly eat a sock, because all signs point to it being public domain.
No.8075
>>8074it's fine, Japan doesn't even protect their anime from piracy
No.8076
Going with the previous posts about blogs and things that vaguely resemble generals I've put into place a rule on /jp/
- The board will have no threads designed to contain thoughts and ideas
This is to say that a thread that contains thoughts and ideas is not a problem, but one with an OP that is crafted as containment of things that could be threads has to be dealt with.
No.8077
>>8076This rule is not one that will effect existing threads, but an assurance that /jp/ won't contain general like threads
No.8078
Just to recap, these are the existing guidelines for the three main discussion boards.
/qa/ - Questions & Answers
- A metaboard with an anime twist
- Posters should try to be mature when discussing content
- Longer winded discussions about anime or imageboard meta are preferred
/jp/ - 2D/Random
- Content must relate to japan through either text or images
- Explicitly sexual content is not allowed, but better tolerated if spoilered
- The board will have no threads designed to contain thoughts and ideas
/sum/, /spg/, /aut/, /win/ - Season Boards
- All topics are permitted.
- Boards will cycle every season.
- Even though it's random, try to be polite.
No.8079
>>8078>The board will have no threads designed to contain thoughts and ideasThis is rather vague, I must point out. A better, but maybe not yet perfect way of putting it might be, "The board will have no threads expressly designed for discussion on specific topics"
I don't particularly like this change, but regardless...
No.8080
>>8079I'll think about the reword. Otherwise if you were planning on making a pseudo general in /jp/ then why would you even post there? I think you're mistaken about what /jp/ is about and your blind hatred of rules is preventing you from seeing this as a good thing.
No.8081
Also ˜ got converted into "˜" on a few of the posts with SJIS for some reason.
No.8082
>>8081because tilde were set wrong and didn't think a change would effect it all. give the site 10min to fix itself
No.8083
>A metaboard with an anime twist
>The board will have no threads designed to contain thoughts and ideas
These guidelines don't make sense, you should clarify what is meant by "metaboard", and the second guideline is poorly worded, as already pointed out.
No.8084
>>8080>blind hatred of rulesI said, "
this change", might I point out. I don't have much to say, nor do I want to, so I'll point to my original thought on this matter (>>>/poll/225), with some addendums of course.
>so long as the only rule is "don't be obnoxious," there's going to be an ongoing tension between what is and what isn't allowed.To this point, I think it's fairly obvious this was the case, otherwise there wouldn't have been any rules or guidelines added. Simply, it is my concern that the inverse could also be undesirable. That the literal understanding of this rule, would impede creative threads, such as >>>/jp/21680, and if taken particularly literally would result in threads that evolve into more general discussion such as >>>/jp/20137 being moved to another board. If the intent is to prevent threads such as >>>/jp/21511, then it should be worded as such. Frankly, I am just unconvinced that a rule is necessary, and to what extent it will be enforced. I am tired and in varying amounts of pain, so I will say nothing more on the topic, especially since, I think, posts such as this are better reserved for dedicated meta threads.
No.8085
>>8084The rules written are what I think works, and as I see more things I have a better idea of what is true. If it's unclear that's because putting my thoughts into words is a pain.
If people make container threads everywhere that will be a problem.
No.8086
The threads you linked, 1 was moderated and the other was crosslinked from /qa/(I would rather it was there). The other irrelevant(actually I probably would be incline to move depending on how long it lasts. it fits with qa threads better). Though they're blurry cases sometimes this is the how rules are and why courts exist. Unless stated as a delete rule(typically all global rules), they aim to clarify purpose to get everyone unified and if written well justify things not being deleted but moved. I think you believe a rule to be a justification for the death penalty and fear (would impede creative threads,) and would rather I go back to a singular "don't be obnoxious" rule. This is why I say you don't like rules
No.8087
frankly I could cop out and use the meaningless word general thread to describe it and ride on the oldfag ideals that X is bad and Y is good, but these meaningless bandwagons aren't healthy and I'd rather describe what I think a general is.
No.8088
There is a very clear reason why a lot of people dislike "general threads", and you seem to be misunderstanding it on a fundamental level. Condensing threads to avoid spreading discussion across multiple threads that talk about more or less the same thing is not and has never been a bad thing. The problem with general threads is that they tend to attract people from outside the board who only go there for one specific thread, creating a new community separate from the board's community. None of the threads you are trying to "fix" present this problem - the only thread on the entire site that does is the happenings thread, which ironically will never be dealt with.
No.8089
>>8088This rant is why I don't want to bring it up.
No.8090
>>8088>There is a very clear reason why a lot of people dislike "general threads", and you seem to be misunderstanding it on a fundamental level.I'm not trying to appeal to emotions and populism. Irrelevant.
>Condensing threads to avoid spreading discussion across multiple threads that talk about more or less the same thing is not and has never been a bad thing. Debatable
>>8088>The problem with general threads is that they tend to attract people from outside the board who only go there for one specific thread, creating a new community separate from the board's community.Having more users is good. You have a criminal misunderstanding of how things work and why a blog thread is not exactly this.
>None of the threads you are trying to "fix" present this problem - the only thread on the entire site that does is the happenings thread, which ironically will never be dealt with.The rule prevents things from being posted on jp. Read my posts before you latch onto the single word you understand.
No.8091
>>8090I don't have the will or the energy to argue with you. I already made my point, and if you're not interested in understanding what people mean when they talk about "general threads", well, that's too bad.
No.8092
>>8091I don't have the patience to understand people who base their arguments around new people being bad because they use a thread more than others.
The idea is a differentiation between qa and jp. You don't understand this and want me to talk about your irrelevant feel good crap.
No.8094
unfortunately the best place to argue with vermin isn't on-site in his thread about development since he brings a billion different prejudices and presuppositions to the discussion. if you really want the most headache free experience I'd recommend #kissu on rizon
No.8095
If you want fairness then yes a named IRC is better.
Even if you use any platform to discuss meta with me, know that I know you are a cretin who is trying to use my resources to control others rather than create a more productive environment.
No.8096
Anyways, the rules are just something which gives people the right to arbitrarily delete and report things that are low quality.
I'm not going to remove a thread about pizza because it's not 2d/random
No.8101
Made a change so SJIS will render better on mobile and smaller resolutions.
made some simple bug/css fixes
No.8105
Issued a token via direct message.
No.8110
link shortener dictionary expanded to contain 1296 blocks with possibility of false positives.
This expands the existing filter list by ~200%
No.8111
considering options for an OCR that activates under certain conditions
No.8114
not very descriptive error message
No.8115
i think the hypens are funny
ｷﾀ━━━(ﾟ∀ﾟ)━━━!!
No.8117
>>8115Aaa, I see it now... It's a windows thing
No.8118
huh it's actually even stranger... it only occurs on windows+FF
No.8119
That's so hecking weird. Windows Firefox is the only browser I know that this, but ━ is not consistent in the same enviornment
if ━━━ is done with an ASCII character in front of it then it behaves like a small dash,
but if you prefix it with a UTF-8 character it becomes a heavy dash ﾀ━━━
No.8120
I altered the Kita to use light dashes instead of heavy, but the process of fixing it for existing posts isn't automatic and too much of a pain.
New posts should work fine
No.8121
Why do we not have (You)?
No.8123
>>8122Uhh I meant the "Anonymous (You)" in the name.
No.8125
>>8123I have no idea what that feature is,
here I implemented a feature that sounded like what you wanted anyways. v4.5.2
No.8127
>>81264chanX put a red border around owned posts, but not that.
No.8129
>>8127Don't use 4chanx. Maybe I should...
>>8128Yeah.
No.8130
>>8129it doesn't work in this UI
No.8132
Put (You) citelinks as an option
Added ownership signifiers to owned posts
Kissu-Fr 4.5.4
No.8134
How do I disable (You)?
No.8135
>>8134unless there's good reason, custom CSS
No.8137
>>8134Custom CSS
I don't feel like adding that feature as yet because it will involve making more options tabs.
.owned{
display:none;
} No.8139
>>8138yes i agree with that
No.8140
k, it's adjusted.
If people want a different font-size or character they have to present examples. I'm kind of meh about owned-post tracking features, but they're in there anyways because some like it.
No.8141
yeah, I'm not adding any more features because I have zero motivation to perfect owned posts tracking and it's introducing messy things into the software/beginning to show off the flaws of my code and backend customizations.
Noticed that owned-post tracking doesn't work with noko active which is a problem that is a pain to solve because confirmations of successful post occur after the post has been sent into the system and this will require some messy or ingenious idea that consumes brain cycles that can be spent elsewhere(Trying to boost site activity/creativity/spam resiliance).
No.8142
Alright, I fixed the feature, but please stop recommending new stuff.
If you want something then please give me a detailed proposal.
No.8143
The OP has been updated with the following:
Important Note: If you want a new software feature (or a really any sort of change) provide me with a detailed proposal. It must be at least 3 sentences long and tell me why it's needed.
No.8145
>>8144Bug fixes are different from features. bugs have to get fixed.
No.8147
>>8146>word/wordsI forgot to escape a . character in a regexp
No.8149
The only tough thing that's getting held up now is the image-text reader to automatically mitigate some of that spam when there's no mod activity.
Otherwise another batch of fixes done.
No.8150
For ~48 hours I've added
https://www.w3counter.com to the site to gauge some statistics about bots vs live clients and some other details. I want to see if having this information leads to constructive decisions for the site, but don't want to commit to anything yet.
It autoblocks if you have ublock which I assume most people have. AWIOS has no negative press.
I'm looking into some other self hosted options or most extreme case writing something myself for the kissu's specific use case if I like the idea. possibly something like this
https://umami.is/ No.8151
Removed 3rd party analtyics. Proved useful in visualizing traffic, but it seems like 100% of people here use ad block so it was showing only referrals and search engine crawlers.
Trying to get Umami to work(it's a bit of a pain in the ass with systemd) but it's pretty nice visualizations. It allows for public statistics pages so I'll put those URLs possibly in the thread OP for statistics nerds.
No.8153
>>8152I'm thinking these pages might make the site break GDPR Compliance or put it into a grey area... I'll give it some thought but these pages will probably go down unless I have a second thought.
No.8154
I get a notification even when I myself reply to my own post.
No.8155
Worked out some details.
Kissu's going to use .webp on both UIs with fallbacks to low quality(25%) .jpg
New apple devices have webp and webm support, but the old ones don't. There's one user on this site who uses an old iPhone so I have to account for that.
This will mean that while the server has duplicate thumbnails(~5kb more per file), the .webp change will mean that users load less speeding up content loading.
In order: png(existing), jpg(iOS fallback), webp(new)
No.8157
>>8156Doesn't mean anything as yet because I'm just saying that it's something in the works.
Simply I'm trying to make images load faster.
No.8158
oh
No.8159
Just doing a quick look at some of these files it looks as if WEBP offers same quality thumbnails than PNG with a 50%-25% better filesize.
No.8160
Kissu must be one of the first (if not the first) imageboard to use .webp thumbnails
https://haiji.kissu.moe/jp/thumb/1629868332617.webphttps://haiji.kissu.moe/jp/src/thumbtest.pngThat's a 71% reduction and there are no compression artifacts nor does the alpha channel get discarded.
Using the saving on the harddrive I added a fallback .jpg fileformat at very low quality for Apple's slowness to adopt new standards
No.8162
I wish it were easier to use citation chains.
No.8163
>>8162the only problem I have with it is that not everyone cites when they reply to people
No.8164
if you can clarify I'd be glad to help, but I don't just know what your problems are as soon as you press down keys on your board.
Found a mistake in how I wanted to serve JSON files, might have been causing some obscure and hard to replicate issues. Think the node.js server might be a bit malconfigured as well, but screw nodejs
No.8165
Not sure if anyone else has been having the same problem, but every now and then the quick reply box fails to show up for some reason.
No.8166
>>8165where do you typically place the QR box?
No.8167
It's notable that if it's in the corner of the screen, then you resize the window it won't correct, but after refresh it will.
This can be fixed.
No.8169
>>8168A similar experience I've had is
>>8167 so I guess it's related.
But I haven't experienced or had anyone report anything that suggests a crash/error in displaying the QR.
Are you leaving out an important bit of info, namely that I think you're using a non mainstream browser...
No.8170
>>8169>crash/error in displaying the QRI'm not really sure what's happening. What typically happens is that the quick reply box simply doesn't show up when clicking on a post to reply to it, although I think the reply window at the bottom of the page does update. It typically happens on the index, and then carries over when going into a thread. But, once in a thread, the quick reply box can be reset by clicking the "Reply" button off to the side, whereupon it'll just show up and fix itself.
It's just waterfox. I'm not sure I'd consider it non-mainstream insofar as it still being a fork of a major browser.
No.8171
>>8170It's an important detail. Browsers that are not chrome/firefox are not designed to work on modern websites.
No.8172
>>8170OK, I think I actually understand what might be it. Not browser related.
You probably have the footer-form focused which by design prevents the QR from being opened.
No.8173
but then again, if you say it happens on index that doesn't really make sense.
No.8174
potentially the activeElement field doesn't behave the way I expect it to in that browser. In any case, I'll disable that and adjust the QR reposition to see if it fixes it
No.8175
Alright, I put in some adjustments to how the QR behaves.
Also realized that select quoting wasn't working on footer-form
No.8176
>>8172>You probably have the footer-form focusedThis reminded me, very rarely I'll refresh a page -- typically after opening up my browser -- only for the page to seemingly fail to load, all but the footer. This fixes itself by refreshing the page again, but I figure I might as well mention it happening, at least on my end.
I might be misremembering, but I'm pretty sure I've had this happen on both my PC and phone - waterfox and Firefox mobile.
No.8177
Are you saying that you also didn't get a notice mentioning the other UI?
No.8178
>>8177Uh. Well, I've only ever noticed that happen when on threads or the index and then refreshing, so I'd already have gone past the homepage to see that notification. So... "Technically", I wouldn't see the notice, but I wouldn't see it anyway...
No.8179
>>8178I see, you don't know what I mean so that helps a bit
No.8190
I restartedd the site for a thing
No.8192
I added a colored draft to the startup window(+News) and new-draft favicon. Old ui uses what's existed in the past because I figure those of you on it don't want new things
No.8193
hoping that these changes give a better first impression when someone starts the site
No.8195
New icon looks pretty decent after edits. I wasn't happy with how it looked on my laptop, but desktop seems pretty nice.
Site has a dark theme on mobile for applicable browsers to make it fit better with the default theme.
Just waiting for more bugs. I think the QR form needs to be a bit less tall on mobile because the keyboard comes up far too high when the post-queue is stacked on top of it.
No.8206
I may have to make a ven-diagram to explain the boards because I still think it is a valid issue that the organization isn't very clear from the outside looking in.
/qa/, /jp/, /sum/, /cry/, /b/, and /ec/ (all of them) have overlapping content and while I don't think that a diagram will help too many people(maybe make things worse) it will help me to visualize an issue that a new poster might feel.
No.8209
So ridiculous that iphones auto zoom into input fields
https://stackoverflow.com/questions/2989263/disable-auto-zoom-in-input-text-tag-safari-on-iphone#answer-46254706that's contrary to any other web design paradigm and bad UX. Taking control of the page away from the users who should be in control and developers if they don't account for it..
No.8468
I'm looking at CSS and behaviour improvements for the catalog. Things such as making the
R: and B: sections more compact, placing more threads into a catalogue line, and fixing some interactivity bugs or potentially revising how expanding a catalog thread works.
The end goal is to bring it in line with how 4chanX's large setting looks (
https://imgur.com/tJQ72CO.png )
Does anyone have any opinions on how it works right now? You click on the container of a catalog post to expand it and when you mouse out it will deselect it.
No.8470
>>8469ok, that's rather strange. I went to bed around that time so I would have seen anything before hand...( thread limits go down to 5 amoung other thing when theres no mod activity in the hour )
No.8472
I see that senran poster made a bunch of replies. That shouldn't be triggering the thread filters though, but I imagine it caused it. There's nothing in my logs that suggests much else.
These filters are board specific.
No.8474
This is proving hard to replicate...
No.8480
Alright. I found the problem. Shouldn't happen now
No.8482
Catalog updated.
No.8490
>>8488Also added in a bold font to the sidebar to try and encourage /all/ to the main board new posters use
No.8492
If anyone thinks the bold /all/ - Overboard is too distracting or obnoxious I have a solution to make it better, but if no one comments then I won't do it.
No.8493
Having it be bold slightly confuses my brain. Normally when there's a list of things and one goes bold it's to indicate that that's the currently selected option.
No.8494
>>8493Think I agree with this, kinda defies common logic.
No.8495
>>8493hm, that's actually a good point. It might be better to even have bold be the board that you're on, but it messes with my idea of how to resolve the issue.
No.8496
>>8495I think anyone that uses the site long enough should generally figure out /all/ is the main board. Maybe having some indication on the homepage would work?
No.8497
>>8496Why do you expect that people would use my shitty site for longer than 10 seconds if I don't hand hold them?
No.8499
the bold text on the sidebar for current board might actually be more ugly than I anticipated
No.8501
Will kissmin ever add iqdb support? And no, I'm not using browser add-ons because I'm not a nerd.
No.8502
>>8501iqdb was dead at the time.
you can add it with custom.
but i probably will configure it today.
No.8503
thinking about solar shielding backups for kissu... sounds expensive as heck though
No.8504
It seems like insuring data would cost 100 dollars on I think... But solar activity induces electricity in electronics while an emp does something else... pain
No.8505
Actually... could back kissu and source code on Optical Disks(bluray) for much less... magnetic activity doesn't effect these since the data is burned onto it and the amount of magnetic energy required to damage these would kill a person.
Write the site's data onto 9 discs a year.
Probably buy new ones every 5 years and then the site would be fully archived.
Optical disks currently go up to 50GB in size so if the site's storage were 12x larger it would still be possible.
Flash drives are another option apparently.
https://www.newegg.ca/verbatim-16x-4-7gb-dvd-r/p/N82E16817507004?Description=blank%20dvd%20discs&cm_re=blank_dvd%20discs-_-17-507-004-_-Producthttps://www.newegg.ca/verbatim-6x-50gb-bd-r-dl/p/N82E16817130189?Description=50gb&cm_re=50gb-_-17-130-189-_-Producthttps://briantomasik.com/backing-data-geomagnetic-storms-emps/https://www.canada.ca/en/conservation-institute/services/conservation-preservation-publications/canadian-conservation-institute-notes/longevity-recordable-cds-dvds.html No.8506
k, bought myself a bunch of DVD+R
Taiwan manufactured, Anywhere from 20-50 year of lifespan according to the conservation-institute of Canada
https://www.newegg.ca/verbatim-16x-4-7-gb-dvd-r/p/N82E16817130992?Item=N82E16817130992In addition to the automatic backups of SQL and thumbnails that get sent to the luna server, my sporadic backups onto local HDD storage devices, I'll do a yearly backup onto read only DVDs.
No.8507
>>8501IQDB added. IQDB doesn't support .webp though and the .jpg fallback on this site doesn't exist on every image.
So IQDB works on anything posted after Aug 27th
No.8508
>>8506>20-50 year of lifespanisn't that for pressed discs only? i've heard that recordable discs have much less lifespan
No.8509
>>8508Burnable media uses triggerable dyes inside of holes. Overtime the dyes react with materials and lose their readability. Silver and gold layers over the holes prevent the rate that this happens.
Pressed is stamped so it should last much longer. I dunno if it uses dyes or not. I think not because it's hard to find information on when pressed CDs go bad. Maybe they never deteriorate in ideal storage.
The max lifespan of a DVD+-R probably assumes that the dye never loses it's readability. In best storage conditions some people say DVDs can last 200 years...
No.8510
of course this is very much a matter of calculation, estimates, projections and averages. The CD-R format has been around for 33 years. My source says that these should last
20-50 Years so it seems like this number is probably accurate and verified from actual examples.
No.8512
>>8506wow, a 24 hour delivery for standard price. love newegg
No.8514
May have replicated the issue with the QR not appearing
Seems to be an error I got which did this
No.8515
More specifically relating to selections and parent nodes.
In other news,
September 11, 2021 has been recorded to 9 DVDs
No.8516
>>8110I had a false positive show up (don't know exactly why but it is an page with a bit of an ugly looking URL so I imagine that has something to do with it) although I was able to get around it by posting an archived copy. Can't post the link here for obvious reasons but it was the original URL of this page:
https://archive.is/RONf.
No.8517
http://pespmc1.vub.ac.be/GBRAIFAQ.html
No.8518
forgot i have bipasses to spam filters... time to start up private tab
No.8519
Err... what did you just do? The site changed back to the default theme by itself and I can't find the button to switch it, the toggle-new button doesn't work and the banners page is dead.
No.8521
does uploading files to the server cause this sort of thing to happen...
No.8522
Anyways, i resolved this false+
No.8523
>>8520oh I think you might be omitting some information... was this on the happenings thread? Happenings thread has a unique way that the UI works.
No.8524
yeah, so you switched to the old UI from /qa/thread/4165 which changed your URL to /qa/res/4165, but 4165 has a unique rule that it ignores your UI cookie(for archiving) so it stuck to the old UI and then when you changed page it started working again
No.8525
>>8524I've resolved what I assume to be your issue
No.8526
Similar threads, found at the bottom of thread pages, rotate on every refresh now.
Previously it was showing the same threads with lots of replies, now it will show less responded ones as well due to a random shuffle function. The algorithm will still prioritize larger threads however.
No.8527
>>8523Nah, the layout on the whole site looked to have reverted to default software settings with all custom elements (themes, options menu, new UI, banners page, etc.) completely missing. Changing or refreshing the page didn't do anything. I didn't alter any settings and it revert to normal by itself after five minutes or so.
No.8561
uh, not sure why it wasn't thought of sooner, but if youtube videos are basically just a thumbnail instead of an embed before being clicked, why can't you spoiler them? like, i'm sure you could spoiler them through the mod interface or whatever, but if you click the spoiler button icon thing by the submit button, it doesn't actually spoiler youtube videos.
No.8562
>>8561problem is brenan or czack decided that it would be a better idea to put embedded data into it's own seperate SQL column.
In layman terms the systems that control files are not the same as the systems that control embeds. This meaning that adding this feature requires reworking the database design.
No.8563
Beginning the draft process for a Rustlang core designed for Kissu. Will be an API that can have bindings with Python/Lua/etc scripts for a speedy performance and ease of application design. Personally built Twig template, MySQL parsing engine(I don't trust any existing Rust packages/libraries).
Will use preexisting security tools for logins and MySQL server because the people who design these are better at this than I am.
This core going to be the main component of all server projects going forward (some statistics features, removal of vichan) so it will be able to handle all the activities of imageboards in an abstract way that allows for scripting freedom.
Core will be called Amaterasu, Kissu applications Tsukiyomi and statistics program Susanoo.
I intend to paywall statistics and improve payment processing.
This hypothetical idea acts as a combination for hazuki-Golang(API server, proxy banner and AI), Kissu-NodeJS(New UI starter), Vichan/NPFChan/Tinyboard(Post server), Vichan-UI-generation(Old-UI page generation).
This is the start of a probably 2+ year long(on and off development) idea to create an immortal server designed for Kissu that allows the site to handle any level of traffic and allow admins to operate all the functionality without shell access.
No UI changes will occur as a result of this, on the mod side or user side. The old Vichan templates will continue to work on it.
No.8568
>>8567I thought I fixed 2
No.8570
2 is probably the result of having incognito tabs open and such, since I can't replicate it.
I suppose I can put in a condition to check for your own post replies, but I don't consider it a major issue.
No.8571
K, I put in a fix for #1.
No.8572
>>8570I don't have any incognito tabs open though. One thing I forgot to mention is that the two notificaions don't come at the same time, there's a gap of half a minute or so between one and the other.
>>8571Thanks.
No.8580
>>8572>One thing I forgot to mention isThis sounds more familiar, but I haven't noticed it. Still think I fixed it a while back and there's some privacy setting that's preventing local storage being shared between tabs, but I'll have a look at it when motivation returns.
No.8582
fixed thanks
No.8640
Maybe I'm just blind, but I keep trying to upload pic related to
>>>/ec/3586 only to get the message "File already exists!" preventing me from posting. I've looked over the thread a few times and don't see it, and searching the filename finds no results so I'm not sure what's up.
No.8641
>>8640it's vichan idiocy, you can't post the image if there is an identical one anywhere on the same board.
the warning should link you to where it is.
No.8642
huh hmmm. let me try
No.8643
>>8641Oh, yeah. I see what you mean. How annoying... I know 4chan does that, but isn't it for stuff in the same thread?...
No.8644
It's a setting that I decided might be a good idea. It makes people look around the board for threads they may have missed. But in some cases it does seem like an inconvinience. I'm open to opinions on it.
No.8645
i've enabled it in thread only. try now and see if it works
No.8646
>>8644I can understand preventing the same image being posted in the same thread, but preventing posting of the same image across an entire board seems a little overboard.
Also, I think I should mention that I was pretty confused what was going on since the notification had disappeared a few times in a row of me trying to post before I even noticed that there was a notification saying what was happening, and a few more times of trying so that the notification would appear so that I could read the whole thing. In my opinion errors such as this where posting is being prevented, the error message should be non-disappearing so that the user can actually read what's happening.
>>8645It works, but upon further investigation I decided the image wasn't a good fit for the thread after all. Whoops.
No.8647
>>8646Makes sense. I'll add it to my list of changes
No.8651
>>8650that gives me 10 thus reported bugs, so I guess I'll fix them now
No.8662
>>8582Fixed 10 bugs I found and were reported after this point.
theatre.kissu.moe update for a not-notable potential bug. Also updated from the fork.
No.8676
.txt is an "unknown filetype".
No.8680
>>8678
var min = Math.floor(sec / 60);
if(min <= 60){
time_string = min+" m" + (long_names ? "inute" + (
min > 1 ? "s" : ""
) : "");
}
I noticed this issue with on 7days vs 1week, but didn't apply it to the rest
No.8751
On Vichan imageboards(probably lynx too), the report feature exists not for reporting illegal content, but things the community doesn't like.
This is a bit of a fault in the system. There needs to be an illegal report option, so I'm going to see about adding this in the next 24 hours.
Reporting will in the future have an illegal option that makes mods get notified of illegal content faster.
No.8755
Updated site to 4.10.0
Major change is how reports are handled, both on client and backend, to give faster responses from mods.
No.8773
>>8772why isn't that on by default?...
No.8774
>>8773The dragable feature is too foreign to people using conventional imageboards
Dragability is more of a unique usage which can make the general usage, watching content, more complicated.
It covers posts below the expanded making it harder to read newer replies.
Don't think it looks as nice as pushing text over
No.8795
Something rather minor but after clicking the "Reply" button on a thread (new UI), the popped up quick reply box is not automatically in focus.
No.8796
>>8795I think that's a pretty big usability issue that I've gotten used to.
No.8797
>>8796I've just never used the "Reply" button so I noticed when I used it for the first time. I always click on a post number to get the quick reply box. But I just realized that I had gotten so used to it that I didn't even notice I still have to manually focus the quick reply box.
No.8798
To be fair, on imageboards, the "reply" on threads more accurately means "view whole thread" than it does "reply to thread".
No.8799
He specified reply button, different from the reply link or even just reply.
Though if someone wanted to be technical they could say the QR button, but it's not a technical discussion, but a simple suggestion.
No.8800
>>8799oh. I'm not sure what they mean then because clicking the Reply button pretty clearly does bring up the QR box. I think I'm missing something here....
No.8804
>>8800I'm not saying it didn't bring up the QR box but that the QR box that appears after clicking on the "Reply" button on the side panel is not automatically in focus. "Automatically in focus" meaning user has to manually click on the QR box that appeared to start typing anything.
No.8821
fixed the focus issue
Fixed a few other issues, but mostly changed how the flash player works.
No.8828
Is there an image size limit? I can't post 6400x6400 image because of error: "Failed to resize image!C Details: Killed"
No.8829
>>8828Means system ran out of memory, so yes. I might need to impose one.
>>8827probably not. What I imagine happened is you couldn't download a certain json file in time so it gave you a 404 page. This might be a bit intuitive. Also debugging phone problems is a major pain.
No.8831
Sachiko Birthday 2021 CSS Art Attributions:
Cursor: Official Signature
Background "Mascot":
https://www.pixiv.net/en/users/174742Disembodied Reply Head:
https://www.pixiv.net/en/users/1657510Falling Sachiko:
https://twitter.com/frenchmaid_ No.8833
/b/'s catalog hasn't been updated since 01/19/21, is it normal?
No.8834
>>8833What do you mean exactly? You're not seeing newer threads since then in catalog?
I had that happen to me, but clearing my cache for kissu.moe cleared it up. Hmm...
No.8835
>>8834Newer threads simply didn't display in the catalog, and it only happens on /b/ not any other boards.
I tried opening on another computer and initially all threads appeared, but after I changed the "Sort by" option, the exact same thing happened, and the "Sort by" option doesn't actually take any effect.
When I refreshed the catalog page, I saw the new threads flashed for an instant but then immediately disappeared. So I think they are actually there but somehow something made them hidden.
No.8836
>>8835ok... that's strange
Also I assume you mean the older UI
No.8837
>>8836Right, new UI doesn't have the problem here.
No.8838
>>8837well, admin #2 got the issue yesterday so I have some ideas on why, he uses Waterfox and assume your browser is more standard
No.8839
>>8838Tried on Chrome, Firefox, Pale Moon, all have the same problem.
No.8840
The error is a bit misleading. I think I replicated it by changing the sort order.
You have a sorting order selected and something about sorting is causing it to crash. If you go into the dev console you should see: Uncaught DOMException: Node.removeChild: The node to be removed is not a child of this node implying something in Vichan's catalog.
No.8841
Got it,
>>>/b/5912Due to various changes in how markup is handled, some pages have broken HTML from mod edits which causes catalog sort to crash.
Also am going to add debug errors to mobile mode so I can better figure out some issues on that.
No.8842
Sachiko's birthday is over in all timezones ;_;
If you wish to keep it active longer (or forever?) here is the css code thingie for old UI/vichan:
https://github.com/ECHibiki/Kissu-Files/blob/master/sachiko.custom.cssBut, you will need to save the images from the /test/ sticky and upload and link them yourself in the future since they won't be there forever.
No.8845
pencil and ruler, it means you're trying to sketch up your post and measure it before posting.
No.8846
ohhh, makes sense I guess
No.8847
looks as if making threads didn't work while I was sleeping
No.8848
Birthday CSS is added. It's quite well developed that it will end up being part of both Kissu and Vichan stylesheets when I get around to adding it.
No.8849
>>8848So you're going to make it a permanent addition? Nice. I think this theme rocks, and I'll probably have it as my 'default' from now on.
No.8850
>>8849yeah. I'll just have to get around to making some small modifications.
as usual all the special themes are placed on this repository for your own addition:
https://github.com/ECHibiki/Kissu-Files/blob/master/birthday.custom.css No.8851
I should add in the attribution that this was made by cool and I just did alterations to make it work on the site.
No.8857
>>8856Oops didn't mean to sage
No.8862
kissumass.css theme needs to be adopted to the kissu-ui. Also some javascripting to make it default when someone is on /xmas/ yet allow them to toggle it off if they request.
No.8863
Also an issue where the non .webp fallbacks aren't being used when someone closes a thumbnail
No.8864
>>8856Now that "Nekomimi" CSS is the default (for me at least), will there be a new CSS the next anniversary? Or will nekomimi CSS just be mandated for that specific time frame?
No.8867
>>8864I think it would be better for us to add more things to the day(birthday hats for example) until it's got too much stuff, but one year is too far away for us to start planning.
If /xmas/ is any example we kept the same theme and are expanding it's functionality(enable and disable, adopt it for the kissu-ui)
No.8868
/xmas/ theme added to the kissu-ui page. Designed to only appear on /xmas/ and be disabled if you don't like it.
kissumas.css won't be added to the kissu-ui rotation
No.8869
>>8864Well, I do CSS stuff because it's fun to mess with so there will probably be a new one. But yeah, planning a year ahead is definitely not something I ever do. It's kind of unfortunate that the big stuff is all at the end of the year, barring any one-time events. Halloween (was too lazy to do it this year), Sachiko, Kissu Birthday, Christmas, New Year's. I thought about doing something for that one German holiday that's mentioned in Madoka, the uhh.. Whataburger Eve or something that's in July or something.
No.8870
>>8869>or something that's in July or something.I think I have a new year's resolution. I need to stop doing this
No.8871
>>8870Stop what? "something something"?
No.8872
>>8871Yeah, I just say "something" too much. Four times in that little post for example.
No.8873
>>8872Now that you mention it... Though the first two "something"s can't be replaced with anything else I think. The third "something" too. The last one could be replaced with "I think" or something of similar nuance. I'm curious though, how would you rephrase that post following your new year's resolution?
No.8874
>>8873Not sure, but I don't want to derail the thread any more. I guess I could answer in a seemingly related /qa/ thread soon
No.8884
>>>/aut/1195Why was this thread moved?
The discussion is civil, it is a theoretical discussion on human behavior and has nothing to do with news or politics. >>>/qa/81421 also discusses human behavior and social media but that is apparently treated differently.
No.8885
was political in nature or in a direction that could focus on political theory. One of /qa/'s topics is internet culture. Discussion about social/political organizations doesn't fall into that category.
No.8886
>>8885>was political in nature or in a direction that could focus on political theory.Ponder this: your view that the thread
could become X, can implicitly bias it into becoming as much. Pre-emptive moderation as such may in practice then lead to what you are trying to avoid.
No.8887
>>8886I'm not here to ponder, I'm here to act.
No.8888
>>8887Then act when it is necessary to do so. Discussing history and anthropology never used to be treated this way.
No.8889
>>8888Ah yes, I remember the good old days when kissu had political commentary that I didn't shit on.
No.8890
>>8889If your view of what falls under the purview of politics is so broad it includes, "I read some stories about mafias and gangs around the world", then everything is politics, and you might as well move anything and everything to /aut/.
No.8892
>>8885>in a direction that could focus on political theoryDo you have any clear guideline of this? Cause any discussion of history or any act of person, even in a pure theoretical context, could "lead to that direction", even though no one apart from you think so.
Might as well just plainly say "Discussing any real or hypothetical person other then yourself is banned".
No.8893
>>8892If you are so interested in reading what I've written about the site rules, then you should first off have read the section on the rules page that says that board guidelines and organization doesn't result in bans.
Although I haven't put much effort into the rules page recently since it's obvious that the people who complain about rules don't even read the page in the first place.
No.8894
>>8893Then the contention at hand is that, for /qa/:
>A board for imageboard meta, anime discussion and generic bloggingIs too strict, particularly considering that, frankly speaking, the vast majority of topics discussed on /qa/ already do not fit these criteria.
No.8896
I need to think for a bit about how exactly to convey my thoughts.
That thread discussed autocracy and Mafia which is an explicitly political topic and the unwritten rule is that politics goes to seasons
No.8897
Tapping the watch icon causes something or other to crash:
TypeError: this.props.summaryPosts is undefined -- {"componentStack":"\nSummaryWindow@https://kissu.moe/js-ui/ui-bundle(4.11.8-debug).js:1:417149\ndiv\nSideBar@https://kissu.moe/js-ui/ui-bundle(4.11.8-debug).js:1:387759\ndiv\naside\nsection\ndiv\nThreadWindow@https://kissu.moe/js-ui/ui-bundle(4.11.8-debug).js:1:464454\nPostSSRRouteDecider@https://kissu.moe/js-ui/ui-bundle(4.11.8-debug).js:1:361807\nRoute@https://kissu.moe/js-ui/ui-bundle(4.11.8-debug).js:1:209199\nRouter@https://kissu.moe/js-ui/ui-bundle(4.11.8-debug).js:1:205355\nFailWindow@https://kissu.moe/js-ui/ui-bundle(4.11.8-debug).js:1:444165\nClientRouter@https://kissu.moe/js-ui/ui-bundle(4.11.8-debug).js:1:309946"}
Agent: Mozilla/5.0 (Android 11; Mobile; rv:93.0) Gecko/93.0 Firefox/93.0
Time: Sat Dec 18 2021 00:50:48 GMT-0500 (Eastern Standard Time)
No.8898
Also, there's no "Hide/Show" thread option on mobile.
No.8899
>>8897I can't replicate, but I can figure it out from that.
>>8898how exactly would you unhide a thread on mobile.
There's no room.
No.8900
Uploading audio file with embedded album art is broken with error "Undefined index: height"
No.8902
nvm, it probably tried to OCR the image for spam links in the thumb.
No.8905
>>8901>>8902Previously I tried several with different album arts and none worked
It's working now anyway, thanks
No.8906
>>8905there's some systems that automatically trigger different antispam measures and I think one of them activated.
No.8931
>>8928I'd rather you discuss this in the feature centered thread not to distract from this one. But I'll give the heads up and say it's likely not going to happen because inserting latex into the pages looks like a pain.
Perhaps on commision. Though, I have a growing list of bugs, mostly relating to /poll/, that I have to get around to.
No.8942
Bug: Yentext is the wrong color when you preview your post in any of the themes except "Kissu."
No.8944
>>8942>>8943Disregard my bug report, turns out the problem was that I was still running the old userscript.
No.8945
>>8944Wait, I turned off the userscript and it kept happening. I did have the "color:purple" thing in
>>8943 set, though. Was that a default at some point? I don't recall setting it. I removed it and it fixed the problem.
No.8946
>>8945It was in the initial release and removed after a week
No.8948
hey V, how does the suggested threads section work, exactly? is it purely random or does it follow some algorithm?
No.8949
>>8948it picks 20 threads based on the first few who post on it(after passing some thresholds). Then inside of threads the order is random
No.8952
>>8951which browser is this? I don't get it on firefox?
No.8953
/secret/ is completely borked on pale moon. Might be the browser's problem though.
No.8954
>>8952Brave, so chromium.
No.8955
>>8953Hide the post with the lines of 死ね until I fix the issue.
>>8954I guess I'll try installing brave, but since I don't see it in chrome then it's probably a more situational thing
No.8956
I can get it by flinging my mouse as fast as possible while trying to move a window, but only on chromium browsers, so I've replicated it. I'll see about a fix.
I'll be fixing things after new years. Except the /secret issue that I'll have to get to sooner than later
No.8979
Captcha hasn't been working for me. The box shows up, but none of the images or buttons to do the captcha.
No.8980
>>8979yeah, the ssl cert changed and I forgot to move it over to the other server.
I'll to that now
No.8981
I stand corrected, it's because i used the url 4taba2 to reverse proxy
No.8988
>>8987I thought the background image was headless but apparently it was just me not using my browser in full-screen.
No.8994
palemoon issue resolved
No.8996
In regards to the audio/video controller, there are various issues with the design choices by the Mozilla team and (especially) the chrome team made. While it might be possible to use CSS to fix the firefox controller, the chrome controller is quite bad.
>>>/qa/78818 looks horrific on chrome and on firefox you have a limited control over the audio volume and no trackbar when it could fit in our containers.
I'm going to create a custom controller. This is an expansion on the techniques I learned during my audio experiments as done with the Christmas cards
- on mouse over of the volume icon it will pop out a vertical bar to set volume
- if the player is smaller than 175px it will show pause|progress|volume
- if player is >= 175px will show pause|progress|trackbar|volume
- if player is > 255px will show pause|progress:max-length|trackbar|volume
And I suppose if there's something extra cool that you want I can consider it such as adding an equalizer, but idk.
No.8997
>>8996>And I suppose if there's something extra cool that you want I can consider itFor audio files, assuming their metadata isn't nuked, it might be nice to have a small rotating ticker of the artist, album, and song title.
No.8998
reminds me that I should account for a place to put features should I ever have the motivation. Chromes ' : ' button on the side of players could hold a variety of extra information and is easy to fit into a 175px bar.
>>8997 could be put into an "About" dropdown
No.9003
>>9002hm.. yeah a recent change would do that yeah. Guess I'll undo it.
No.9004
>>9002new posts will be fixed, old ones will have the issue. Since it's the summary system it will be quickly flushed and repaired.
Anyways, here's the prototype for the audio player to replace the existing one for audio.
https://codepen.io/echibiki/pen/mdBKJGEAside from making the Audio-UI more usable for Kissu's audio frames, it also makes the sound fade out smoother so you don't get a clicking sound when you press stop.
It doesn't give buffering messages, but tracks the bufferable audio. Not sure about how it handles on mobile yet.
No.9005
>>9004Added time tags, so you hover over the time bar and it returns a second.
For now this is only for audio and I'm concluding the prototype.
My laptop should be repaired soon so I can get back to the scheduled bug fixes and then implement it into the software.
https://codepen.io/echibiki/pen/mdBKJGE No.9006
worked pretty well only issue was that i needed to click it a second time maybe because for some reason it didn't work the first
No.9007
>>9006There are many things that are clickable. which one are you talking about?
No.9008
well it runs in private tabs of Edge, Chrome, PaleMoon and Firefox so it was probably nothing.
No.9009
>>9007The first time I tried it I needed to click the generate audio button twice, but now it just works.
No.9010
actually did find a bug in palemoon, what a surprise that outdated browsers are holding me back from progressing onto other things.
>>9009ok, nothing major
No.9016
There was actually a little quick fix i left in which resulted in people being able to make not polls to /poll/. Decided I'll leave it in for now.
No.9017
Issues with poll creation should be resolved as of Kissu-Fr 4.11.14b-debug
No.9022
>>9021yes, i noticed it in the morning,,, but as soon as I posted it vanished. I made a change to accomodate for PaleMoon being a shitter and it's introduced a hard to track bug
No.9025
Well, I hope it happens again so I can see it and do a proper test because it looks like it shouldn't happen
No.9027
Speaking of dumb PM glitches, the animated banners will occasionally freeze up for some reason.
No.9029
>>9027can't do anything about your browser not being able to play .gif consistently...
No.9038
Imageframes now (4.11.20) resize their container when adjusted, so no longer will there be a giant empty box when it's resized.
No.9039
>>8951Does this frequently happen to you? I really have to force it for it to happen.
Fixing it means I'm going to have to dive into this guy's code since I didn't implement the dragability myself
https://www.npmjs.com/package/react-draggable No.9040
>>9039Yeah, I checked and it happens every time. Looks like the click registers on the background of the QR box before the QR box itself? Even without an image behind it, there's a half a second lag between me clicking the box and moving my cursor and the box following the movements.
If its not a wide spread problem then I don't really care if you don't fix it. In fact,
>>8951 was the first time I ever noticed that happening. I don't move the QR box usually.
No.9041
>>9040Brave: Version 1.34.80 Chromium: 97.0.4692.71 (Official Build) (64-bit) ?
Next question, what OS is that? My cursor doesn't look like this.
I'm trying to reproduce it, but no luck.
No.9042
actually, i am getting it now.
No.9043
>>9040Insert this into your customCSS
img, a {
-webkit-user-drag: none;
-khtml-user-drag: none;
-moz-user-drag: none;
-o-user-drag: none;
user-drag: none;
}I'll have to find a nicer solution than completely disabling dragging on certain elements or specifically target brave browsers somehow. Until then this works for me.
No.9044
your stupid feedback thing appears on archive.is nice
No.9045
>>9044Don't think I've had many positive experience coming from archive.is.
https://web.archive.org/web/20220107183812/https://kissu.moe/ No.9047
>>9025Happened again >>>/jp/27838
No.9050
>>9047alright, just have to theory craft on why it vanishes on post then I can solve it. I know why it makes a funny post though.
>>9048what the fuck. this is truly bizzare.
No.9051
>>9048The only theory I can think of is that I'm downloading files to do a server restart and in that moment you couldn't access the server files that make posts..
No.9052
>>9050>>9051>this is truly bizzare.I should add: for whatever reason, after trying to reupload the file I was trying to upload, I got an error saying something to the effect of "file not selected." despite having one attached. I tried 2 or 3 more times before changing browser where I was actually able to upload the image, although if it was something server side maybe it was just that minute or so of changing browsers was enough for the post to go through and the browser made no difference?
No.9053
>>9052where is the image?
No.9055
>>9054server probably had no ram. can't be for certain. all I know is that I have been downloading all the files for the past hour or three
No.9056
>>9047think I found the problem, missing escape character in the regex which golang didn't throw any errors for
No.9057
>>9043Yeah, it doesn't happen with image in the back anymore. Thank you for the quick fix!
No.9058
Next update is going to add in a fix that adds the CSS whenever someone drags the QR box. So you can remove that from customCSS if you want to drag other things and not concern yourself if you clear cache.
Update also going to add in the custom audio player. Feel free to suggest/report on it.
Kissu-Fr 4.12.1-debug
I have two bug items left for flash resizing and a site navigation issue.
Also we're going to be doing a server rebuild meaning I'm backing up everything and reinstalling the servers. This should buy us the last stretch of storage space before the server upgrades.
No.9059
>>9056AAAAAAAAAAAAA That's not the fix
No.9060
I'm doing some aggressive testing for at least 10 minutes that will make browsing the site on the kissu-ui uncomfortable.
No.9061
>>9060K, I put in another fix for that issue. Hopefully this one does it, but it's starting to look like a rare golang issue, though I shouldn't speak too soon because often times it's user error...
But still, the function gets called once and if it's processing other tags(it is) then it certainly should be reading the regex to detect cites
No.9062
mother FUCKER
No.9063
>>9047 put in another fix
No.9064
I've redone my palemoon fix, which caused this issue in the first place, to use a way that doesn't heavily alter the golang regex tag parsing. HOPING that this issue doesn't trigger again.
Going to be doing a full rebuild of everything on the site so posts might be slightly slower
No.9085
bleh, I wonder if I should just investigate this and give up on the idea of a complete vichan rewrite
https://github.com/deuill/go-php No.9086
>>4175site's down huh... guess the github page is enough for them
No.9087
>>9085If it were me, I'd rewrite it bit by bit rather than all at once, swapping out individual components written in PHP with Rust/Go equivalents, until eventually there's no PHP left. In my mind, replacing post.php now and then maybe something else later is way easier than uprooting the entire site.
No.9088
>>9087The thing about kissu is that it's already written in Golang and PHP, but I do the communication through curl calls which is slow and has overhead. Not like there's a better option since the PHP code is used by php-fpm processes handled by NGINX. Connecting the two together somehow might be beneficial.
No.9089
If I think about it more logically the two points that consume the most CPU and RAM are constructing the vichan templates and retrieving from the database. There are other infastructure related issues, but I'll put those aside since they're improved by better hardware. Although, the quick fix for database speed is more ram.... Communication between PHP and Golang is problably not a big deal... though it could be better...
If I switch the template writting over to golang then things will be a bit faster. Vichan's DB is set up better than I could probably do. Not sure I can make the structure any faster for data access. Probably need to rewrite queries to be better for mysql.
Best I can do on a limited economy/budget.
No.9090
>>9089Doing some research it's apparent I know next to nothing about database optimization.
A single google search yields me some ways to make it faster.
No.9091
Posting looses a 1-2 seconds on these queries
Posting looses a 1-2 seconds on these queries
[code] User@Host: ommitted[ommitted] @ localhost [127.0.0.1] Id: 3
# Query_time: 0.303444 Lock_time: 0.000526 Rows_sent: 300 Rows_examined: 2694
use vichan;
SET timestamp=1642089894;
SELECT id,thread,subject,email,name,trip,capcode,body,body_nomarkup,time,bump,files,num_files,filehash,sticky,locked,cycle,sage,embed,slug,zombie,'qa' AS `board` FROM posts_qa WHERE thread IS NULL UNION ALL SELECT id,thread,subject,email,name,trip,capcode,body,body_nomarkup,time,bump,files,num_files,filehash,sticky,locked,cycle,sage,embed,slug,zombie,'jp' AS `board` FROM posts_jp WHERE thread IS NULL UNION ALL SELECT id,thread,subject,email,name,trip,capcode,body,body_nomarkup,time,bump,files,num_files,filehash,sticky,locked,cycle,sage,embed,slug,zombie,'win' AS `board` FROM posts_win WHERE thread IS NULL UNION ALL SELECT id,thread,subject,email,name,trip,capcode,body,body_nomarkup,time,bump,files,num_files,filehash,sticky,locked,cycle,sage,embed,slug,zombie,'cry' AS `board` FROM posts_cry WHERE thread IS NULL UNION ALL SELECT id,thread,subject,email,name,trip,capcode,body,body_nomarkup,time,bump,files,num_files,filehash,sticky,locked,cycle,sage,embed,slug,zombie,'poll' AS `board` FROM posts_poll WHERE thread IS NULL UNION ALL SELECT id,thread,subject,email,name,trip,capcode,body,body_nomarkup,time,bump,files,num_files,filehash,sticky,locked,cycle,sage,embed,slug,zombie,'b' AS `board` FROM posts_b WHERE thread IS NULL UNION ALL SELECT id,thread,subject,email,name,trip,capcode,body,body_nomarkup,time,bump,files,num_files,filehash,sticky,locked,cycle,sage,embed,slug,zombie,'f' AS `board` FROM posts_f WHERE thread IS NULL UNION ALL SELECT id,thread,subject,email,name,trip,capcode,body,body_nomarkup,time,bump,files,num_files,filehash,sticky,locked,cycle,sage,embed,slug,zombie,'ec' AS `board` FROM posts_ec WHERE thread IS NULL ORDER BY bump DESC LIMIT 300;
# Time: 2022-01-13T16:04:55.488676Z
# User@Host: ommitted[ommitted] @ localhost [] Id: 2
# Query_time: 1.576677 Lock_time: 0.001532 Rows_sent: 30 Rows_examined: 60246
SET timestamp=1642089895;
SELECT *, 'aut' AS `board` FROM `posts_aut` UNION ALL SELECT *, 'b' AS `board` FROM `posts_b` UNION ALL SELECT *, 'cry' AS `board` FROM `posts_cry` UNION ALL SELECT *, 'ec' AS `board` FROM `posts_ec` UNION ALL SELECT *, 'f' AS `board` FROM `posts_f` UNION ALL SELECT *, 'jp' AS `board` FROM `posts_jp` UNION ALL SELECT *, 'poll' AS `board` FROM `posts_poll` UNION ALL SELECT *, 'qa' AS `board` FROM `posts_qa` UNION ALL SELECT *, 'secret' AS `board` FROM `posts_secret` UNION ALL SELECT *, 'spg' AS `board` FROM `posts_spg` UNION ALL SELECT *, 'sum' AS `board` FROM `posts_sum` UNION ALL SELECT *, 'test' AS `board` FROM `posts_test` UNION ALL SELECT *, 'trans' AS `board` FROM `posts_trans` UNION ALL SELECT *, 'win' AS `board` FROM `posts_win` UNION ALL SELECT *, 'xmas' AS `board` FROM `posts_xmas` ORDER BY `time` DESC LIMIT 30;
# Time: 2022-01-13T16:04:55.790531Z
# User@Host: ommitted[ommitted] @ localhost [] Id: 2
# Query_time: 0.272866 Lock_time: 0.003257 Rows_sent: 1635 Rows_examined: 4905
SET timestamp=1642089895;
SELECT id,subject,email,name,trip,capcode,body,body_nomarkup,time,bump,files,num_files,password,sticky,locked,cycle,sage,embed,slug, 'aut' AS `board` FROM `posts_aut` WHERE `thread` IS NULL UNION ALL SELECT id,subject,email,name,trip,capcode,body,body_nomarkup,time,bump,files,num_files,password,sticky,locked,cycle,sage,embed,slug, 'b' AS `board` FROM `posts_b` WHERE `thread` IS NULL UNION ALL SELECT id,subject,email,name,trip,capcode,body,body_nomarkup,time,bump,files,num_files,password,sticky,locked,cycle,sage,embed,slug, 'cry' AS `board` FROM `posts_cry` WHERE `thread` IS NULL UNION ALL SELECT id,subject,email,name,trip,capcode,body,body_nomarkup,time,bump,files,num_files,password,sticky,locked,cycle,sage,embed,slug, 'ec' AS `board` FROM `posts_ec` WHERE `thread` IS NULL UNION ALL SELECT id,subject,email,name,trip,capcode,body,body_nomarkup,time,bump,files,num_files,password,sticky,locked,cycle,sage,embed,slug, 'f' AS `board` FROM `posts_f` WHERE `thread` IS NULL UNION ALL SELECT id,subject,email,name,trip,capcode,body,body_nomarkup,time,bump,files,num_files,password,sticky,locked,cycle,sage,embed,slug, 'jp' AS `board` FROM `posts_jp` WHERE `thread` IS NULL UNION ALL SELECT id,subject,email,name,trip,capcode,body,body_nomarkup,time,bump,files,num_files,password,sticky,locked,cycle,sage,embed,slug, 'poll' AS `board` FROM `posts_poll` WHERE `thread` IS NULL UNION ALL SELECT id,subject,email,name,trip,capcode,body,body_nomarkup,time,bump,files,num_files,password,sticky,locked,cycle,sage,embed,slug, 'qa' AS `board` FROM `posts_qa` WHERE `thread` IS NULL UNION ALL SELECT id,subject,email,name,trip,capcode,body,body_nomarkup,time,bump,files,num_files,password,sticky,locked,cycle,sage,embed,slug, 'spg' AS `board` FROM `posts_spg` WHERE `thread` IS NULL UNION ALL SELECT id,subject,email,name,trip,capcode,body,body_nomarkup,time,bump,files,num_files,password,sticky,locked,cycle,sage,embed,slug, 'sum' AS `board` FROM `posts_sum` WHERE `thread` IS NULL UNION ALL SELECT id,subject,email,name,trip,capcode,body,body_nomarkup,time,bump,files,num_files,password,sticky,locked,cycle,sage,embed,slug, 'win' AS `board` FROM `posts_win` WHERE `thread` IS NULL UNION ALL SELECT id,subject,email,name,trip,capcode,body,body_nomarkup,time,bump,files,num_files,password,sticky,locked,cycle,sage,embed,slug, 'xmas' AS `board` FROM `posts_xmas` WHERE `thread` IS NULL ORDER BY `bump` DESC;
# Time: 2022-01-13T16:04:56.746230Z
# User@Host: ommitted[ommitted] @ localhost [] Id: 2
# Query_time: 0.514141 Lock_time: 0.000455 Rows_sent: 20 Rows_examined: 29469
SET timestamp=1642089896;
SELECT *, 'aut' AS `board` FROM `posts_aut` WHERE `files` IS NOT NULL AND `BODY` != '' UNION ALL SELECT *, 'b' AS `board` FROM `posts_b` WHERE `files` IS NOT NULL AND `BODY` != '' UNION ALL SELECT *, 'f' AS `board` FROM `posts_f` WHERE `files` IS NOT NULL AND `BODY` != '' UNION ALL SELECT *, 'jp' AS `board` FROM `posts_jp` WHERE `files` IS NOT NULL AND `BODY` != '' UNION ALL SELECT *, 'poll' AS `board` FROM `posts_poll` WHERE `files` IS NOT NULL AND `BODY` != '' UNION ALL SELECT *, 'qa' AS `board` FROM `posts_qa` WHERE `files` IS NOT NULL AND `BODY` != '' UNION ALL SELECT *, 'spg' AS `board` FROM `posts_spg` WHERE `files` IS NOT NULL AND `BODY` != '' UNION ALL SELECT *, 'sum' AS `board` FROM `posts_sum` WHERE `files` IS NOT NULL AND `BODY` != '' UNION ALL SELECT *, 'win' AS `board` FROM `posts_win` WHERE `files` IS NOT NULL AND `BODY` != '' UNION ALL SELECT *, 'xmas' AS `board` FROM `posts_xmas` WHERE `files` IS NOT NULL AND `BODY` != '' ORDER BY `time` DESC LIMIT 20;
[/code]
No.9092
As far as I know the queries can't be optimized..
keys of
+----+--------------+-----------------------------------+------------+------+------------------------+-----------+---------+-------+------+----------+---------------------------------+
| id | select_type | table | partitions | type | possible_keys | key | key_len | ref | rows | filtered | Extra |
+----+--------------+-----------------------------------+------------+------+------------------------+-----------+---------+-------+------+----------+---------------------------------+
| 1 | PRIMARY | posts_aut | NULL | ref | thread_id,list_threads | thread_id | 5 | const | 105 | 100.00 | Using index condition |
| 2 | UNION | posts_b | NULL | ref | thread_id,list_threads | thread_id | 5 | const | 215 | 100.00 | Using index condition |
| 3 | UNION | posts_cry | NULL | ref | thread_id,list_threads | thread_id | 5 | const | 150 | 100.00 | Using index condition |
| 4 | UNION | posts_ec | NULL | ref | thread_id,list_threads | thread_id | 5 | const | 150 | 100.00 | Using index condition |
| 5 | UNION | posts_f | NULL | ref | thread_id,list_threads | thread_id | 5 | const | 30 | 100.00 | Using index condition |
| 6 | UNION | posts_jp | NULL | ref | thread_id,list_threads | thread_id | 5 | const | 150 | 100.00 | Using index condition |
| 7 | UNION | posts_poll | NULL | ref | thread_id,list_threads | thread_id | 5 | const | 142 | 100.00 | Using index condition |
| 8 | UNION | posts_qa | NULL | ref | thread_id,list_threads | thread_id | 5 | const | 301 | 100.00 | Using index condition |
| 9 | UNION | posts_spg | NULL | ref | thread_id,list_threads | thread_id | 5 | const | 126 | 100.00 | Using index condition |
| 10 | UNION | posts_sum | NULL | ref | thread_id,list_threads | thread_id | 5 | const | 139 | 100.00 | Using index condition |
| 11 | UNION | posts_win | NULL | ref | thread_id,list_threads | thread_id | 5 | const | 59 | 100.00 | Using index condition |
| 12 | UNION | posts_xmas | NULL | ref | thread_id,list_threads | thread_id | 5 | const | 68 | 100.00 | Using index condition |
| NULL | UNION RESULT | <union1,2,3,4,5,6,7,8,9,10,11,12> | NULL | ALL | NULL | NULL | NULL | NULL | NULL | NULL | Using temporary; Using filesort |
No.9093
upgrade to mysql 8 with full innodb expanded cache size not much improvement. Possibly worse
No.9094
>>9093Oh, maybe I spoke too soon.
posts_qa now only takes 0.06S and my perception of speed posting is better enough that I don't think I'm imagining it
No.9101
made two small changes.
- Clicking on the cat in the capcode does something
- You can expand the QR text box from either the right side or the bottom in addition to the existing bottom-right icon.
Next changes are some updates of our software such as captchouli. Except PHP8 because it doesn't look like it has payoff given my increasing interest in ditching PHP. Related is analyzing how long it takes for HTML templates of pages to build.
Also I am hoping that I stumble across a bug where navigation does not work. The URL changes but pages stay the same. That's why the version numbers still have -debug at the end.
No.9102
I don't even notice the slow posting you're talking about. I make a post and it gets posted, after that it's done, it's out of my hands and the server will process it automatically. Who cares if you have to wait a bit before you can refresh the page and see your own post...
No.9103
>>9102Well... In theory, it matters at scale. An unoptimized solution may be fine for a smaller community, but if/as things grow that slowness will only compound. Besides, it's just preferable to have things perform quicker anyways.
No.9104
>>9102yeah, doesn't really matter, but there is a way to use communication platrforms that requires instant(under one second) feedback such as IRC. If Kissu were to get that sort of performance then there would be no need to use discord/irc/liveboards over the system here.
There are things about efficiency and scaling up that are worthwhile, but this is more abstract.
No.9114
Template building is inconsiquential
Build post_thread.html: 0.004040374 seconds
Build thread.html: 0.001412901 seconds
Build post_thread.html: 0.001575339 seconds
Build post_thread.html: 0.000622027 seconds
Build post_thread.html: 0.002187153 seconds
Build post_thread.html: 0.00071731 seconds
Build post_thread.html: 0.000746706 seconds
Build post_thread.html: 0.000609239 seconds
Build post_thread.html: 0.000566081 seconds
Build post_thread.html: 0.000623017 seconds
Build post_thread.html: 0.000710531 seconds
Build post_thread.html: 0.000775186 seconds
Build post_thread.html: 0.00059533 seconds
Build post_thread.html: 0.000661931 seconds
Build post_thread.html: 0.000685007 seconds
Build post_thread.html: 0.000613555 seconds
Build post_thread.html: 0.000367887 seconds
Build index.html: 0.000712868 seconds
Build mod/archive_list.html: 0.000432297 seconds
Build page.html: 0.000259245 seconds
Build themes/rss/rss.xml: 0.000340275 seconds
Build themes/catalog/catalog.html: 0.003558448 seconds
Build themes/catalog/index.rss: 0.002807624 seconds
Build post_thread.html: 0.000921932 seconds
Build post_thread.html: 0.000934841 seconds
Build post_thread.html: 0.000722056 seconds
Build post_thread.html: 0.000926266 seconds
Build post_thread.html: 0.000568823 seconds
Build post_thread.html: 0.000393111 seconds
Build post_thread.html: 0.000704053 seconds
Build post_thread.html: 0.000710426 seconds
Build post_thread.html: 0.000831989 seconds
Build post_thread.html: 0.000541476 seconds
Build post_thread.html: 0.000262487 seconds
Build post_thread.html: 0.000887318 seconds
Build post_thread.html: 0.000667229 seconds
Build post_thread.html: 0.000755937 seconds
Build post_thread.html: 0.001272209 seconds
Build themes/ukko/index.html: 0.000866665 seconds
Build themes/index/index.html: 0.000389548 seconds
Vichan has built in optimization to make this smooth. Thinking a bit harder the API engine doesn't have any optimization like this. Probably suffers from speed issues due to this.
No.9115
Yeah, I'm pretty sure I know of a major optimization I can make to the golang API server.
Nice that I've basically finished designing the site so I can actually go back and look at what I did sloppy.
No.9116
think this upgrade is going to make posting go to under one second for non-image posts, but I have to set up a dummy database and test server on the main site so it will take me a bit of time.
No.9117
The image in >>>/qa/83396 is rotated sideways on pale moon. If it's a browser-side issue, you don't need to (and indeed probably shouldn't) fix it, but I still think it's worth looking in to.
No.9118
>>9117Either it's reading this tag and other browsers don't, or it's not reading this tag and other browser do.
Not a priority issue
No.9119
PaleMoon doesn't read exif orientation and while it could be sensible to rewrite the image in the correct orientation it doesn't appear to be an issue effecting anything other than people's photography, which isn't a major site component.
https://repo.palemoon.org/mcp-graveyard/UXP/issues/1601?lang=zh-CN No.9120
I'm going to be uploading some experimental code during the stream and running some tests on the kissu-UI API. I'll point out the two threads I'm using so you can hide them, but thread creation and (unlikely) posting could be a little wonky as in threads not showing up or multiple duplicates appearing at the same time. Possibly crashes. Don't bother reporting any to me in that time frame.
Reason I'm doing it on the site is because I don't have a full development environment set up and I've done enough testing on my limited one that it's unlikely that pages will explode with errors.
No.9134
Storing optimized time values for operations
Trial 1
['{"Page":"build","Number":"5558","Board":"b","Category":"thread"}' '{"Page":"build","Category":"catalog","Number":"5558","Board":"b","Options":""}' '{"Page":"build","Number":"9149","Board":"b","Category":"properties"}']
build 495.127594ms (unaltered)
build 324.229499ms (altered)
build 343.400976ms (unaltered)
['{"Page":"build","Category":"overboard","Number":"5558","Board":"b","Options":""}' '{"Page":"build","Category":"summary","Number":"9149","Board":"b","Options":""}' '{"Page":"build","Category":"home","Number":"5558","Board":"b","Options":""}']
build 1.048950064s (altered)
build 679.720795ms (altered)
build 569.857718ms (altered)
Trial 2
['{"Page":"build","Number":"5882","Board":"b","Category":"thread"}' '{"Page":"build","Category":"catalog","Number":"5882","Board":"b","Options":""}' '{"Page":"build","Number":"9150","Board":"b","Category":"properties"}']
build 408.760266ms
build 318.564994ms
build 372.223012ms
['{"Page":"build","Category":"overboard","Number":"5882","Board":"b","Options":""}' '{"Page":"build","Category":"summary","Number":"9150","Board":"b","Options":""}' '{"Page":"build","Category":"home","Number":"5882","Board":"b","Options":""}']
build 1.459589376s
build 1.006235501s
build 709.6905ms
Trial 3
['{"Page":"build","Number":"6426","Board":"b","Category":"thread"}' '{"Page":"build","Category":"catalog","Number":"6426","Board":"b","Options":""}' '{"Page":"build","Number":"9151","Board":"b","Category":"properties"}']
build 578.914277ms
build 362.66053ms
build 380.621354ms
['{"Page":"build","Category":"overboard","Number":"6426","Board":"b","Options":""}' '{"Page":"build","Category":"summary","Number":"9151","Board":"b","Options":""}' '{"Page":"build","Category":"home","Number":"6426","Board":"b","Options":""}']
build 1.649670211s
build 601.593662ms
build 601.996869ms
No.9137
Unoptimized
Trial 1
['{"Page":"build","Number":"6193","Board":"b","Category":"thread"}' '{"Page":"build","Category":"catalog","Number":"6193","Board":"b","Options":""}' '{"Page":"build","Number":"9136","Board":"b","Category":"properties"}']
build 445.856118ms (unaltered)
build 1.511872274s (altered)
build 521.85504ms (unaltered)
['{"Page":"build","Category":"overboard","Number":"6193","Board":"b","Options":""}' '{"Page":"build","Category":"summary","Number":"9136","Board":"b","Options":""}' '{"Page":"build","Category":"home","Number":"6193","Board":"b","Options":""}']
build 4.299701033s (altered)
build 533.114344ms (altered)
build 555.392242ms (altered)
Trial 2
['{"Page":"build","Number":"6186","Board":"b","Category":"thread"}' '{"Page":"build","Category":"catalog","Number":"6186","Board":"b","Options":""}' '{"Page":"build","Number":"9139","Board":"b","Category":"properties"}']
build 336.30166ms
build 1.526173125s
build 358.760967ms
['{"Page":"build","Category":"overboard","Number":"6186","Board":"b","Options":""}' '{"Page":"build","Category":"summary","Number":"9139","Board":"b","Options":""}' '{"Page":"build","Category":"home","Number":"6186","Board":"b","Options":""}']
build 3.374175136s
build 662.375161ms
build 603.671679ms
Trial 3
['{"Page":"build","Number":"1528","Board":"b","Category":"thread"}' '{"Page":"build","Category":"catalog","Number":"1528","Board":"b","Options":""}' '{"Page":"build","Number":"9144","Board":"b","Category":"properties"}']
build 605.778204ms
build 1.722577337s
build 369.152895ms
['{"Page":"build","Category":"overboard","Number":"1528","Board":"b","Options":""}' '{"Page":"build","Category":"summary","Number":"9144","Board":"b","Options":""}' '{"Page":"build","Category":"home","Number":"1528","Board":"b","Options":""}']
build 3.981038488s
build 562.800892ms
build 471.384885ms
No.9141
>>9138>>9120You should be voting after the stream is done anyways since it doesn't store your past answers each time
edit: oops
No.9154
>>9137>>9134This method uses JSON modification on the assumption that certain actions will happen. Instead of querying the database and doing rebuilds every time.
Does not yield good results on small queries. Somewhat expected I guess since it needs to do two file IO functions and one small database read as apposed to one file IO and one smallish read. Guess I'll see if it ever beats the three trials on
>>9137 while I have the tab open
No.9158
ah crap, i discovered a typo that was slowing the entire thing down
No.9162
K, the combination of optimization and discovery of a fairly major bug has given us these stats
['{"Page":"build","Number":"6288","Board":"b","Category":"thread"}' '{"Page":"build","Category":"catalog","Number":"6288","Board":"b","Options":""}' '{"Page":"build","Number":"9161","Board":"b","Category":"properties"}']
build 149.206856ms
build 36.421313ms
build 346.792749ms
['{"Page":"build","Category":"overboard","Number":"6288","Board":"b","Options":""}' '{"Page":"build","Category":"summary","Number":"9161","Board":"b","Options":""}' '{"Page":"build","Category":"home","Number":"6288","Board":"b","Options":""}']
build 167.621666ms
build 1.381298ms
build 128.121682ms
No.9164
would be great to have sorting by last post in old ui if it's something easy to implement
No.9171
>>9164It's unlikely that I'll ever be touching vichan again. For free anyways
No.9173
The vichan UI that is. I modify the backend a lot.
No.9175
Keep getting this error after submitting a captcha on /ec/
Mobile Debug: SyntaxError: JSON.parse: unexpected character at line 4 column 1 of the JSON data -- Fatal error: Uncaught Error: Call to undefined function do_filters() in /var/www/html/post.php:1209 Stack trace: #0 {main} thrown in /var/www/html/post.php on line 1209 {"error":"Caught fatal error: Uncaught Error: Call to undefined function do_filters() in \/var\/www\/html\/post.php:1209\nStack trace:\n#0 {main}\n thrown in \/var\/www\/html\/post.php<\/strong> on line 1209"}
No.9176
>>9175strange, i'll make a captch
No.9177
I can't replicate it but someone reported this once before and I assumed it was because I was uploading files at the time. But since it happened again there must be something a bit more tricky going on.
No.9178
yeah, ok. I looked at line 1209 and I can see that something is off.
No.9179
OK it should be resolved. When a mod hasn't posted or done any activity it triggers a set of stronger antispam features. I made a mistake in not including a PHP function file.
No.9180
Wow, it really got a lot quicker.
No.9181
yes. right now the only thing slowing down the site is image processing speed and the image transfer speed between servers.
Our tentative upgrade is going to be double the Ram and CPU with 250GB of storage. So images will be way faster and file size limits will likely be raised.
No.9193
Here's the file in question if you need it:
https://files.catbox.moe/mmhv6h.webp No.9194
Yeah. I know what caused that. I'll get it back up
No.9197
>>9164>>9171In response to this chain
https://invoice.stripe.com/i/acct_1IY8JvGGyFWk7gzP/live_YWNjdF8xSVk4SnZHR3lGV2s3Z3pQLF9LelJYUVdyUWpjdVlsSkp3Z1YyN2hqcUNpMHNHaEti0100yI2V3vBCIf there's a feature I don't think is a good use of my time, then I'll create an anonymously addressed invoice that anyone can pay. The payment will guarantee that I start working on it as soon as I see it's requested with a dollar sign in front of it.
This is feature requests on commission. If there's other software unrelated to Kissu that people want from me then we can work out other deals. My email is in the footer of the site.
No.9198
>>9197If you want me to attach your name to the feature(For example Berun-Last-Reply.js) then you can also request that. This doesn't actually give you any ownership of the file, it's just a way you can buy virtual land on kissu.moe through the software... Hey maybe I should make every html file have an NFT...........
No.9207
Seems as if I've caught every bug under my radar. Hopefully stays that way but it never does. Puts me into a situation where all the software design is stable again.
So I'm going back to RnD
Banner program ought to be moved off of laravel.
I mentioned a simple user statistics page for funders a few months back.
Also want to eventually do the rustgo transition.
So going to draft the Rust server concept for the 3 services but until I actually make the statistics program not much point in making it.
No.9226
Listing feature specs and update plans for various issues raised in the above post.
Bullet point overview of below
- Creation of an imageboard engine is unlikely to happen unless I can live off of the site. Still Drafting will help cut down on development time if it ever happens.
- More likely to happen is I look into the possibilities of PHP-Go to remove the CURL communication done between API and post server(I've yet to measure the impact it has, but it can't be great).
- Laravel should be discarded as it will be impossible to update and Laravel is one of those frameworks that I could see people finding major exploits on.
- Statistics program is something I've wanted to add for funders as a tie in to >>8551 and to potentially discard umami.js which requires a few gigabytes to run and is now blocked by ublock origin making it practically pointless.
- Mostly pointless is that the Kissu-UI pages are served out of a NodeJS server. My language plan for Kissu is for it to be Rust, Golang and Python only. Python has good data processing, Golang is a fast language that is far simpler than Rust, Rust is fast secure and safe. The server does barely anything as apposed to the past where it was prerendering the UI. It can be redone whenever I have a few hours.
Summarization of features in a potential in house imageboard engine.
Implementation of server would require advanced level knowledge in multiple fields and require multiple precursor projects. To acquire enough knowledge to create a more optimized versions of existing projects may be a futile process. Still it's a fun concept.
- Multithreaded connection pooling
- MySQL IO
- Image Processing(Thumbnailing, exif stripping, perceptual hashing, OCR)
- Controlled caching of information from MySQL
- Be bindable/usable with Python, Golang and PHP. This should be a scripting language to quickly implement ideas on top of the engine.
- Templating in Twigg
- Account creation and security
- JWT Authentication(for banners.kissu and potential mod UI redesigns if ever)
More reasonable solution than server rewrites
Could be seen as a gateway project for a server rewrite, merge Go and PHP so that our two servers are no longer communicating with CURL.
- Likely will be fairly simple job
- Need to rewrite the PHP file that handles communicating with the Golang server
- Go server would only handle the PHP requests such as post.php, score.php, poll.php and so on. HTML files would continue to be served by NGINX.
- Most complicated part is what will break in the process such as the libraries that PHP uses.
Proposed banners changes
I'm a bit uncomfortable with Laravel. It's a fine tool, but it's impossible to update now. If a CVE becomes known it may be very hard to fix it resulting in the service being offline indefinitely. We're removing laravel.
- In line with the 3 main scripting languages I want new Kissu to be using for server infastructure it will be rewritten in Golang, Python or Rust.
- The program uses some ease-of-use features such as database creation and routing that will have to be taken into my own hands.
- JWT authentication will need to be learned about since it uses this to keep accounts logged in.
- Adding in board specific banners
New Statistics Program
Offered to people providing for the site, it will give them statistics about kissu and their current token/IP address.
- In Python, Django or Flash. I know flask, but Django is more production oriented Lets just do golang
- will replace umami.kissu as it takes up 1GB storage for node modules, is rather heavy and is blocked by ublock.
- Combination of React elements for interactability with static fallbacks for people who like simplicity.
- Timelines of the site's performance.
- Information about your post counts.
- Host the variety of interesting and harmless information found on umami https://puu.sh/IDZb0/e08c305be9.png
- Mods may have greater levels of information from this page.
- Private and not shared.
No.9232
>>>/poll/1856Should the text be moved to before the poll? I've noticed myself not noticing the text a lot.
No.9233
>>9232you can revote whenever though
No.9243
I don't think it's something on my end, but backlinks don't seem to be showing up in this thread
>>>/win/667 No.9244
the thread was moved so it probably made a mistake. I'll look into it in a bit
(noting that former thread ID was 28491)
No.9245
resolved the issue(as can be seen from the move) >>>/trans/1862 but I'll have to manually go through each post in the /win/ thread to restore cites.
No.9251
I just had an idea, instead of using the unix timestamp as the default name for saving files why not use a kissu timestamp that starts when kissu went online or when /qa/ got good.
No.9252
>>9251Ehhhh but then people would think we're tryharding with old timestamps and that wouldn't be fun.
No.9253
>>9252everyone else can suck it or switch to kissu time
No.9254
>>9253Everyone can tell
us to suck it in turn. If we're gonna use a different time, it should at least have some prefix that makes it different from the standard. Something like "fgh-13426..."
No.9255
>>9254just use the timestamp backwards or append a "k" to the end
No.9256
>>9255K at the start and we're good.
No.9267
>>9266I think it makes sense.
No.9268
>>9267Yeah, it does help. Yet it changes how people use the site, so while it's a convenience I'm concerned about that aspect.
No.9269
>>9268That's very true. Now that I think about it, it might be a good idea to keep it how it is now.
No.9270
>>9268Afraid that people might start fishing for (You)s by making low effort troll posts? Yeah I don't want that either.
No.9271
>>9270Sort of. not exactly that but kind of a similar thing. It's like balancing a game, certain aspects get buffed it has an effect on the balance and certain moves/techniques/builds become more prevalent. In the end it might make the site meta too much like 4chan or any of the nchan's rather than kissu.
No.9272
"File too big" error should probably be changed to "Image dimensions too large" or something like that.
No.9274
>>9272I got caught up on that issue before when I was running some tests. What you say is a good error response
- Error message changed, I don't have a test file though so you can verify it's good if you want
No.9275
>>9274Is it total dimensions (H + W) or per axis? The image I tried uploading (
https://files.catbox.moe/3r5msl.jpg) is 6250 x 4103, which in total dimensions is greater than 10,000px, but less than if it's supposed to be per axis.
No.9276
>>9275I'll check.
I also removed the annoyingness with ) getting caught into the URL
No.9277
No, the dimensions were getting detected correctly, but the error message was using the wrong value. There are config files scattered all over the server so creating descriptive error messages is a bit of a pain.
6000px is what I set it as. Any more than this is begins to drain the server RAM too much
No.9279
Speaking of image dimensions, it would be nice to have a "fit to view" option for image expansion
No.9281
>>9279something like that shouldn't even be optional. it should just be the default
No.9282
Not really any bug or issue in particular - I'm not familiar with how things work exactly on the backend with the gradual move away from Vichan - but I remember from my experience with Vichan that wordfilters can be set up to give multiple results instead of just one to one wordfilters. Maybe it's not such a big priority, but it could be practical for obfuscation to have multiple permutations, or otherwise for more fun wordfilters (assuming we ever get any). Something to think about I suppose.
No.9283
>>9282I know how to do that even without vichan. But I'm unsure how that would make wordfilters more effective
No.9284
>>9283The way I imagined it, maybe you might think otherwise, is that if one or word or phrase can end up being filtered into any number of other words/phrases, then it becomes more difficult to interpret what the person was actually talking about, whereas if there's just one word filtered into one other word it becomes obvious what the person was talking about that got filtered. Perhaps it's not going to absolutely prevent people from figuring out what the word/phrase was that got filtered, but -- from the reader's perspective -- it's going to be a lot more abstract unless and until someone manages to evade the filter.
That's my thinking at least. I don't really have any strong feelings about it, it was just something that occurred to me at random.
No.9297
Does /ec/ have autonoko now?
No.9298
>>9297for a long while
>>6594>>9284I can believe this, but I wouldn't change any of the existing filters. I'll consider it if we add filters later.
No.9299
I wish hidden posts/threads were more hidden.
In my mind, hiding something should make it as easy to forget about as possible, and the way it is now makes that kind of hard. With threads, hiding them actually makes them harder to ignore.
No.9301
I have no idea what you're talking about with threads being harder to ignore
No.9302
Threads are hidden from /all/, the recent feed, the homepage, the similar thread list and you have the gall to say (without giving me any sort of constructive ideas) that it's not good enough?
No.9304
>>9300>>9301>>9302If it were up to me, I'd move the hidden threads button off the navbar and put it somewhere on the page proper, maybe on the top-right with the board-list. This way, it'd be a bit more out of the way but still easily reachable. I'd also change it from 'x Threads/Posts Hidden' to 'Hidden Threads/Posts' and make it show up even if you haven't hidden anything, making it easier to forget if you've even hidden something to begin with.
I'd like to see these changes, but I'm also afraid they might be a bit extreme, which is why I didn't open with them.
No.9305
>>9304ok. that is a fair consideration. Then I could move thread search into the sidebar possibly or other items that are currently at the top of the page which could be used while scrolling
No.9317
posts which are "[text in here]" get turned into "[-]" on the recent posts ticker.
No.9318
>>9317Hmm. the "[text in here]" part got removed on the recent posts ticker. So, I guess they get removed and the "[-]" means there's nothing there.
No.9319
>>9318information inside of [-] is discarded on the feed.
No.9320
Why is only yandex linked and not saucenao as well?
No.9323
I think the part about making it it a generic toggle and not showing the counts is a good idea though. I'm not sure about having it be a permanent fixture.
No.9324
>>9322My logic was that you'd see it less because it doesn't scroll with you. The sidebar's always there unless you close it, meaning the hidden threads button is also always there. When it's moved to the page proper you only see it when you're at the top of the page. But I can definitely see where you're coming from.
>>9323I'd personally make it a thing you can toggle in options. It feels pretty inline with some of the other things you already have in there, like (You) backlinks and relative time.
No.9325
>>9324I'm going to do the bare minimum at the moment and see if it continues being an issue. It will stay up all the time and show
0 threads hidden. I think this will also act as a way to remind people that they
can in fact hide threads instead of wanting to derail them.
No.9331
A recent post just reminded me, does the new UI have the ability to filter posts? Never tried it
No.9332
>>9331regex filters don't exist. Actually adding them is probably easy. Biggest concern is how to fit them on the options box. Possibly a new tab and a textbox like 4chanX
No.9335
i feel like the options field shouldn't be saved by default. i've accidentally saged a few times now. i feel like it should be intentional every time on a site that gets as few posts as this!
No.9337
>>9336only because i feel like the posts are generally worth more than the posts i see on other sites
No.9338
that's part of why i keep coming back!
No.9340
>>9335If I make a change to sage it will flip the other way very hard. That kissu appears slow is actually to it's benefit.
I have thought about something like 30% of the time a sage won't go through, this would be very annoying and piss everyone off, but produce the kind of results you're looking for.
>>9332Won't add regex filters someone says they're wanted.
No.9341
I've noticed that the worst posts coincidentally are bumping the thread.
No.9342
bump
No.9347
when I open
https://original.kissu.moe/ I get this web server error message. The URL worked an hour ago. The message says:
Warning: require(/var/www/html/wp-blog-header.php): failed to open stream: No such file or directory in /var/www/html/index.php on line 48
Fatal error: require(): Failed opening required '/var/www/html/wp-blog-header.php' (include_path='.:/usr/share/php') in /var/www/html/index.php on line 48
No.9348
>>9347it just fixed itself...
No.9350
yea, it probably went to index.php which I haven't used in years.
Deleted that file so shouldn't happen again
No.9353
I'm gonna run a cleanup utility soon and stuff. If it goes wrong the files might vanish temporarily(they're being backed up though)
No.9367
Best to test for the absence of all modifiers (control, shift, alt, meta) in addition to whether it's a left click. You never know what useful feature someone's browser might implement. I think shift+click is usually new window, and alt+click is a download in some browsers.
No.9368
I have to delete the other posts here because the images weren't moved correctly or something
>>9367Yeah, thanks for the tip. I'll be sure to add it onto the hundreds of other things to look at that could wind up crashing the site or lead to exploits.
No.9370
works on my machine™
No.9372
ctrl clicking doesn't work for me either btw
No.9374
This isn't a bug, it's a lack of a feature
No.9376
I know for a fact it won't be ported over, but I miss the tree view from the old UI. I know citation chains kind of do the same thing, but IMO the tree view does it better.
No.9377
>>9376I don't really use my chain feature because I think it stands apart from the alternative system of modifying the inline view,
I think my chain idea has potential if people who like condensed views can articulate the point better. Maybe it's intrinsically flawed though. I don't really want to add more features to the UI unless it seems like the site's losing out to a competitor.
No.9378
Thinking about it, the chain and tree ideas are very different functions even though I'm trying to make the chain concept replace the need for trees/inline
No.9379
Did /trans/ get wiped?
No.9380
In the past few months? Not to my knowledge. It's not like it has more than 5 pages though
No.9381
After further investigation it seems like just the index is broken. Catalog seems to work fine. Maybe it's just an old UI thing because it seems to work fine in a private window with the new UI.
No.9382
>>9381I see, you should have specified vichan-ui, because I don't touch that "thing" anymore. Well I guess I can have a glance at why it's not working
No.9383
I dunno if it will keep working, but I regenerated the pages
No.9385
It wasn't really a fix, I just slapped some duct tape on it. I'm barely even maintaining the UI of vichan.
No.9387
I'll be adding some documentation to the site for the Vichan JSON API, Kissu JSON API and the Kissu-UI API after I finish it. Will place a link to it in the FAQ and here. I have to modify that page as well so might as well do that in the process. Documentation will specify between work in progress API and stable API.
Kissu UI will be receiving these signals.
document.body.addEventListener("qr-menu-off", this.handleQRMenuOffEvent)
document.body.addEventListener("qr-comment-off", this.handleQRCommentOffEvent)
document.body.addEventListener("qr-new-comment", this.handleQRCommentNewTextEvent)
document.body.addEventListener("options-new-tab", this.handleQRCommentNewTextEvent)
Giving off these
let navigation_event = new CustomEvent<LocationDetail>('kissu-location-change', { detail: location_detail });
let open_event = new CustomEvent<boolean>('option-menu-state',{ detail: state });
let open_event = new CustomEvent<boolean>('qr-menu-state',{ detail: state });
let markup_event = new CustomEvent<MarkupDetail>('qr-markup-clicked',{ detail: markup_detail });
let change_event = new CustomEvent<BudDetail>('qr-bud-changed',{ detail: bud_detail });
Will be documented and will provide the API.ts file that acts as a controller.
No.9388
Not a huge priority, but it would be greatly appreciated if the QR box didn't disappear until after a captcha has been submitted. It's pretty disorienting if you try to post something, end up needing to do a captcha and then seeing the QR box vanish. I'd also submit that it's the case -- at least for myself -- that I often think about my post a bit more when I have to do a captcha, so the QR box vanishing is rather problematic if I think of a change I want to make to my post.
No.9389
>>>/qa/85428Ideally it could be, but it's come out of a discussion on something anime related and mutated into discussions about gender stereotypes which is really a pain to deal with for everyone on this site who isn't interested
>>9388I'll consider it. I'm multiple days behind on my API update so it will happen later on, likely next month.
No.9391
>>9390The captcha writes tmp files. Said tmp file probably was erroneously deleted.
No.9392
The images on /cry/ are all giving 404 links if nobody noticed yet.
No.9393
>>9392They are slower than normal, but there's no 404s to my knowledge
No.9394
nothing suggesting excessive usages so probably provider related. Either mine or the server's
No.9395
/cry/'s images seem fine on new UI, but not there on old UI
No.9396
>>9395for lack of a better explanation something "odd" has happened with vichan's page builder that has caused each image to try and open a /b/ url. It seems like it was perhaps associated with a full rebuild I did from the command line?
It's back working to my knowledge, there's probably some other boards so I'll adress that
No.9425
The thread search doesn't seem to include titles on /ec/. For instance, searching "alice" returns no results.
No.9427
>>9425Weird, it works on other boards
No.9429
It seems as if there's a bug with the search feature bleh... I'm glad I started using git forks instead of doing everything on the same branch because I stalled on my API work and the bugs have reached a point where I have to fix them.
No.9446
Can't upload banner. Getting "Server error with request".
No.9447
>>9446Resolved it, thanks for report
No.9453
>>9452fix'd
I just "edited" them and resubmitted them without any changes and I guess it was like updating them. If anyone else sees posts like that let us know. I'm pretty good at computers.
I'm leaving the OP here as-is so vern can look at it >>>/qa/77320
2
No.9454
>>9453Thanks. After a more thorough look I found a few more in that thread:
>>7025>>7162>>8922>>8932>>8972>>8991 No.9456
i'll just fix it on the server
No.9457
Hiding a post should also hide all replies to said post.
No.9459
>>9457alright, i'll add it to me "do when you have motivation" list
No.9496
Newly uploaded banners on the banner submission page lack a file extension. This doesn't seem to pose any problem in terms of displaying the banners since the MIME type isn't being stripped, but clicking on the image causes these now extensionless files to be downloaded rather than opened in a new tab unlike previous banners which still have a file extension.
No.9497
Hidden threads are stratified by catalog halves. This means that not all hidden threads are actually shown at once.
For example, if you hide a thread on Pages 1-6, and another on pages 7-10 (Also, why is it 1-6 and not 1-5?) When you click the "[ + ]" to show hidden threads, you will only see one thread instead of two depending on which side of the catalog you're on. Likewise, based on which half you're on, you will only see X Threads Hidden based on which side you're on. So, for example, you might see "3 Threads Hidden" on one side, and "5 Threads Hidden" on the other. Similarly, because the board is split into halves, posts are removed from each half resulting in less pages per half. Suppose you hid 15 or more threads, this will resulting in you losing the last page of a given half. So, if you had 15 or more threads hidden on pages 1-6, you would lose page 6. Page 6 would cease to appear in the Index to click on, and manually going to /board/6 shows a page with no threads at all.
No.9498
>>9497it downloads catalog page 1 /api/threads/b/catalog/1.json, but knows the total number of threads in /api/properties/b.json
This makes it not have to download the full catalog /b/catalog/full.json but as a result I can't know what you've hidden on catalog page 2.
So yes, I know this exists because I don't have a solution other than to use full.json for everything.
Why 6 instead of 5? Idk, wasn't important.
No.9499
The obvious solution to me is that people who hide should just use the catalog, but I wish I had a way to make it work better with the index
No.9506
Please check 4taba, urgently.
No.9507
Didn't mean to sage. To users, don't go to 4taba, there's very illegal spam.
No.9508
>>9507another day another dolar
No.9509
On the topic of making hidden posts more hidden, again, it would be nice if posts in hidden threads didn't still count as new posts in the title when indicating "(X)" new posts.
No.9510
>>9509I guess I overlooked this. This is probably the last thing of this nature.
No.9517
>>>/ec/7762there was nothing to post. Just the captchouli pop-up, but instead of captchouli filling in, it was just the outline where it would normally be. no images, no text, no "some error has happened" message. nothing. just a box inside the window.
No.9547
I often wonder if there's a non-insignificant cost to having the default be for all pages, even on the old ui, to auto-update.
No.9558
the mood to work on the site is improving. bug list is at ~30 items now and I'm having ideas for new features
No.9559
also need to follow through with my server upgrade ideas and pick between my three server host options (continue hostinger, a2 or another one that I'll keep unnamed
No.9562
expired polls probably should probably have their options be greyed out and unselectable. Would also probably make sense to have the "poll expiration" thing be visible by default rather than having it behind the results.
No.9563
>>9562good usability idea
No.9569
>>9568maybe. I can look at it.
There's a bug that I noted down where markup creates it's own separate block. If it's unrelated, there are some things with text blocks that just exist and we(4chan, vichan, myself) just hope will only happen once every ten thousand posts.
No.9570
>>9558>and I'm having ideas for new featuresNice
No.9571
Site's loading pretty slow for me right now
No.9572
never had this problem
No.9573
Well, you're in a similar geolocation as another guy who comments on the site being slow sometimes. Perhaps it's cloudflare or your service provider
No.9574
It's loading a tad slow for me as well.
No.9576
>>9575I think it's probably a cloudflare issue, maybe? Somehow, by changing my DNS from cloudflare to my ISPs default my speed on speedtest.net went from 100mb/s to 500mb/s.
No.9577
>>9576I haven't experienced the issue even though Mod has brought it up a handful or two of times so I'm guessing that it's something to do cloudflare in "your region"
No.9578
>>9425fixed
also since I can't find an error that required some console.log statements, the site version will end up having -debug removed soon
No.9584
It's even slower than before now...
No.9585
>>9584I can't help you fix your comcast issues
No.9599
>>9242cared
>>9562adding it, lets see how it looks
No.9601
>>9242What is this SFX for?
No.9602
>>9601Can't remember. There was some nyaa sound effect somewhere on the site, but I can't remember where.
No.9603
k, polls changed for better UX
debugging calls removed(possibly better application performance idk)
No.9604
>>9601Uhh, there was an option for playing a sound on post on new UI. I'm not sure if nyaa was used for that or if was the "yay". Let's see...
No.9606
>>9604It's the Show by Rock "yay"
No.9607
Oh, you already changed it. It was the sound effect on clicking the admin icon.
No.9609
The polling implementation isn't handled very well for this feature. There will be a bug, but I won't fix it unless it becomes a problem because fixing it could either be a pain or not be a pain, and i don't feel like doing hard thigns
No.9612
>>9608>>9611feel like I found an easter egg
No.9614
Speaking of polls, it might also be valuable to have a "hidden poll" option where the results are only known once the poll ends.
No.9616
>>9615Yeah, I figured. It probably wouldn't be used much anyways. Thought it would mainly be useful for the seasonal stream polls.
No.9628
Something I might do that takes effort is adding in the 4chanX reply chaining...
The reply chain window kissu has doesn't fulfil the same role that I envisioned it doing
No.9632
>>9509Seems like this can't be done without reworking the counter system.
Didn't account for this issue when adding the feature.
I'll keep it around on my issues list as something on the complicated side
No.9633
>>9557done with other CSS fixes
No.9636
>>9568I made an update to fix this problem.
I can't cause such a thing to happen, but I'm half expecting someone doing something dumb with markup will cause pages to break someday.
No.9659
>>9658Didn't even know that existed.
No.9660
>>9658I have an idea! How about a sequence of no body posts get ignored except the first post until a non no-body post shows up?
No.9661
>>9660yeah, that might work better. Show the first and last posts in a series of empty posts and omit the in betweens.
No.9664
>>9663yeah, i guess there is a bit of a feature overload...
I basically only browse the site through that feed nowadays so I assumed everyone was on some levels similar to me. Don't even use /all/ half the time.
No.9674
So I made it that if two posts that count as either [-File-] or [-Video-] are posted next to one after another in the same thread then it will get condensed into becoming a [-Image-Set-] item.
No.9678
>>9677ye, needed some sort of identifier to show when new posts are made, but the visual language is kind of contradictory here for reasons you just said.
Some alternatives could be:
[-Image-Set-] :8000:[-Image-Set-] {8000}[-Image-Set-] [8000][-Image-Set-] "8000" No.9679
>>9678What about the usual
>>8000 or something similar?
No.9680
>>9679asymmetry doesn't look nice in that menu
No.9704
>>9509how the hell do I resolve this...
No.9705
Would be easier if I had websockets, but I think I've thought up a solution...
Will be the last bugfix on the kissu-ui agenda
No.9706
fixed some unintuitive markup issues and information display bugs. Still have another rare bug to fix with the hover previews on posts with [read more].
No.9708
Looks like there's spam links getting posted as replies on /trans/. I only see them on >>>/trans/2148 so far.
No.9709
>>9708that was just verm testing
No.9710
>>9709Oh. False alarm then.
No.9712
>>9711noticed, but decided it's not noteworthy enough to fix. Although, maybe hidden could untoggle when nothing is hidden
No.9713
and the hidden count is correct, as as I intended, however I think it should change
No.9714
>>9713What do you mean by working as intended? Is it supposed to show the number of all threads hidden site-wide regardless of who hid it?
No.9716
>>9711I put in a partial fix to this issue
>>9714and resolved this one
No.9749
the properties.json file is going to be added to in order to fulfil this request
>>9509https://kissu.moe/api/properties/all.json and etc will have a
post_feed field that will keep track of the last 10 actions on a board
Will have fields:
type PropertyFeed struct{
No int `json:"no"`
Resto int `json:"resto"`
Board string `json:"board"`
Sage int `json:"sage"`
Deleted int `json:"deleted"`
}
This way deleted and new posts inside of hidden threads/Boards will not have an effect on the new post counter placed on top of threads
No.9757
I'm not fully testing your feature request. Tell me any bugs
No.9777
Caught up on all the UI issues that were brought up. Dismissed/shelved some feature ideas.
No.9783
all reported issues have been resolved. So then, back to the assignments I had previously started... UI-API, userscript, stats pages
No.9814
Current list of things in the log in no particular order:
- API to modify UI behavior and allow modification of the QR
- Modify URL downloads to be client side(store as option that's default off, will be needed if kissu goes to single server config)
- Kissu Stats page
- Donor user-statistics page+login
- Donation reward of a User-Script with various complicated to implement behaviors
That's basically all my medium term goals.
Most notable thing to take in is that I view the current site as stable enough to release a User-Script API and am looking at services surrounding kissu rather than modifying it directly.
No.9815
Long terms:
Kissu Vichan replacement(includes many adjustments to existing features)
> Delete-post/thread API Optimizations> Mod action API Optimizations> Remove thread items(or disable) in a locked or 404 thread> Doing major rework on vichan's markup to help make Fr easier (https://imgur.com/GMm33B6 , https://imgur.com/qSimILr , urls in preview )
> OCR scanning color profiles No.9816
Backlog - Likely never to be touched:
> Catalog revision
> mobile notifications
> full Rust rewrite of backend
> Rust IRC server and integration into site
> More SEO optimizations
> Infinite scrolling opt-in
> FA Brands
> IndexedDB where appropriate instead of localStorage(with localstorage fallback)
> Reply hidinging chains
> Hiding/unhiding a post should also hide all replies to said post.
> Advanced menu. Replaces the preview into something allowing advanced features such as thumbnail selection on webm
No.9837
UI dev API looks to be in a good starter state. Will be released as Kissu-Fr 4.16.0. Not sure if I'll make use of it now or later since I intend it for vip features. Stats page goal has some an all purpose nature to it.
https://github.com/ECHibiki/Kissu-Fr-API-DocsBefore uploading making a thread about it(and an update to the FAQ)Have to tend to:
- Modify URL downloads to be client side(store as option that's default off, will be needed if kissu goes to single server config)
and - Improve the navigation error messages(timeout, 404, 403, 500 etc.)
No.9838
perhaps I should document the JSON API as well in another repo
No.9845
FAQ updated to reference developer tools.
JSON API Documentation:
https://github.com/ECHibiki/Kissu-JSON-API-Docs/blob/main/README.mdKissu-UI API Documentation:
https://github.com/ECHibiki/Kissu-Fr-API-Docs.
4.16.0 will have UI-API and an update to error messages when files are not found.
No.9846
Kissu will be beginning server transfers sometime at the end of this month or the next.
I'm planning to spend roughly 28$ / month on leasing server infrastructure. I'll leave the details a bit ambiguous for security purposes but the intention is for us to have better Cloudflare performance and more thorough backups. If Kissu were to for some reason go down I want to make it so that only up to 24 hours of data loss is the max at any time. No loss of original data, no loss of thumbnails... A complete restore.
I'll make a thread about it when it's relevant
No.9847
Any thoughts/plans on the latest image formats like AVIF and JPEG XL?
No.9870
It would be nice for the mobile UI to have the ability to hide posts. It's inconvenient needing to go into Desktop mode just because the mobile UI post dropdown is missing the option.
No.9871
>>9870making me find a place to put the hidden toggle .....
No.9872
ye, i have no idea where I would put the unhide toggle. I suppose I can add in the hide toggle in the dropdown, but you won't be able to readd it
No.9873
actually, this screams of bad UX. So I'm not sure how to approach your concern
No.9874
Kissu has no way to move replies to threads without a mod going through each of them individually.
Because of this reply chains can't be preserved. I would need to create a new mod feature to resolve this. It is very different from the /trans/ issue
No.9875
Just letting you know if you make a multiline quote and put a space before a URL, it tries to turn it into a link. Very important bug, please fix immediately.
No.9876
*multiline code snippet
No.9877
>>9876>>9875I see it. Mostly a matter of correcting the regex
No.9878
ah, I actually had to move a line of code up the execution chain so I wonder if some other markup aspect is going to be broken now. Oh well... someone will find the bug in like 3 months
No.9880
>>9870I'm thinking about it a bit more and would it make the mobile version worse than desktop if I were to give it hidden post stubs while the desktop version allows for full removal?
No.9881
Actually I might have a solution along with some other improvements
No.9882
I've notated down that the mechanism for File URL uploads is going to be changing in the future.
- From the footer reply form it will likely vanish or become a button much like on the QR form
- In the QR form the client will download the image if the image source allows it in their page rules.
- If the image source does not allow clients to download through javascript it will fail and I will have an error message directing people towards a userscript which will do the job for them.
No.9884
near future mobile modifications:
> Hiding on mobile same was as desktop
> On mobile swap catalog and update on the index
> Increase visibility of update-thread in threads
> Correct the position of options on home page
> mobile notifications(should be doable somehow)
No.9885
Misc list of changes that are lower priority than what I've said above:
> Catalog show index page numbers
> In locked/404'd threads prevent replying
> More SEO optimizations
> Delete-post/thread API Optimizations
> Mod action API Optimizations
> Hiding/unhiding a post should also hide all replies to said post.
> create a better system for luna/haiji transfer.(store on haiji then move it to luna after the post confirms, use a generic subdomain and have that subdomain proxy to the correct server) Basically a CDN
No.9886
Also there is another list which splits into two scenarios, but this one is private. It depends on another person's future
No.9894
>>9892i got it,
>>9878took less time to notice
Should be fixed for good now. Let me just rebuild that thread
No.9910
made a few changes to try and remove some generic bots.
No.9911
Daily backups confirmed working.
Next actions are going to be the mobile fixes(primarily hiding items)
Also some moderator tools such as allowing for mods to move a set of replies and maintain the post links
No.9913
Updates Kissu-Fr 4.16.11
Released when I finish today
Fixes:
some QR/Comment related bugs
Mobile changes confirmed:
More functionality such as hiding and better placements of stuff such as updating
When there's a new post related to your settings you'll get one of the blue notifications
Sidebar removed mode:
More tabs at the bottom of the page for unhiding and etc.
No.9924
How can I get old viewing mode on mobile? The new one is pretty yabai.
No.9926
AKA. I have no idea what you like about it. Looks terrible and poorly designed for mobile in my opinion.
Anyways. Done for today.
No.9927
The image at >>>/qa/90841 is broken for some reason - the thumbnail looks fine but it gives a 403 error when I try to expand it. Maybe something to do with it being a .bmp file rather than a more common format?
No.9928
>>9927ye, probably no rules for the server to display bmp
No.9929
I'm typically against things such as .bmp and .wav because they eat away at drive space. But since it's just one file I'll let it pass.
No.9935
WHY DID NO ONE TELL ME IT IS IMPOSSIBLE TO SUBMIT A POLL THROUGH THE CATALOG!!!
No.9938
>>9937It must have to do with which oldui /all/ page it's on. I suppose I can look at it for 30 minutes and if I can't fix it then I won't bother
No.9955
>>9935You can't even create polls through the new UI anyways, and never have been as far as I know. It's also kind of always broken.
No.9956
>>9955I made the "summer" thread just yesterday
No.9957
Even before that I made the poll about UI choices.No idea what you're talking about.
No.9962
Fixed some bugs with polls. Modified behavior when a cite/recent feed/homepage post doesn't exist.
Also trying to resolve an issue with a permanent buffering graphical effect.
No.9964
>>9956>>9957Nice dodging the fact that you STILL cannot make polls through the new UI. Regardless, my point was about there being rendering issues with options from time to time, but I'm pretty sure this is when
someone decides to modify the options after the fact or do something beyond what polls should be normally.
No.9966
>>9964oh boy... How do you think I make the polls. Do you think I do it through a moderator login?
No.9968
I think you're starting to go crazy. I suggest using the new UI. It helps make your brain stronger
No.9969
Anyways, allow me to just reaffirm you for a second that the old vichan pages will be around on the site forever(I believe) because Google has some issues caching the site due to how it's built. So I'm having the webcrawlers directed to old pages.
No.9970
but let me just emphasize, I am doing it for the sake of webcrawlers in order to get the site to index better on search engines, not because i think It's good, but because it's simple enough for them to understand.
No.9971
Current known buglist fixed. Resuming this month's backlog
> Catalog show index page numbers under threads
> In locked/404'd threads prevent replying(remove forms and openables)
> Mod feature to move reply chains into other locations and preservve links
No.9973
>>9972I can't fix your slow internet for you
No.9974
>>9972It's probably just taking too long to load, have had that happen to me in the new UI.
No.9975
Dude, I have no idea why vichan doesn't do what it's supposed to do.
Why don't you publish a report on
https://github.com/vichan-devel/vichan/ No.9977
>>9973>>9974It's not an internet speed issue. It doesn't do anything no matter how long I leave it, whereas if I push the same 'click to expand' button from the index, rather than the last 50 posts page, then it loads almost instantly.
No.9978
>>9974Kissu ui is a bit different because it triggers the change and then it requests a file while you wait.
Vichan just locks up your entire browser until it's finished(BUT IT'S SO GOOD!!!) so you won't see the effect change until it finishes locking up your environment.
What probably happened is he failed to fetch the file from the server because (vichan is such a piece of shit that it requests 100kb files for updates) that he can't do anything.
No.9979
>>9977maybe you should use a functioning user interface
No.9980
sorry, I meant to say that every time you hit update on the shit-ui you're requesting a 1.92MB file, but I suppose that's the price of freedom
No.9989
>>9982
nvm, brotly encoding is built into cloudflare and it's doing good enough
No.9990
>>9982
>the shitty code that 8chan wrote .
What's that?
No.9991
>>9990Vichan is the prototype that became 8chan. Most of what it uses is built off of this engine. Eventually the two projects diverged to the more lightweight version that can be run on the cheapest of servers(vichan) and the board creation engine Infinity.
Infitity got some updates and lots of typical 8chan dev drama.
Vichan got nothing and is stuck with buggy code, ancient UI and the occasional walking time bomb exploit.
I don't give a rats ass about anything to do with vichan aside from the fact that it's free and it's what we've been using for 2 years until I begin to write up a replacement. Even if I do replace it then the UI pages will stay because they're free and serve a purpose. That's literally the only reason it's around. To serve a purpose. The purpose is not anyone using it, the purpose is not my enjoyment with maintaing it. It is because it doesn't require javascript. That is it.
No.9996
>>9990>what's thatDudcode
No.10014
Am considering putting the [u url] tag and the [det]+[sum] tag into the markup-helper. But Maybe just the URL and I remove the [s custom] field to give to keep the bar the same length
No.10017
>>10014I think that's a great idea.
>>10016I honestly don't know why there's two Ms to begin with. Being able to turn the markup menu off and on has no practical benefit, at least not with how it's set up right now.
No.10018
>>10017looked neat and wanted to add it. Same story as every other feature minus the few that have been requested.
No.10022
Markup hotkey binds have been revised to make more sense with the helper icons.
Markup helper options: Hidden custom, added URL markup option and Dropdown markup option
Changes
New:
alt+A
alt+D
Revised:
alt+B glowblue
alt+S strong(formerly shared bind with R and known as bold)
alt+I italics(formerly shared bind with E)
alt+8 Spoiler
alt+- Strikethrough No.10023
have a bunch of rendering issues to fix now
[u https] and other markup rendering inside of code block on previews https://puu.sh/J4RBM/a89c7c8d0f.png
URLs not rendering in preview https://puu.sh/J4RYd/a8f927afeb.png
content in a [det] shouldn't trigger a [read more]
URL inside of a [det] didn't get linkified
[u https] should only activate if it's a link in the form of [u https://(subdomain.)?domain.(TLD)(/asdf)?] ...
Markup should be ignored if it has no text body to it
a few feature modifications:
push markup clicks into undo/redo stack document.execCommand("insertText", false, "asdf");
new reply notices shouldn't ignore tabs
in reply form beside spoiler put a QR preview button
Stalling on the new features I have written down. Hopefully I won't feel like correcting/improving anything after this current batch
No.10027
In progress:
- Issue with markup undo(ctrl+z) not fully functioning
for some reason- new reply notices shouldn't ignore tabs
- update delay
- a mistake in build threads?
- silly float hack issues in previews
https://haiji.kissu.moe/test/src/1654944156024.pngresolved until I find bugs in them:
- keybind update Strong -> Bold , Blue -> bLue
- markup rendering inside of code block on previews
- URLs not rendering in preview
https://puu.sh/J4RYd/a8f927afeb.png- URL inside of a [tag] didn't get linkified
- prevent URL mismatch on submissions
- content in a [det] shouldn't trigger a [read more]
- Markup should be ignored if it has no text body to it
- in reply form beside spoiler put a QR preview button
- qr expand L and U side
No.10028
dammit, the preview in the post form is causing issues with undo
No.10030
Implementing my own undo-redo feature because it will be too hard to redesign existing function around default browser functionality.
No.10044
Putting in an upgrade soon to reduce the amount of bandwidth moving in and out of Kissu through page updates.
A small <100 byte file will be associated with each .JSON file and when checking for updates it will first go to this file before checking the main one.
This means lurking on the site will no longer consume as much data. If you use vichan, that's your own problem and suffer the consequences
No.10045
>>10030Undo functionality finished but not added to the site yet:
Undo markup insertion.
Undo text stored even if you change to a different item in the reply queue.
Undo text stored on page navigation.
No.10077
do you mean the | ?
Yeah, that's there because of a reason.
Although perhaps it shouldn't be appearing in the preview
No.10078
Oh nice... you reminded me that I forgot to get rid of the possibility to make invisible text with that feature
zawawawawawawawawawawaw
No.10082
Kissu-Fr Version 4.18.0 release
Notable changes:
¥ Undo and Redo will work always.
if you input markup you can undo it. If you link a post you can undo it. If you change item in post queue it will undo it. and so on.
This entire post was ctrl+z and then ctrl+shift+z so it shouldn't destroy anyone's posts by accident
¥ Instead of updates being done by checking main files(for example the catalog or this thread) it will check a 90 byte info.json file which will be downloaded and then if there are changes you will download the up to 200kb file after.
This means lurking on pages will not use as much bandwidth for either server or you. Retarded imageboard devs create a convoluted method of adding websocket functionality to imageboards. This is better.
¥ Fix a bunch of markup related bugs
¥ Fix a bunch of QR preview related bugs
¥ Add the the bottom post form the ability to preview posts and dismiss images.
This form is spam-bot resistant so I'm more inclined to add it to other places as people have requested in the past.
¥ Fix issues related to read-more and previews
Next I have to upgrade the mod tools to allow them to move multiple replies into threads while preserving the cite links.
Soon I will begin tearing apart vichan for the tsukuyomi engine
No.10088
>>10087as of
>>10027 the 11th
No.10101
Added in notifications that will pop up when you enter a locked thread
> This thread is locked!
And if you're inside of a thread that gets deleted it will output
> This thread is 404 not found!
No.10107
Added in a mod feature to move replies and maintain cite links
No.10109
>>10107>Added in a mod feature to move replies and maintain cite linksTHANK YOU
No.10110
messing around with the ban message settings for a bit:
One person thinks that ban message colors should be saturated to stand out from the rest of the posts.
I think they should be discrete since ban messages are just there to serve as a warning or for a mod to say that they think I'm being a dumbass.
Previously ban messages were minimized by default, but I decided they should be default maximized
No.10112
>>>/test/8046
I seem to have somehow created a thread that gives a 404 error for me when I try to view it on the old UI, but still shows up in the index/catalog and displays fine on the new UI.
No.10113
exhibition bans sounds 4chan as hell to me
No.10114
>>10113Basically yes. I don't want to give off 4chan associations with the feature usage or have people use it for it's intended purpose.
But it's usage between mods to basically show that one mod disproves with another is fine. This is what I'm trying to achieve with the public ban feature. Not for mods=gods comments.
((´・ω・)つ(・(・kneading tits) No.10115
>>10112I found a bug in deletes. Was it a delete through vichan UI?
No.10116
>>10115Yeah, I tried to delete it but it gave me some error message and after that it was stuck in limbo.
No.10117
>>10116I don't use that UI so bugs can go unnoticed since actions on the UI go through different pathways on the server. Important to include the details in the future
No.10119
last feature I'm adding to Vichan (by extension both UI).
After this, the UI and server software will be stable while I begin destroying vichan forever to write my own software which handles both UI styles.
Feature in question will be the ability to pick the frame for webm thumbnails. Easy to do on the old UI. I'll just add a new input field.
For new UI it will be integrated into the preview where a slider can be set giving you a preview of whay your thumb will look like.
No.10121
small tweak:
Character limit for Read-more on replies: 500 -> 800 characters
Character limit for Read-more on OP: 1000 -> 1600 characters
No.10123
Thread archive doesn't seem to work anymore, now threads 404 immediately after being bumped off.
Is it intentional or some error?
No.10124
not intentional, but I also see threads in the archive so I have to test
No.10125
>>10123I see what you mean
No.10126
Looks like I forgot to account for something with the linux permission levels so it's failing. I'll test again in a bit after my server file manager slowly sets permissions
No.10131
some boards will be seemingly broken for the current hour
No.10135
>>10123yeah, a matter of permission levels
No.10138
>>10136that means it crashed
No.10139
>>10137It now works when I test it
...
In trying to check for duplicates, the server runs out of memory.
If you want to know the nerd stuff
Description="banners"
Restart=always
User=b-data
Group=nogroup
WorkingDirectory=/var/www/banners.kissu.moe/
ExecStart=/var/www/banners.kissu.moe/cb2
MemoryLimit=100M
CPUQuota=10%
WantedBy=multi-user.target
This means I had it set for the banner program to only have 100Megabytes of memory max. This was based on a budget if everything on the site was running at once. There might be certain banner filesizes which will cause the program to crash. I think I have some large preexisting banners to check with.
No.10140
turned service on and off to test memory limits. Won't crash with <2MB files
No.10141
Something relatively dry, but my current view of the /f/ board is that content will be semi-permanent like on /b/ or maybe /qa/ where the two boards support much longer time limits on content
No.10142
Will I support HTML5.... HMMM.... we might have a subdomain eventually for this... Or maybe the /f/ board will be moved into a subdomain, we'll have to see what the future holds
No.10144
>>10143In the unlikely event that this is a browser specific problem or whatever is the cause I don't know, this happens with both Firefox and Tor. The delay isn't there on vichan UI.
No.10145
>>10143Using the elements on the page for various purposes is an interesting idea. React gets you into a situation where posts feel as if they are isolated from one another so fetching from a central resource keeps the code cleaner and less out of control. So picking the JSON resource was the first thing that came to mind. However, each post is also stored into an array of observers/listeners by the Flux/Redux style design, so I should be doing as you say.
>which is a confusing mismatch making it feel less responsive than it is for users who are used to the old style.that's your fault for being used to a standard that is going extinct on the internet. Every page on the internet written by someone who's not 60 years old is doing this now.
No.10146
It's fucking hilarious to be considering vichan as a standard for not wasting your data when every n seconds it pulls the entire HTML page out to do what it needs to do. Vichan doesn't even have it's operations integrated into JSON APIs.
It has these optimizations precisely because it is so shitty. But it is not worth disregarding how it worked around being such a terrible design
No.10147
have to use WASM FFMPEG to solve a use case requested by someone
No.10152
FFMPEG.wasm might be a write off.
- A security vulnerability forces it to be one core or I have to disable a bunch of site features
- Single core has errors I can't find a resolution to
- Only work around is reloading/rebuilding ffmpeg every time someone want a thumbnail
- It might be causing an occasional tab crash.
- Not sure the project is even maintained
Have to get to a situation where ffmpeg.wasm single core can be used(by end of day) or I'm going to have to use ffmpeg.js.
ffmpeg.js is slower though. By how much I don't know.
No.10153
after i release video thumbnail preview I'm going to make a poll asking if people want increased accuracy...
No.10154
bleh, actually i still have time so i'm going to make it so that the inacurate <video> version is attached to the preview system and then clicking on a certain icon will take them to a new tab which they can use the FFMPEG version to more accurately pick a timestamp if they're being memey. On hitting submit in the new tab it will transfer the info from the FFMPEG tab into the creator's tab.
No.10158
Second part is proving to be very advanced/complicated functionality. I don't expect any imageboard in the current decade to possibly do this idea in any comparable level.
No.10170
>>10145>that's your fault for being used to a standard that is going extinct on the internet. Every page on the internet written by someone who's not 60 years old is doing this now.Well they're wrong and I'm right.
No.10187
Nice feature, and thanks for having it work on the old frontend as well.
No.10188
>>10187By luck really. WebAssembly got completely gutted a few years ago following the Spectre and Meltdown CPU vulnerabilities which ruined multi-threaded applications(2017) and they've only started trying to make it work again(2020).
I can't have kissu.moe/thumbnailer/ work inside of the pages because youtube hasn't enabled a header, meaning that I'd have to use single threaded FFMPEG. And I tried to do exactly that except FFMPEG WASM hasn't had an update in a year and single-threaded breaks.
https://blog.logrocket.com/understanding-sharedarraybuffer-and-cross-origin-isolation/This tool is really a thing of modern web design
- multi threaded Web Assembly
- React Hooks
- Indexeddb API
- Broadcastchannel API
No.10803
I want to comb through the banners to find dead links. Can I have your permission?
No.10804
>>10803sure, you can check them