« first day (529 days earlier)      last day (3044 days later) » 

12:00 AM
RELOAD!
[Hosch250/Rubberduck] 49 commits. 41162 additions. 23455 deletions.
 
Oh, I see now.
 
[Hosch250/VSDiagnostics] 11 commits. 1453 additions. 169 deletions.
[retailcoder/Rubberduck] 40 commits. 9559 additions. 7765 deletions.
[rubberduck-vba/Rubberduck] 48 commits. 6 opened issues. 1 closed issue. 23 issue comments. 41397 additions. 23463 deletions.
[skiwi2/OGNext] 2 commits. 3 opened issues. 2 closed issues. 2 issue comments. 572 additions. 14 deletions.
[skiwi2/OGNext-Userscript] 2 commits. 1 opened issue. 1 closed issue. 200 additions. 50 deletions.
 
Guys @ Rubberduck been busy I see
 
[Vannevelj/VSDiagnostics] 6 commits. 1 opened issue. 5 closed issues. 5 issue comments. 392 additions. 52 deletions.
 
@SirPython ^^ search the page for "React-templates"
 
12:05 AM
That React code doesn't seem right.
render: function() {
return (
{this.props.items.map(function(item, i) {
return <li key={i}>{item}</li>
}}
 
For templating though we could just use something really simple like Markup.js though github.com/adammark/Markup.js
There doesn't need to be any logic beyond "put this section here, then that one after it, then that one and then display the assembled page statically"
 
I'm don't think we need to use a separate library, as React can already do that ^^ (just not directly in HTML)
 
Ok
In any case, I do think JavaScript is a better tool for manipulating DOM and HTML than Java is
 
Unless I don't understand what templating. Also, I think that sounds right for the logic.
What would be the benefits of changing to JavaScript from Java?
 
We've been using templating quite a bit for the HTML client with AngularJS
You know how there are different templates, best example IMO is for cards
 
12:11 AM
Oh, like the pre-stored HTML?
 
Yep, exactly
 
React definitely has support for that.
Easier support than Angular, I'd say.
 
Just load in the data whenever the server passes it down and the inject the template into the document
In the case of the static website it's even simpler, since it's all static
 
That's true. We could even render the HTML server-side.
 
Which is what we do now, I think, with the Java app
It just feels sluggish with Java, for some reason... stats.zomis.net/io-web
Try navigating between different pages. It feels like it shouldn't take near that long
But it could also have to do with @SimonForsberg's server load
Eventually, when the traffic justifies it, we can get a dedicated web server, though that's going to be a while
 
12:16 AM
It doesn't take that long for me, but maybe my connection is better/I'm closer.
 
OK
My connection is not that great
 
I can't find it now, but IIRC, Dan once said that rendering the HTML on the server side when using React could help create really fast webpages (especially if they are static)
 
Anyways, to me the primary downside of the Java app, as much as I think it's a cool app, is that it needs Tomcat 8+ to run, there are no other supported platforms at the moment
While JS works in practically all environments, right?
 
Except IE ;)
 
Well, nothing works in IE ;)
Vanilla JavaScript does but that about sums it up
 
12:21 AM
Is TOmcat the server?
 
Yeah
brb
 
Okay. Just out of curiosity, is there something wrong with Tomcat? Because, if it's what we are already using and it's working for us, then do we need to switch from Java?
@Marc-Andre Congratulations!
 
Well, not very many commercial web servers support Tomcat 8+ (AFAIK) which is my primary concern
However, as of right now, yeah, it works fine, and we can probably put this on the back burner until we improve/refactor the HTML client
 
12:41 AM
@DanPantry Do you know what's good practice for keeping global variables in React?
For example, user data?
SO isn't being very helpful.
It would be troublesome to just keep passing them in through props, unless that is the way to go.
And this data isn't initially constant, so it can't be stored in a separate JS file to be imported later.
Flux looks a little complicated since I'm somewhat new to React, still.
 
Not sure why I didn't do this a long time ago
 
1:10 AM
Oh, thanks!
Yikes, I just gave myself a start with that ping.
 
 
1 hour later…
2:21 AM
@Phrancis as far as I know the Java webapp does not need tomcat 8+
@SirPython on what are you congratuling me? I'm on mobile
 
 
4 hours later…
6:07 AM
 
 
1 hour later…
7:17 AM
@SirPython
libraries used for data flow tend to be some variant of Flux
I prefer alt, but redux has been getting a lot of attention recently
Remember that React is just a view layer, you shouldn't really be storing any data "with" it at all
 
7:31 AM
Actually, now tht I'm reading about redux it seems pretty rad..
 
 
2 hours later…
9:57 AM
hey
@Phrancis I'll see if I can find time to look into those audit thingies, else feel free to remind me in a while
 
 
1 hour later…
11:03 AM
@Phrancis Java doesn't manipulate the DOM. In the Java webapp, we're just generating the DOM initially, when the page is loaded.
@Phrancis I believe that no matter what framework or language we will use, we will most likely host it ourselves anyway. So it doesn't matter that most webservers don't support Java like that
@Phrancis I don't want to let you down or anything, but don't have your hopes up too high. The core code of Cardshifter is very complicated. Even for me and @jacwah :)
But you will probably be able to help with something anyway :)
 
 
1 hour later…
12:23 PM
@SimonForsberg That's the idea :)
 
@SirPython This is actually how React works, it uses a flavour of javascript called JSX
The code you linked gets transpiled into something that looks like this:
render: function render() {
  return this.props.items.map((item, i) => React.li({ key: i }, item))
}
That won't actually work though because you're returning multiple elements from render (which is a no-no). You'd need to wrap it in a <ul> or createFragment call.
 
 
2 hours later…
2:09 PM
hey everyone! :D
 
Hey @Marc-Andre !
 
Hey @Phrancis I see that you're now motivated for the website :P
I still don't understand why you say that the java web app need tomcat 8+, I don't remember it being the case. I may be wrong.
 
2:33 PM
It's entirely possible that I'm remembering it wrong
@Marc-Andre Do you think we should make the new page as the landing/index page to cardshifter.com by default, or make it a separate area of the site specifically as landing page when someone clicks the ad?
 
The true landing page ! We want people to see the prettiest things first !
We will have to make sure to fit the new style of the landing page to the cardshifter site (not the cyborg css)
This would be an alternative to the java app (and we could be hosted by githhub but no more dynamic things there)
I'll check tonight if the java web app need tomcat 8+
 
3:18 PM
@Marc-Andre Alternatively, maybe it was a problem related to the site actually being tied to the cardshifter.com domain name, instead of redirecting to stats.zomis.net/io-web/ ...
 
3:53 PM
I'll check if the project itself need it. If it does not than it's a server problem which is not a problem in itself since if we change the host for whatever reason we will have no limitation (just tomcat 7+ ) (that I'm sure! )
 
4:09 PM
hey again
 
hey
 
let me fix that bug where planets were deleted by delaying the release a bit and seeing if it appears again ;)
(Great method, please trust)
 
4:51 PM
@skiwi Much bug, many repro, such great
 
Groovy ftw
 
Nice
 
I'm having trouble figuring out what's going on though... Looks like the code under test is correct, but the hard-coded String is wrong. wat
Ah, I put the arguments in the wrong order
 
5:52 PM
@Marc-Andre ^^ any thoughts on this? Should we use a pre- and/or post-processor for CSS?
 
Css is so out of my league that I don't know! Never understood css enough to have a valid opinion
My opinion is we should do the best we can with what knowledge we have at the moment. Since CSS is not my strong suit, I say until we have someone who know what he's doing we should keep it simple
 
Are you meaning to tell me that out of all the smart devs in this room, I'm the CSS "expert"? lol
 
6:07 PM
No I'm telling that of all the people who worked on the website you're the css expert :P
 
@Phrancis Yes
 
Haha
 
I know some CSS too, but I don't really have an approach on how to create a layout
And hae never used those fancy media queries
 
I need to check out SCSS/Sass, it looks like they could make writing consistent and extensible CSS easier
 
Maybe some "real" web dev would be better in CSS than me or other smart devs from this room :P
 
6:21 PM
Yeah... @DanPantry ?
28 mins ago, by Phrancis
@Marc-Andre ^^ any thoughts on this? Should we use a pre- and/or post-processor for CSS?
 
6:39 PM
@SimonForsberg How's your NN progress?
 
Did I get pinged?
What's up?
(Can't scroll up atm, in game)
 
@DanPantry Yes you did
 
@DanPantry Hey, yes I pinged you, just wanted your opinion on this
 
Stop playing LoL
 
19 mins ago, by Phrancis
28 mins ago, by Phrancis
@Marc-Andre ^^ any thoughts on this? Should we use a pre- and/or post-processor for CSS?
 
6:41 PM
@skiwi |:
@Phrancis I would say yes
 
I don't even know what game you are playing :P
 
preprocessors give you a lot of maintainablity
I prefer SASS, personally.
If you're using webpack, SASS + CSS modules = heaven
that said postcss and LESS have their place
so it's really which one you prefer
PostCSS is much more useful for cross-browser compatibility as it autoprefixes stuff for you
And it's alos more flexible
SASS is a prettier version of LESS which is microsoft's version of SASS
@skiwi It is LoL
 
Hehe okay
 
6:47 PM
UI changed again?
 
yep
about half a year agfo
 
Looks a lot like Starcraft to me
 
@Phrancis that's funny, because Riot hired one of Blizzard's old top devs
 
7:04 PM
@SimonForsberg Did you read the whole Neural Networks chapter? Or skipped through to the algorithm ;)
I'm really not used to reading books anymore
 
@skiwi too busy today. Going to work on my assignments for the course, after I have finished discussing site business in the 2nd.
@skiwi I read it all, yes.
the whole chapter, not the whole book :)
 
I was having to have the NN coded this evening, that might have been a little bit too optimistic
 
s/having/hoping/
 
that ^^
 
@skiwi I wanted to be done with my image recognition on 3 days ago (on Sunday), that was optimistic.
I underestimated how slow my network was.
 
7:10 PM
And the fact that you don't know yet if it's correct doesn't help either
 
yeah... I need to train it much more, and to analyze it, see if the errors from what I'm expecting are reduced each iteration or not.
 
http://www.commitstrip.com/en/2016/01/13/questions-you-should-never-ask-a-coder-in-an-interview/
CommitStrip
Questions you should never ask a coder in an interview
CommitStrip
1452712530
 
 
1 hour later…
8:24 PM
@Phrancis ^^ Those are awesome cards
 
8:40 PM
Hmm @Simon reading the theory and a real brain has the notion of "neurons dying off if they are unused ultimately", maybe we could add that to a neural network to filter out unused neurons :D
 
8:50 PM
Next up after being more advanced in all this (and have coded what I wanted) is to really look into LTSM networks, as they are very often used
(Long-Term Short Memory) I believe
For your case of recognizing minesweeper flags Hopfield Neural Networks might even be more relevant (on first sight at least)
Interestingly 9.2 seems to describe what I have programmed, which is not that surprising as I got my inspiration from the lecture slides, which are based on this book
 
 
1 hour later…
10:03 PM
@skiwi I'm pretty sure that there's some learning algorithm for making the best Neural Network architecture that already does that
I haven't quite read up on the chat history here today, hope to do so tomorrow.
 
10:30 PM
@DanPantry I know, that's what I was saying. Maybe I worded my sentences wrong, though.
@DanPantry This makes sense. Also, thanks for the links!
 
10:46 PM
in The 2nd Monitor, 2 mins ago, by Captain Obvious
0
Q: Git Hook to Generate Commit Summary Based on Comments

Jacob YoungI am writing two Git hooks (bash scripts) to preform text-searching and commit message formatting, and need some review regarding best practice and simplicity of the scripts. The project uses two Git hooks (prepare-commit-msg and post-commit) to grab comments with // @commmit and automatically g...

 
@Phrancis That seems interesting.
 
10:59 PM
All these JavaScript examples I'm seeing online are in ES6. I wonder if it's time to start learning it.
@Marc-Andre I may have misunderstood "planning", but I was congratulating you on your future child.
 
@SirPython Ohh well there is no child yet but we're trying so it's close ;) !
 
That is good!
 
11:23 PM
@SirPython They're all in ES6 because most people use a transpiler to transpile from Es6 -> es5 now
also because react is written exclusively in es6
 
I've heard about a transpiler for 6, but I didn't know that React was written in Es6; that's cool!
 
11:35 PM
@Phrancis You were right as of now the project can only be hosted on tomcat 8 on Eclipse. I'm looking for why at the moment
 
200_success vs. rolfl: 6467 diff. Year: -369. Quarter: -369. Month: -369. Week: -448. Day: -375.
200_success vs. janos: 15616 diff. Year: -1154. Quarter: -1154. Month: -1154. Week: -793. Day: -405.
Loki Astari vs. Simon Forsberg: 2512 diff. Year: -154. Quarter: -154. Month: -154. Week: -129. Day: +19.
Mat's Mug vs. Simon Forsberg: 3402 diff. Year: +116. Quarter: +116. Month: +116. Week: -57. Day: 0.
 
I know why : it's because we are on Dynamic Web Module 3.1 and not 3.0
 
11:50 PM
@Marc-Andre OK. There is no workaround I assume then?
 
reading on this
Done it it's tomcat 7 friendly
Hummm the logo is not loading and I can't see why
it's a been a looong time
http://localhost:8080/io-web/-> logo
#WTF
Want me to push my modification on master @Phrancis ?
 

« first day (529 days earlier)      last day (3044 days later) »