#jsmess 2017-03-07,Tue

↑back Search

Time Nickname Message
00:00 🔗 DFJustin http://interbutt.com/temp/evpatoria-screenshot.png
00:00 🔗 DFJustin also it loses the text after reloading but not the symbol selections
00:07 🔗 db48x weird, it didn't find any matches at all
00:08 🔗 db48x it would have given them a green box if it had
00:10 🔗 db48x do you get a red box while you're selecting?
00:15 🔗 DFJustin yes
00:18 🔗 db48x well, it's not 100% broken then :P
00:23 🔗 db48x I guess in the dom you don't see anything like <div class="highlight-rect symbol-5" style="top: 35px; left: 207px; width: 16px; height: 28px; border-color: green;"></div>?
00:24 🔗 db48x it would be interesting to step through and look at whether or not matchSymbol and pixelValue are working correctly
00:25 🔗 db48x I must sleep
00:25 🔗 db48x DFJustin: but thanks for the help :)
00:25 🔗 db48x SketchCow: SlimerJS should actually be able to run the games; that's more than just "slightly" better than using PhantomJS :)
00:26 🔗 DFJustin slightly better than his current kludge of using firefox I think
00:27 🔗 db48x ah, perhaps :)
00:28 🔗 DFJustin no matches in Inspector for border-color
00:43 🔗 db48x DFJustin: I should sleep, but if you add a breakpoint in the highlight function, does it ever get hit?
00:43 🔗 db48x if it does get hit, what part of highlight fails to function correctly?
00:46 🔗 DFJustin does get hit, dunno why it doesn't work
00:47 🔗 db48x hrm. well, that narrows it down a bit
00:50 🔗 db48x does it fail to append the element to the dom, or does it mess up the class or style info?
00:51 🔗 [X-Scale] has joined #jsmess
00:52 🔗 X-Scale has quit IRC (Ping timeout: 240 seconds)
00:52 🔗 [X-Scale] is now known as X-Scale
00:53 🔗 DFJustin I'm not sure I know how to answer that, I'm not actually a web dev
00:54 🔗 bai db48x: I'm finding it hard to select accurately on my laptop
00:54 🔗 bai like the selection area seems to be offset by a few pixels from where I actually click
00:54 🔗 bai up and to the left
00:55 🔗 DFJustin oh you know what, highlight() is being called with the red selection rect but not the green found a thing rect
00:56 🔗 db48x bai: you're using chrome, and you've zoomed the page
00:56 🔗 db48x DFJustin: ah. that's harder
00:56 🔗 bai I haven't zoomed the page no, but I do have a high-dpi display
00:56 🔗 bai maybe I can offset the high-dpi by zooming :D
00:57 🔗 db48x heh
00:57 🔗 bai wwwoooaaaaaahhhhhhh
00:57 🔗 bai haha I dunno what I did now
00:58 🔗 db48x bai: I added a comment about hidpi to the bug
00:58 🔗 bai I zoomed in a bit and tried a selection and now my whole screen is filling up with grids
00:58 🔗 db48x :)
00:58 🔗 db48x you selected a very small "symbol" that probably only has white pixels :)
00:58 🔗 bai yeah that's probablyw hat happened
00:59 🔗 db48x if you delete it before it finishes finding them all, you'll see another bug I haven't fixed
00:59 🔗 bai yeah I tried deleting it but it kept going :D
00:59 🔗 db48x :)
00:59 🔗 DFJustin sym, generator, and func are undefined in highlightMatches()
01:00 🔗 db48x DFJustin: ok, that's weird
01:00 🔗 db48x go up a level on the call stack and see what's going on there?
01:01 🔗 DFJustin it's just window.setTimeout, doesn't show any code
01:01 🔗 db48x ah
01:12 🔗 db48x time to sleep
01:12 🔗 db48x bai, DFJustin: make sure you let me know when you solve it all :D
01:12 🔗 bai hah
01:13 🔗 db48x you doubt your ability to solve it?
01:13 🔗 bai I doubt my ability to solve it before ADD kicks in :P
01:14 🔗 db48x well
13:22 🔗 hook54321 has quit IRC (Ping timeout: 244 seconds)
13:45 🔗 hook54321 has joined #jsmess
20:38 🔗 Ipggi I just had a Chrome bug confirmed where as a test case a piece of ANSI art breaks its encoding engine. I find a little joy that a text encoding standard created for minicomputers in the 1970s is creating work for developers of the world's most successful web browser, some 40-years later. :)
20:45 🔗 bai nice. I did some web-based ansi stuff a few years ago, that was a huge pain
20:45 🔗 bai I think we did it all with canvas though, not text rendering
21:02 🔗 Ipggi It is a huge pain, isn't it! Particularly having to deal with its select graphics rendition control, with it's unknown lookahead requirement and then having to keep track of the its formatting, cursor positioning and display toggles.
21:57 🔗 bai yeah. it's amazing people were able to coax such beautiful art out of such basic systems, but it turns out they left us a huge complicated mess in the process :D
22:03 🔗 SketchCow Oh, there's tons of hell with it
22:04 🔗 SketchCow Example: ANSI Artists often (? occasionally) put black text on black background to slip in insults or messages
22:06 🔗 SketchCow It's super complicated.
22:27 🔗 Ipggi Yeah it is. I find it fascinating with old ANSI like this from 1988, which is nothing more than coloured text. But the writer didn't want anyone to modify it, so they obfuscated the whole text using repeated save cursor position controls and turned it into a cryptic mess
22:27 🔗 Ipggi https://defacto2.net/file/view/a8286b
22:28 🔗 bai crazy
22:31 🔗 SketchCow This is a very old situation
22:31 🔗 SketchCow but
22:32 🔗 SketchCow You will find people point to the edge cases claiming this is why modern attempts to save old things is a failure
22:32 🔗 SketchCow They're wrong
22:32 🔗 SketchCow edge cases exist to risk oblivion
22:32 🔗 SketchCow that's why they exist at all
22:37 🔗 SketchCow Also, WebAssembly officially open on Firefox
22:37 🔗 SketchCow https://blog.mozilla.org/blog/2017/03/07/lots-new-in-firefox-game-changing-webassembly-support/
22:37 🔗 SketchCow So, my attitude about webassembly -
22:38 🔗 SketchCow We did a lot of testing, thanks to bai, and we know it'll work. It's also still flaming hot new out of the box
22:38 🔗 SketchCow So while I want us to support it, I also don't want it to be a firefox only thing.
22:38 🔗 SketchCow And it also blows into the testing harness question db48x has brought up, etc.
22:39 🔗 SketchCow I think the solution is "use webassembly coming as a resons to shore up Emularity/JSMAME/otherstuff, clean it up, etc."
22:49 🔗 Ipggi I do like that on webassembly.org Use Cases page specifically mentions, Platform simulation / emulation (ARC, DOSBox, QEMU, MAME, …).
22:55 🔗 bai SketchCow: basically doing wasm would mean a second copy of each system - we'll still need to keep asm.js versions around for some time and automatically fall back when necessary
22:55 🔗 bai there's a wasm-compaible interpreter for the full binary wasm builds but it ran waaaaay slower last time we tried
22:57 🔗 bai I mean, asm.js-compatible
23:04 🔗 SketchCow Agreed
23:05 🔗 SketchCow I mean, we'd obviously have to double-set until what time it was no longer a thing to do in javascript
23:05 🔗 SketchCow Which might be never
23:06 🔗 SketchCow I just know we have a bunch of "it works" and it'd be good this year to really clean things up so we don't have panics like we did with the DOSBOX/Chrome thing
23:06 🔗 SketchCow It helps a lot we have Ipggi around witha second obvious use case.
23:09 🔗 bai yeah maybe we should get an official build system set up for the archive.org stuff, so you could just do a git pull and a rebuild rather than having to get whoever's available to build one and send it over
23:10 🔗 bai I mean, I say "system" but what I really mean is "a directory on a computer that archive.org runs"
23:11 🔗 SketchCow I'd be happy to get that
23:12 🔗 SketchCow It'll be some docker with 1tb or something
23:17 🔗 bai that should be fine
23:26 🔗 SketchCow Let's see if I survive 31 hours of air travel first
23:29 🔗 bai good luck :D when's your flight?
23:38 🔗 SketchCow 10pm my time

irclogger-viewer