[13:35] *** Sgeo has quit IRC (Read error: Connection reset by peer) [13:35] *** Sgeo has joined #jsmess [14:01] *** Sgeo has quit IRC (Read error: Connection reset by peer) [14:01] *** Sgeo has joined #jsmess [16:17] *** azakai_ has quit IRC (Read error: Operation timed out) [18:46] I just proposed the visor [18:47] There is like about the vistor [18:47] visor [18:47] I should reach out to defacto2 [18:47] Also will annoy db48x abut this [18:59] I have updates on the PCE/macplus fullscreen bug. I am able to reproduce it locally, using a copy of pcejs that I downloaded off of archive.org, and a slightly-modified copy of Emularity's example_macplus.html. Replacing pcejs with a custom-built copy, appears to fix it, although I am unsure exactly why. [18:59] dll: Thanks for chasing it down [19:03] If y'all could use any further help with that bug, I can try to look into it further. Otherwise, I'll assume that someone needs to recompile it and upload it. [19:04] Side note: pcejs does not compile out-of-the-box with the latest Emscripten. I have a hack-ish patch to pcejs' build scripts that fixes it, or at least it does for me. :-) [19:07] (Emscripten no longer supports the '-s BINARYEN_METHOD='suchandsuch'' parameter. Instead, it takes in a combination of the following: '-s ASM_JS=N', '-s WASM=N', whereby N is either 0 (disable) or 1 (enable)) [19:07] That setting is specified in pcejs' build file, scripts/pcejs_build_env.sh [19:08] oh interesting, how recent is that change? I was just trying to use that the other day, and the docs certainly still mentioned it [19:10] It appears to have been dropped in Emscripten v1.38.23. Emscripten's ChangeLog.md lists that as being released on 01/10/2019 [19:10] I saw it in the docs too, which threw me off for a while. :-/ [19:11] Also, trying to use -s ASM_JS=0 was leading to other errors. -s WASM=0 worked fine, though. [19:19] Who did the compilation of the pce.js we use now [19:21] dll: Can you give me a link to the pce-macplus.js you generated? I can snap it in and see if something explodes. [19:22] I think you have to generate the asmjs version either way, then if you're building wasm, the asmjs version gets converted to wasm [19:23] There's no wasm of the pce for what it's worth [19:23] Like, we never did that [19:23] oh ok [19:24] I mean, I'm all for doing it later [19:25] But if dll hands us a replacement .js I'm up for shoving it in [19:25] And keeping a backup of the old one. [19:27] SketchCow: it has some extra printf-debug code in there. Is that ok for now? [19:28] For now. [19:28] I expect you'll make a refined version the more you look at it. [19:28] Ok. One moment... [19:39] https://www.dropbox.com/s/rdsp4qltkekq5y5/pce-macplus-from-DavidL-1.zip?dl=0 [19:40] That's an asmjs only version. It does add a .js.mem file, to note. [19:40] ... which I vaguely recall there being a way to embed into the .js file. I can try that, if you like. [19:40] ... or give a wasm version. [19:41] bai can sanity check. [19:43] I mean, it's better when it can all be in one [19:44] So if possible, yes [19:44] Ok, one moment... [19:45] Here's a single-file, asmjs-only version: https://www.dropbox.com/s/qgmwlogeomgldpe/pce-macplus-from-DavidL-2.zip?dl=0 [19:51] zopfli'ing the crap out of it [19:51] Ha, your version is smaller. [19:52] 250k to the current-in-use's 260k [19:54] Cool, although, I am unsure exactly why. ;-) [19:54] Ok, either we've destroying Macintosh functioning or we made it better [19:55] Is that just in regards to the file-size reduction, or have other issues come up? [19:58] You are encountering my sense of himor [19:59] Ok :-) [20:00] So, the good news is you fixed the fullscreen issue. [20:00] Bad news is it appears whatever the system does, a bunch of items where it's not a hard drive but a floppy disk fail. [20:00] ugh [20:00] I mean, not bad, it didn't explode. [20:01] If you point me at an example, I can try to diagnose [20:01] Otherwise, I'll pick something at random :-) [20:02] Sure, one moment. [20:02] https://archive.org/details/mac_Shuffle_Puck fails [20:02] Ok. I'll start digging with that. Thanks! [20:05] https://archive.org/details/AppleMacintoshSystem753 works [20:05] That uses a .img [20:05] https://archive.org/details/mac_Shuffle_Puck uses a .dsk [20:06] Also, the frame rate seems to have dropped, perhaps there's a missing compile signal - that seems to be what the console is saying. [20:06] "Looks like you are rendering without using requestAnimationFrame for the main loop. You should use 0 for the frame rate in emscripten_set_main_loop in order to use requestAnimationFrame, as that can greatly improve your frame rates!" [20:09] I saw that earlier. I wonder why the old copy didn't bring that up. Might a fork of pcejs have been getting used? [20:10] I used the copy from https://github.com/jsdf/pce as a starting point [20:12] Interesing... just had breach attempts from the firefox. [20:12] Firefox Monitor: malformed breach entry. [20:12] Site: [20:12] {"Name":"ElasticsearchSalesLeads","Domain":"","schema":1547745032683,"PwnCount":5788169,"AddedDate":"2018-11-17T09:04:54Z","BreachDate":"2018-10-29","id":"30af4ecc-4d08-4d51-8d47-3213e4abec18","last_modified":1547745048641,"_status":"synced"} FirefoxMonitor.jsm:180 [20:12] Firefox Monitor: malformed breach entry. [20:12] Site: [20:12] {"Name":"CrimeAgencyVBulletin","Domain":"","schema":1547745057658,"PwnCount":942044,"AddedDate":"2017-03-21T03:12:40Z","BreachDate":"2017-01-19","id":"4eb74f39-32ce-4f9f-b69b-3a4aac1a6751","last_modified":1547745063355,"_status":"synced"} FirefoxMonitor.jsm:180 [20:12] Firefox Monitor: malformed breach entry. [20:12] Site: [20:12] {"Name":"MasterDeeds","Domain":"","schema":1547744994293,"PwnCount":2257930,"AddedDate":"2017-10-18T11:01:46Z","BreachDate":"2017-03-14","id":"6e7daef1-57f8-4171-a37e-6a705ac2f72c","last_modified":1547745000600,"_status":"synced"} FirefoxMonitor.jsm:180 [20:12] Firefox Monitor: malformed breach entry. [20:12] Site: [20:12] {"Name":"YouveBeenScraped","Domain":"","schema":1547745092571,"PwnCount":66147869,"AddedDate":"2018-12-06T19:11:27Z","BreachDate":"2018-10-05","id":"846aa4bf-98e9-4ac7-96c1-a6a3b62db690","last_modified":1547745316104,"_status":"synced"} FirefoxMonitor.jsm:180 [20:13] Firefox Monitor: malformed breach entry. [20:13] Site: [20:13] {"Name":"Estonia","Domain":"","schema":1547745025686,"PwnCount":655161,"AddedDate":"2018-06-11T09:41:17Z","BreachDate":"2018-06-07","id":"874f9748-d504-49c9-9510-c0d995fe8abe","last_modified":1547745032674,"_status":"synced"} [20:13] Quickly seeing if it's us, or my firefox sucks [20:14] huh [20:14] Ah, I think firefox is just being a boop [20:14] Ignore [20:14] It's just horseshit [20:14] Ok [20:14] Anyway, all the hypercard stacks work, so that's good [20:15] I'm wondering who did the compiling of the pce.js for us. [20:15] I mean, without a doubt - it's been over a year and a half since the last compile. [20:22] My attitude is this. [20:23] It "mostly" works. We get full-screen back and it's absolutely working for a majority of items. That means we're on a very good track. [20:23] I can leave this brokey version up for a day while we sort this out. I suspect there's some optimizations/fixes you're missing, as well as the problem with .dsk files [20:23] I can hear it's slower, probably due to the problem it's mentioning [20:24] I'll alert db48x this is going on [20:24] I'm still digging, on my end. More to come... [20:24] ... regarding floppy disk loading [20:51] Ok, this is bizarre. pcejs's C code is flat-out not able to read that floppy disk. It fopens it ok, but fails at reading it. [20:53] I have a copy of non-js PCE around here somewhere. I'll try throwing the floppy image at that, and see if it loads. [20:53] Would you like a copy of the other js code? [20:53] The one we were using? [20:55] If it's the C source code to it, then yes, please. :-) [20:55] Ha no [20:56] I did find non-js PCE. I'll see if it behaves differently. [20:57] I'm curious if it's using the same, or similar functions, to load the disk. [20:57] Or if it loads the disk at all. [20:58] I asked db48x to stop in when he has time. He might have insights. [20:58] I can also check the logs to see who was doing the compiling. [21:00] History seems to show it was me and bai fucking around [21:03] non-js PCE does load that disk file. I'm digging further... [21:07] I'm actually not sure if I ever compiled pce, I don't think I have any of the roms or anything for that [21:07] I feel like that was one of our visitors [21:15] The non-pcejs code that I have, which does load floppy disks, has a disk loader that pcejs (from https://github.com/jsdf/pce) does *not* have. [21:16] I.e. there are functions in non-js PCE that do not exist in pcejs, which look important to loading .dsk files. [21:17] I'm really wondering where the old build's code came from. :-) [21:17] ... and what else might be missing [21:18] if I recall, it was forked some time back [21:19] Ok [21:20] I'm going to break for a while. If the location of that fork's source code shows up, I'd be happy to try taking a look at it and compiling from that. [21:22] Before I go, I did find this fork just-now: https://github.com/db48x/retroweb-pcejs-jsdf [21:23] It appears to have the .dsk loader :-) [21:24] Maybe that's the one? [21:46] That'd be it [21:46] That's db48x [21:46] Go for it