| --- Log opened Thu Oct 11 00:00:43 2018 |
| 00:09 | | Degi [Degi@Nightstar-v0f4s6.dyn.telefonica.de] has quit [Connection reset by peer] |
| 00:38 | | Kindamoody is now known as Kindamoody[zZz] |
| 01:05 | | Derakon[AFK] is now known as Derakon |
| 01:26 | | Reiv [NSkiwiirc@Nightstar-ih0uis.global-gateway.net.nz] has joined #code |
| 01:26 | | mode/#code [+o Reiv] by ChanServ |
| 01:37 | | himi [sjjf@Nightstar-1drtbs.anu.edu.au] has quit [Ping timeout: 121 seconds] |
| 01:53 | | himi [sjjf@Nightstar-vo9.7ns.56.130.IP] has joined #code |
| 01:54 | | mode/#code [+o himi] by ChanServ |
| 03:21 | | himi [sjjf@Nightstar-vo9.7ns.56.130.IP] has quit [Ping timeout: 121 seconds] |
| 03:21 | | himi [sjjf@Nightstar-1drtbs.anu.edu.au] has joined #code |
| 03:21 | | mode/#code [+o himi] by ChanServ |
| 04:02 | | Vornicus [Vorn@ServerAdministrator.Nightstar.Net] has joined #code |
| 04:02 | | mode/#code [+qo Vornicus Vornicus] by ChanServ |
| 05:04 | | Derakon is now known as Derakon[AFK] |
| 05:09 | | Vorntastic [uid293981@Nightstar-6br85t.irccloud.com] has joined #code |
| 05:09 | | mode/#code [+qo Vorntastic Vorntastic] by ChanServ |
| 05:10 | | Vornicus [Vorn@ServerAdministrator.Nightstar.Net] has quit [Ping timeout: 121 seconds] |
| 06:04 | | himi [sjjf@Nightstar-1drtbs.anu.edu.au] has quit [Ping timeout: 121 seconds] |
| 06:41 | <&McMartin> | Man, legacy code |
| 06:41 | <&McMartin> | I'm tracing through the SDL video code for managing the window size, and the following values are part of the canvas struct, whether directly or indirectly |
| 06:41 | <&McMartin> | width, real_width, actual_width, draw_buffer_width, canvas_width, visible_width, canvas_physical_width |
| 06:42 | <&McMartin> | Note that under most code paths and configurations real_width and actual_width are not actually the same value |
| 06:51 | <&McMartin> | Oh yeah, there's also custom_width, but that's unrelated to rendering and the user can't actually customize it at runtime |
| 07:18 | | Vornicus [Vorn@ServerAdministrator.Nightstar.Net] has joined #code |
| 07:18 | | mode/#code [+qo Vornicus Vornicus] by ChanServ |
| 07:19 | | Vorntastic [uid293981@Nightstar-6br85t.irccloud.com] has quit [[NS] Quit: Connection closed for inactivity] |
| 07:40 | | Vornicus [Vorn@ServerAdministrator.Nightstar.Net] has quit [Connection reset by peer] |
| 07:51 | | Vorntastic [uid293981@Nightstar-6br85t.irccloud.com] has joined #code |
| 07:51 | | mode/#code [+qo Vorntastic Vorntastic] by ChanServ |
| 08:47 | <&jeroud> | Yay Legacy Code. |
| 08:48 | <&jeroud> | SDL is a disaster. I believe (you'd know much better than me) that SDL2 is less of a disaster. |
| 09:57 | <&McMartin> | This isn't SDL itself, it's code that's using it |
| 09:58 | <&McMartin> | SDL1 in fact seamlessly solves the problem I think five of those widths are attempting to address |
| 09:58 | <&McMartin> | (That of wanting a fullscreen display resolution that the underlying OS does not support) |
| 09:58 | <&jeroud> | Ah. |
| 09:59 | <&McMartin> | This issue cannot even be stated in SDL2's display model, because SDL2 does not provide a framebuffer |
| 10:06 | <&McMartin> | There is a legit need for three different widths when tracking multiple simultaneous scaled aspect-corrected nonsquare-pixel screens. |
| 10:14 | | Degi [Degi@Nightstar-e4puca.dyn.telefonica.de] has joined #code |
| 10:31 | | Kindamoody[zZz] is now known as Kindamoody|afk |
| 10:51 | <&jeroud> | Having thought only briefly about the problem, I'm a little surprised that it's only three. |
| 11:33 | | Emmy [Emmy@Nightstar-9p7hb1.direct-adsl.nl] has joined #code |
| 12:41 | | Vorntastic [uid293981@Nightstar-6br85t.irccloud.com] has quit [[NS] Quit: Connection closed for inactivity] |
| 13:03 | | gnolam_ [lenin@Nightstar-ego6cb.cust.bahnhof.se] has joined #code |
| 13:04 | | gnolam [lenin@Nightstar-ego6cb.cust.bahnhof.se] has quit [Ping timeout: 121 seconds] |
| 13:26 | | Vorntastic [uid293981@Nightstar-6br85t.irccloud.com] has joined #code |
| 13:26 | | mode/#code [+qo Vorntastic Vorntastic] by ChanServ |
| 13:31 | | gnolam_ is now known as gnolam |
| 13:32 | | mode/#code [+o gnolam] by ChanServ |
| 16:10 | | Vornicus [Vorn@ServerAdministrator.Nightstar.Net] has joined #code |
| 16:10 | | mode/#code [+qo Vornicus Vornicus] by ChanServ |
| 16:21 | | Degi [Degi@Nightstar-e4puca.dyn.telefonica.de] has quit [Connection reset by peer] |
| 17:13 | | * ToxicFrog pokes lanterna and/or sshd with a stick. |
| 17:13 | <&ToxicFrog> | This is weird. |
| 17:13 | <&ToxicFrog> | So, at startup lanterna (curses++ for java, basically) does a bunch of terminal initialization stuff, which includes sending DSR 6 (cursor position report) to the terminal and reading the result to figure out the terminal size. |
| 17:14 | <&ToxicFrog> | At the point where it does this, the terminal is supposed to already be in noecho mode, so this is invisible to the user. |
| 17:14 | <&ToxicFrog> | If I run a program using lanternal locally, this works fine. |
| 17:15 | <&ToxicFrog> | If I ssh into my work desktop or my personal laptop and do the same thing, ditto. |
| 17:15 | <&ToxicFrog> | If I ssh into my server, the reply to the DSR gets echoed to the terminal and then once it finishes starting up a bunch of other stuff is busted. |
| 17:15 | <&ToxicFrog> | If I ssh to the server and then from there into the laptop and then try it, it still works fine. |
| 17:16 | <&ToxicFrog> | Not tried yet: logging in to the server at its physical tty and running it there. |
| 17:16 | <@Tamber> | hooray terminfo muckups? |
| 17:17 | <&ToxicFrog> | But so far the common thread is that if it runs on the server, it's busted, and on anything else, even via ssh, it works fine. |
| 17:17 | <&ToxicFrog> | Changing TERM doesn't seem to help, and everything else on the server that does curses-style stuff like tmux, weechat, htop, etc works fine. |
| 17:18 | <&ToxicFrog> | So I don't think the problem is terminfo. |
| 17:18 | <&ToxicFrog> | They're all running the same version of lanterna, clojure, etc because that's all set in project.clj. |
| 17:18 | <&ToxicFrog> | They don't all have the same JRE version. |
| 17:20 | <@Tamber> | hmm |
| 17:21 | <&ToxicFrog> | (it's two different 1.8 versions on the desktop and the server, 1.10 on the laptop) |
| 17:21 | | Vorntastic [uid293981@Nightstar-6br85t.irccloud.com] has quit [[NS] Quit: Connection closed for inactivity] |
| 17:21 | <&ToxicFrog> | They all have the same Leiningen version. |
| 17:22 | <&ToxicFrog> | Oh, and this happens on the server whether or not it's being run inside tmux. |
| 17:22 | <&ToxicFrog> | And they all have $LINES and $COLUMNS set correctly. |
| 17:34 | < [R]> | Seems like they wouldn't read those values anyways though |
| 17:35 | <&ToxicFrog> | [R]: I was wondering if something was going on like, say, it reads those at startup and only falls back to DSR6 if they aren't set. |
| 17:35 | < [R]> | Fair enough |
| 17:36 | < [R]> | Can you copy the desktop JDK/JRE over to the server (even just into your ~) and run the test with the desktop JRE? |
| 17:37 | < [R]> | The nice thing about Java on Linux is it's all self-contained :) |
| 17:39 | <&ToxicFrog> | The server is Nix so maybe but I wouldn't bet on it |
| 17:39 | <&ToxicFrog> | It might be easier to install 1.8.0.181 on one of the other machines and see the problem appears. |
| 18:49 | | Kizor [moryok@Nightstar-e0a4sm.utu.fi] has quit [Ping timeout: 121 seconds] |
| 18:49 | | Kizor [moryok@Nightstar-e0a4sm.utu.fi] has joined #code |
| 20:32 | | Kindamoody|afk is now known as Kindamoody |
| 21:16 | | Degi [Degi@Nightstar-e4puca.dyn.telefonica.de] has joined #code |
| 22:45 | | Reiv [NSkiwiirc@Nightstar-ih0uis.global-gateway.net.nz] has quit [[NS] Quit: http://www.kiwiirc.com/ - A hand crafted IRC client] |
| 22:47 | | Reiv [NSkiwiirc@Nightstar-ih0uis.global-gateway.net.nz] has joined #code |
| 22:47 | | mode/#code [+o Reiv] by ChanServ |
| 22:48 | | Emmy [Emmy@Nightstar-9p7hb1.direct-adsl.nl] has quit [Ping timeout: 122 seconds] |
| 23:02 | | Vornicus [Vorn@ServerAdministrator.Nightstar.Net] has quit [Ping timeout: 121 seconds] |
| 23:06 | | himi [sjjf@Nightstar-1drtbs.anu.edu.au] has joined #code |
| 23:06 | | mode/#code [+o himi] by ChanServ |
| 23:11 | | macdjord [macdjord@Nightstar-grpbnp.mc.videotron.ca] has joined #code |
| 23:11 | | mode/#code [+o macdjord] by ChanServ |
| 23:13 | | macdjord|slep [macdjord@Nightstar-grpbnp.mc.videotron.ca] has quit [Ping timeout: 121 seconds] |
| 23:17 | | Kindamoody is now known as Kindamoody[zZz] |
| 23:29 | <&McMartin> | jeroud: Roughly: width of the rendering context, width of the quad being rendered, width of the texture being mapped to the quad |
| 23:29 | <&McMartin> | Everything else (arguably, even the width of the quad being rendered, but you do that enough you want to cache it) can be derived on the fly as needed. |
| 23:32 | <&McMartin> | SDL2 hangs on to some of the values for you, but if you want to swap textures/configurations within a single window (which you do, for the C128 driving two logical monitors but only one physical one) you need to keep the necessary information stored in the application's context |
| 23:57 | <&McMartin> | An oldie but still a classic: https://i.imgur.com/QpISccR.jpg |
| --- Log closed Fri Oct 12 00:00:04 2018 |