--- Log for 05.01.125 Server: mercury.libera.chat Channel: #rockbox --- Nick: rb-logbot Version: Dancer V4.16 Started: 1 month and 7 days ago 00.08.06 *** Saving seen data "./dancer.seen" 00.13.17 Join dconrad [0] (~dconrad@152.117.104.232) 00.18.12 Quit dconrad (Ping timeout: 265 seconds) 00.58.07 Join dconrad [0] (~dconrad@152.117.104.232) 01.03.00 Quit dconrad (Ping timeout: 276 seconds) 01.39.13 Quit othello7 (Ping timeout: 248 seconds) 01.58.58 Join dconrad [0] (~dconrad@152.117.104.232) 02.03.17 Quit dconrad (Ping timeout: 248 seconds) 02.08.08 *** Saving seen data "./dancer.seen" 02.20.19 Quit massiveH (Quit: Leaving) 02.31.22 Join dconrad [0] (~dconrad@152.117.104.232) 02.36.36 Quit dconrad (Ping timeout: 276 seconds) 03.19.49 Join dconrad [0] (~dconrad@152.117.104.232) 03.24.22 Quit dconrad (Ping timeout: 252 seconds) 03.46.50 Join luscusrex [0] (~luscusrex@user/luscusrex) 04.05.00 Quit jacobk (Ping timeout: 276 seconds) 04.05.03 Join jacobk_ [0] (~quassel@47-186-65-73.dlls.tx.frontiernet.net) 04.07.09 Join lebellium [0] (~lebellium@2a01cb0405d07f00d33a0da62bfc2206.ipv6.abo.wanadoo.fr) 04.08.12 *** Saving seen data "./dancer.seen" 04.08.14 Join dconrad [0] (~dconrad@152.117.104.232) 04.13.09 Quit dconrad (Ping timeout: 260 seconds) 04.56.42 Join dconrad [0] (~dconrad@152.117.104.232) 05.01.26 Quit dconrad (Ping timeout: 265 seconds) 05.55.53 Join dconrad [0] (~dconrad@152.117.104.232) 06.00.07 Quit dconrad (Ping timeout: 244 seconds) 06.08.15 *** Saving seen data "./dancer.seen" 06.58.37 Join dconrad [0] (~dconrad@152.117.104.232) 07.02.53 Quit dconrad (Ping timeout: 252 seconds) 07.56.57 Join dconrad [0] (~dconrad@152.117.104.232) 08.01.36 Quit dconrad (Ping timeout: 276 seconds) 08.08.19 *** Saving seen data "./dancer.seen" 08.36.52 Join ajk [0] (~zem@97-115-88-155.ptld.qwest.net) 08.37.08 Quit hactar|ant (Ping timeout: 252 seconds) 08.57.53 Join dconrad [0] (~dconrad@152.117.104.232) 09.02.26 Quit dconrad (Ping timeout: 252 seconds) 09.32.19 Join dconrad [0] (~dconrad@152.117.104.232) 09.32.56 Join Everything [0] (~Everythin@46-133-12-19.mobile.vf-ua.net) 09.36.27 Quit dconrad (Ping timeout: 246 seconds) 09.36.35 Quit Everything (Read error: Connection reset by peer) 10.08.20 *** Saving seen data "./dancer.seen" 10.31.16 Join dconrad [0] (~dconrad@152.117.104.232) 10.35.23 Quit dconrad (Ping timeout: 245 seconds) 10.38.05 # Build Server message: 3New build round started. Revision f42343b715, 345 builds, 9 clients. 10.38.05 # 3[BugFix] Ipod 6G Classic Shutdown Hang by William Wilgus 10.49.22 # <_bilgus_> ok so Ipods 6G and classic what is a 5.5g? the video? 10.50.02 # <_bilgus_> ^ I think this bug is present on these devices too jedging by the forum post so maybe this needs expanded to whatever a 5.5G is 10.50.20 # yeah 10.50.25 # <_bilgus_> https://forums.rockbox.org/index.php/topic,55036.msg254526.html#msg254526 10.52.48 # Build Server message: 3Build round completed after 884 seconds. 10.52.50 # Build Server message: 3Revision f42343b715 result: All green 10.54.32 # <_bilgus_> well the vide has a similar command bcm_init on lcd_awake() I wonder if all lcd sleep devices should turn on lcd on shutdown 10.54.50 # <_bilgus_> I was sorta worried it might fix these and break others 11.02.46 # Build Server message: 3New build round started. Revision 734cc3e8d1, 345 builds, 9 clients. 11.02.47 # 3[Bugfix] call lcd_awake() unconditioanlly on shutdown for devices with HAVE_LCD_SLEEP by William Wilgus 11.06.08 # <_bilgus_> In theory lcd_awake() would get called anyway if there was something to display so should be fine 11.10.04 # <_bilgus_> I know this bug has been reported several times over the last few years with similar circumstances, (and irate users) hopefully that solves all of it across them 11.12.13 # hmm, could result in the display flashing awake on shutdown. which I suppose it probably does already anyway 11.12.48 # <_bilgus_> yeah but no backlight 11.13.16 # <_bilgus_> well if you set it that way.. 11.13.29 # <_bilgus_> otherwise its just business as usual 11.13.30 # I've pushed an updated lcd driver for ipod6g recently, it changes the lcd sleep/awake behaviour 11.14.15 # <_bilgus_> I really don't think this is new though and its happenig on the video too 11.14.51 # <_bilgus_> I can't find the post RN but Ive been seeing this bug report for probably 3 years now 11.15.21 # <_bilgus_> it starts with my battery runs down with device shutdown 11.15.50 # <_bilgus_> and ends with hold switch on shutdown hangs device 11.17.06 # Build Server message: 3Build round completed after 861 seconds. 11.17.08 # Build Server message: 3Revision 734cc3e8d1 result: 50 errors 0 warnings 11.17.09 # <_bilgus_> user890104, when you get a chance you want to look at https://forums.rockbox.org/index.php/topic,55159.0.html 11.17.29 # <_bilgus_> he has the steps to repro it pretty quickly 11.18.27 # <_bilgus_> If I had the device in hand I'd investigate a bit more but I think it has to do with the clocks being off 11.19.23 # it could be related to https://gerrit.rockbox.org/r/c/rockbox/+/6214 11.19.27 # <_bilgus_> its definately something to do with lcd_awake though as when I just turned on the backlight without lcd_awake it still crashed 11.20.56 # huh, lcd_awake() is defined in the target specific lcd-target.h 11.21.35 # no it's in firmware/export/backlight.h 11.21.44 # wrapped with HAVE_LCD_SLEEP 11.21.59 # so these build failres are for targets that define HAVE_LCD_SLEEP but don't export a lcd_awake() function. 11.22.11 # tk, tsk 11.22.27 # <_bilgus_> it has some weird declares that make them nops I think 11.22.48 # I don't think so, just a simple funciton prototype 11.23.58 # <_bilgus_> odd to have them in the c file but I guess they get added later and it can find them globally? 11.23.59 # presumably the solution here is to define a stub for each of these targets 11.24.13 # <_bilgus_> weak ref be easier 11.24.41 # I'm not fond of those in general, but I think it's a bad idea in this case. 11.24.57 # because this is supposed to be part of the target-specific lcd driver 11.25.19 # I'd rather we get a failure to build than let folks think their driver is complete when it's really not. 11.25.32 # <_bilgus_> I don't see anything in 6214 that makes me go hmm but I guess there could be something thats no longer getting set up proper? 11.26.08 # <_bilgus_> Ah good point, are you alreead working on this or shall I? 11.27.35 # if you don't mind, please take care of it? I just happen to be at the keyboard right now but I have to deal with some plumbing. 11.27.56 # (and can't get to this until later today) 11.28.05 # <_bilgus_> just making sure before I have to abandon something because you beat me to it :P 11.32.26 Join dconrad [0] (~dconrad@152.117.104.232) 11.36.30 Quit dconrad (Ping timeout: 246 seconds) 11.59.04 Join babagu [0] (~babagu@185.34.32.217) 11.59.10 # <_bilgus_> so I found 13 devices that have lcd_sleep and as far as I can see no lcd_awake but only 10 builds failed so its probably those 3 will fail as its a redefinition 11.59.45 # urg sorry i got disconnected before i could readn eventual reply to: "would there be a way on an Eros/Surfans to have sound coming by  the headphone and the lineout jack? considering using those players as rudimentary DJ turntables, but would be hard without this feature (or maybe splitting the lineout to 2 mono...?)" 12.06.13 # <_bilgus_> babagu, if the hardware supports it then yes but likely we do the switching in software so in theory you could disable that and make both active but you'll have to wait for someone who knows the hardware to say for sure 12.06.58 # ah i see... how to investigate that further? 12.07.04 # maybe i will ask on the forum... 12.08.22 *** Saving seen data "./dancer.seen" 12.11.22 # <_bilgus_> probably your best bet of not having to wait all day or longer 12.16.26 # Build Server message: 3New build round started. Revision 5d9b01b9ed, 345 builds, 9 clients. 12.16.26 # 3[Fix Red] define lcd_awake for targets with lcd_sleep by William Wilgus 12.16.51 # <_bilgus_> I got it down to 11/10 I'll just see what falls out 12.17.57 # the hardware can't independently drive the two outputs 12.18.03 # on the erosq 12.18.12 # <_bilgus_> babagu ^ 12.18.23 # <_bilgus_> what about the x3? 12.18.26 # same 12.18.46 # <_bilgus_> damn thats what I suspected but hadn't gotten to pullin it out 12.19.33 # single pcm output on the SoC hooked up to a single i2s stream that's fed into a single codec that feeds into a stereo switch. 12.19.47 # s/single/single stereo/g 12.20.32 # in theory we could treat it as two independent mono streams but that would takesome heavy internal surgery. 12.20.43 # <_bilgus_> so IOW buy two 12.21.15 # yep, and an external eq/fader 12.22.02 # <_bilgus_> you could mix two tracks though I wonder if the device would be fast enough to do two decodings to the stereo stream 12.23.09 Join chris_s [0] (~chris_s@2a09:bac3:2edc:2482::3a3:5b) 12.23.17 # <_bilgus_> you'd have to do it in a plugin and mix the raw pcm there to bypass our frontend that wants a single track fed by the playlist 12.23.50 # nevermind our UI really isn't set up for something like this 12.25.25 # <_bilgus_> Itd be clunky-er at best i'd rather have the two devices 12.26.41 # have a hard time seeing how a generic cheapo laptop with a pair of generic USB DACs wouldn't be vastly better in every way 12.27.22 # Hah, run two copies of rockbox with each pinned to a different device. :D 12.31.24 # _bilgus_: fwiw, I tried reproducing https://forums.rockbox.org/index.php/topic,55036.msg254526.html#msg254526 on an iPod video when it was posted, but wasn't able to / still can't, even before your latest fix, so possibly a different issue,  since iPodVT mentioned, he was also only seeing the shutdown crash on the iPod6G. 12.31.43 # <_bilgus_> I think the idea is no bulky laptop and less power hungry 12.33.02 # Build Server message: 3Build round completed after 997 seconds. 12.33.04 # Build Server message: 3Revision 5d9b01b9ed result: All green 12.34.07 # given that you already have some sort of sound system/PA going on, the "bulk" of the laptop is... negligable. 12.34.21 # (likely mains-powered too) 12.34.54 # <_bilgus_> sometimes provided by the venue though.. or hacked with a fm transmitter 12.35.52 # <_bilgus_> chris_s I don't see any harm in this calling lcd_awake (besides the targets missing it I guess) but everything else is going to be calling it if a message comes up so we just turn it on early 12.36.43 # yeah, seems to work well 12.37.04 # <_bilgus_> there might be a device that ffforces the backlight on out of lcccd_awake but I figure we'll hear about it 12.37.43 # <_bilgus_> my keyboard is doing double keys even ctrl v it is not that cold here. 12.40.03 # wow i see thanks 12.42.12 # yeah that was indeed to move places that already have a PA/soundsystem, and carry only the minimal. and two rockbox devices is what i had in mind, but if each could also provide (even mono!) output to a "phones" line, on which i could independantly listen to 1 or 2, regardless of what is coming though for the mix, that would be close to "perfect" 12.42.13 # setup for DJing! 12.42.44 # (minus how accessible the pitch shift is... the impossbility to use the wheel for fine-grained seeking, and a few things like this, but still...) 12.46.04 Join dconrad [0] (~dconrad@152.117.104.232) 12.46.32 # <_bilgus_> well the two devices I was thinking of won't be able to probably not likely for any of the others,, 12.47.19 Quit chris_s (Quit: Client closed) 12.50.30 Quit dconrad (Ping timeout: 252 seconds) 12.53.45 Quit babagu (Quit: Client closed) 12.53.54 Join chris_s [0] (~chris_s@2a09:bac2:2efb:18fa::27d:76) 13.08.31 # <_bilgus_> devices that do stuff in lcd_awake .. mrobe500 samsung yh820 sansaconnect and the ipods (6G video nano2g) 13.34.27 Join dconrad [0] (~dconrad@152.117.104.232) 13.39.36 Quit dconrad (Ping timeout: 276 seconds) 14.08.23 *** Saving seen data "./dancer.seen" 14.35.11 Join dconrad [0] (~dconrad@152.117.104.232) 14.39.27 # Build Server message: 3New build round started. Revision 00066a4927, 345 builds, 9 clients. 14.39.28 # 3Playlist Viewer: slight simplification by Christian Soffke 14.39.35 Quit dconrad (Ping timeout: 244 seconds) 14.46.55 # <_bilgus_> hmm just noticed that my device has display corruption on startup with this current theme Cleanscreen-SpinalTap and so far it does it back to at least jul 2024 I guess i'll bisect if I ever find a build where it works 14.49.19 Quit chris_s (Quit: Client closed) 14.50.12 Join chris_s [0] (~chris_s@2a09:bac3:2f84:246e::3a1:22) 14.54.34 # Build Server message: 3Build round completed after 908 seconds. 14.54.36 # Build Server message: 3Revision 00066a4927 result: All green 14.56.24 # <_bilgus_> well I guess we've made progress at the end of 2022 it would have crashed 15.00.12 Quit jacobk_ (Ping timeout: 276 seconds) 15.00.22 Join jacobk [0] (~quassel@47-186-65-73.dlls.tx.frontiernet.net) 15.25.26 # <_bilgus_> haha we must have gotten the omit plugins some time in 2020 ok so its been happening since at least the end of 2019 the theme is from 2014 it happens in 3.15 now to do the harder thing of tracking down the bug in the code 15.25.44 Quit Bobathan_ (Ping timeout: 260 seconds) 15.38.14 Join Bobathan [0] (~admin@syn-065-029-248-157.res.spectrum.com) 15.44.24 # Build Server message: 3New build round started. Revision 018ccf4e47, 345 builds, 9 clients. 15.44.24 # 3Playlist Viewer: Fix dir not updated after playlist removal by Christian Soffke 15.47.53 Quit tertu (Quit: so long...) 15.53.18 Join Everything [0] (~Everythin@195.138.86.118) 15.57.34 # Build Server message: 3Build round completed after 791 seconds. 15.57.36 # Build Server message: 3Revision 018ccf4e47 result: All green 16.02.48 # _bilgus_: you mean the theme's been technically broken since before 3.15? 16.03.19 # or are you talking about something else now? kinda hard to follow 16.08.27 *** Saving seen data "./dancer.seen" 16.12.27 Quit bpye (Quit: The Lounge - https://thelounge.chat) 16.15.52 Join dconrad [0] (~dconrad@152.117.104.232) 16.18.41 Join PheralSparky [0] (~S|h|a|w|n@user/shawn/x-4432647) 16.20.34 Quit dconrad (Ping timeout: 260 seconds) 16.30.38 Join afearfultraveler [0] (~afearfult@2a02-8440-9100-decd-46af-28ff-fe56-7863.rev.sfr.net) 16.30.41 # hey :) 16.31.16 # Today I think I corrupted all of my iPod by transferring too much using the default mode (rockbox) to transfer. 16.32.22 # Now, I can't make any changes to the hard drive. I connect it (in disk mode) to my Linux computer, mount it and try to delete my music, uninstall rockbox, etc... and when unmounting, rebooting, it just reverts to the same state 16.32.46 # It's soo weird, because when using rm -rf / Nautilus to delete file it's not giving me any errors!! 16.32.52 # Does someone has an idea :) ? 16.33.11 # Important information: 16.33.21 # - I'm using an Ipod Video 5'th Gen 16.33.32 # - I'm using a 128Gib iFlash adapter 16.34.05 # - I was able to transfer FLACs file onto it, and it suddenly stopped working. I did not know about disk mode at the time 16.34.55 # - When I click to play a file (in the file browser of Rockbox), instead of the name there's some alphanumerical (probably some binary data travelling where it shouldn't be!!) 16.35.48 # - I'm using the daily build 16.49.53 # afearfultraveler: can you try running fsck.fat on the partition? 16.54.30 # user890104, ok so; I tried something destructive because I have a back-up of everything and just decided to overwrite everything with dd :( so I can't answer sorry 16.54.35 # sudo dd if=/dev/zero of=/dev/sda bs=4M status=progress 16.55.48 # ok then, but if you eject and reconnect it, does it show the old contents? 16.55.57 # <_bilgus_> it was broken in 3.15 it doesn't crash but that display glitch is pointing to .. something.. 16.57.07 # user890104, I'll tell you when it's done, it's taking pretty long. Longer than I thought it would 16.57.20 # Build Server message: 3New build round started. Revision 907c91997e, 345 builds, 9 clients. 16.57.21 # 3[FS#10686] Peakmeter on ipod 5g using too much CPU by William Wilgus 16.57.32 # 5133828096 bytes (5.1 GB, 4.8 GiB) copied, 629 s, 8.2 MB/s ==> need to be patient, it's going to be a long night 17.05.44 Join dconrad [0] (~dconrad@152.117.104.232) 17.10.18 Quit dconrad (Ping timeout: 265 seconds) 17.11.32 # Build Server message: 3Build round completed after 853 seconds. 17.11.34 # Build Server message: 3Revision 907c91997e result: All green 17.12.58 # <_bilgus_> hmm the radio skin has something as does the sbs skin but the wps part appears to be fine 17.14.48 # <_bilgus_> oh ok I take that back its in the sbs skin of cleanscreen-spinaltap but the radio skin makes it show different garbled data 17.14.56 # <_bilgus_> guessing bufferovfl 17.16.06 # <_bilgus_> I'll track it down this eve 17.19.13 # <_bilgus_> it looks like different pieces of the skin so maybe something with the preload is bad (this skin does have massive artwork) 17.19.52 # <_bilgus_> depending on when I load it versus the other parts changes what is garbled when the sbs is loaded 17.27.07 Quit Bobathan (Ping timeout: 264 seconds) 17.28.40 Quit lebellium (Quit: Leaving) 17.36.39 Quit chris_s (Quit: Client closed) 17.36.54 Join Bobathan_ [0] (~admin@syn-065-029-248-157.res.spectrum.com) 17.36.57 Join chris_s [0] (~chris_s@2a09:bac3:2f84:246e::3a1:22) 17.38.12 Join dconrad [0] (~dconrad@152.117.104.232) 17.42.50 Quit dconrad (Ping timeout: 272 seconds) 17.46.04 Quit afearfultraveler (Read error: Connection reset by peer) 17.48.48 Quit Everything (Quit: leaving) 18.00.12 # New Windows simulator build round started. Revision 907c91997e, [FS#10686] Peakmeter on ipod 5g using too much CPU by William Wilgus 18.00.14 # https://www.rockbox.org/tracker/task/10686 3Peakmeter on ipod 5g using too much CPU (replacing FS#10527) (bugs, closed) 18.01.03 Join MarcAndersen [0] (~no_znepna@217.74.221.70) 18.01.11 Nick MarcAndersen is now known as Guest2253 (~no_znepna@217.74.221.70) 18.01.53 Nick Guest2253 is now known as MarcAndersen (~no_znepna@217.74.221.70) 18.08.29 *** Saving seen data "./dancer.seen" 18.16.34 Quit luscusrex (Quit: Leaving) 18.23.04 Join dconrad [0] (~dconrad@152.117.104.232) 18.27.28 Quit dconrad (Ping timeout: 252 seconds) 18.41.32 Quit chris_s (Quit: Client closed) 18.42.32 Join chris_s [0] (~chris_s@2a09:bac2:2e57:2478::3a2:54) 18.47.57 # Build round completed after 2865 seconds. 19.07.57 Join dconrad [0] (~dconrad@152.117.104.232) 19.12.11 Quit dconrad (Ping timeout: 252 seconds) 19.45.13 Join tertu [0] (~tertu@user/tertu) 19.49.19 Join massiveH [0] (~massiveH@2600:4040:a982:5400:edfe:9497:6e75:fa57) 20.08.08 Join othello7 [0] (~Thunderbi@pool-100-36-176-164.washdc.fios.verizon.net) 20.08.31 *** Saving seen data "./dancer.seen" 20.08.49 Join dconrad [0] (~dconrad@152.117.104.232) 20.13.04 Quit dconrad (Ping timeout: 252 seconds) 20.28.51 # <_bilgus_> this is weird the miniml reproducible sbs doesn't even load a image its '%V(0,0,-,-,-)%VB%Vf(FFFFFF)%Vb(000000) 20.28.51 # <_bilgus_> # set UI viewport 20.28.52 # <_bilgus_> %Vi(Z,0,14,96,-,-)%Vf(FFFFFF)%Vb(000000) 20.28.52 DBUG Enqueued KICK _bilgus_ 20.28.52 # <_bilgus_> %VI(Z)' 20.29.44 # <_bilgus_> those three lines corrupt the display on load 20.30.56 # could be wrong, but doesn't the VI have to come before the Vi ? 20.34.08 # or maybe not 20.34.34 # <_bilgus_> no says i then I 20.35.02 # <_bilgus_> i wonder if its a one off in the - code 20.35.15 # <_bilgus_> default '-' 20.38.59 # is that supposed to work though? Looking at the SBS it's filling the whole viewport with an image and then tries to draw the UI viewport on top of it ? 20.39.54 # oh I missed the %VB 20.41.15 Join dconrad [0] (~dconrad@152.117.104.232) 20.45.52 Quit dconrad (Ping timeout: 265 seconds) 20.47.45 # <_bilgus_> its not a one off error 20.48.13 # <_bilgus_> might be a more than one off error I suppose maybe its not drawing the bg color 20.48.40 # i haven't looked at the code, but in the docs it says you need to add %V(0,0,-,-,-) when switching to the foreground layer: https://www.rockbox.org/wiki/CustomWPS.html#Viewport_Draw_Layer 20.49.19 # <_bilgus_> so maybe its misplaced? 20.49.53 # I think both is needed, fullscreen backdrop viewport and then a fullscreen foreground layer 20.55.09 # on the other hand that causes flashing of the screen. idk, my impression is that there's still lots of undefined behavior when it comes to using the backdrop layer in SBS - I've tried to avoid using that in my own themes 20.56.23 # Didn't really dare look at the implementation for that 20.58.11 # thar definitely be dragons 21.05.17 # <_bilgus_> yeah so they are missing the second one to switch back 21.08.56 # <_bilgus_> lol thats it 21.09.23 # <_bilgus_> so is there a way to fix it without altering the theme 21.12.03 # <_bilgus_> we know it can't be anything other than %xd or %V(0,0,-,-,-) 21.12.29 # <_bilgus_> so if its something else slip in %V(0,0,-,-,-) 21.27.21 Join dconrad [0] (~dconrad@152.117.104.232) 21.33.43 Quit dconrad (Ping timeout: 245 seconds) 21.50.19 # <_bilgus_> ok so I can detect if you try to call something other than xd Vf OR Vb and fail the theme load 21.51.23 # <_bilgus_> I think I can do similar in the viewport convert function but just add a vp if its not a full screen one 21.58.44 # <_bilgus_> otherwise I guess it will fail checkwps and we will know who else forgot.. 22.05.16 # <_bilgus_> yeah the issue is that I don't have a way to add an element in the middle 22.05.34 # <_bilgus_> I can detect the error though and fail the load 22.06.51 # <_bilgus_> I guess thats the right thing to do, I can probably skip the other tags part thats less likely to cause issues 22.08.34 *** Saving seen data "./dancer.seen" 22.22.26 Join dconrad [0] (~dconrad@152.117.104.232) 22.57.03 Quit dconrad (Remote host closed the connection) 22.59.56 Join dconrad [0] (~dconrad@152.117.104.232) 23.07.39 # _bilgus_: I don't think you can be quite as conservative as g6315 - that would break  FreshOS, Adwaitapod, Themify.., 23.11.40 # <_bilgus_> the issue is I can't really insert a new element without going up the parser stack so rather than display corruption what recourse do we have except to break them and then fix them? 23.13.00 # isn't this disallowing *any* further VB viewports though even after they've inserted a foreground viewport, which should actaully be allowed? 23.13.26 # <_bilgus_> no 23.14.46 # <_bilgus_> curr_vp->output_to_backdrop_buffer is the determination after that your curr->vp is the fg 23.15.06 Quit massiveH (Quit: Leaving) 23.17.27 # <_bilgus_> I'll see if I can write something to the buffer to at least black it out because other than the display corruption on load when it gets going it seems ok 23.18.01 # <_bilgus_> aliased pointers something something.. 23.18.37 # <_bilgus_> but the asan stuff isn't catching iot so might be fine otherwise 23.22.22 # <_bilgus_> use_extra_framebuffer maybe I can black it out ahead of time 23.23.21 Quit jj5 (Quit: Konversation terminated!) 23.23.33 # <_bilgus_> oh it doesn't look like it does anything :/ 23.32.50 # re 6315, once another VB viewport appears it becomes the curr_vp though, doesn't it? So if after V(0,0,-,-,-) on the foreground layer you have another background viewport at some point, you are then forced to follow it up with another SKIN_TOKEN_VIEWPORT_LOAD ? 23.38.07 # The docs seem to imply that it's fine to draw further viewports on the background even after %V(0,0,-,-,-) (" Any viewports you want drawn on the backdrop need the %VB tag" in step 3) unless I'm reading that incorrectly 23.42.50 # <_bilgus_> oh see I was reading it as doing that after the seconv %V 23.43.00 Quit dconrad (Remote host closed the connection) 23.44.01 Join dconrad [0] (~dconrad@152.117.104.232) 23.48.21 Quit dconrad (Ping timeout: 246 seconds) 23.57.28 # <_bilgus_> OK so how about this? the issue is that they are loading a uninitialized background buffer what if I mark empty buffers and set them to 0 before I hand them back? 23.58.02 # <_bilgus_> ultimately the display corruption is superficial