**** BEGIN LOGGING AT Sun Jun 24 03:00:00 2018 Jun 24 05:37:08 rah, I don't have X27 device to check personally, but it was reported that selecting default SIM fixes it, not sure why or if this is just a coincidence, but it could write to nvdata/nvcfg partitions Jun 24 07:35:31 gobuki, thx a lot never the less ;) Jun 24 07:49:05 unfortunately the guide at http://support.planetcom.co.uk/index.php/Linux_Flashing_Guide is not quite working out for me because the flashing tool is showing some error messages Jun 24 08:01:28 on start there is this one: https://kek.gg/i/7jcLpt.png Jun 24 08:02:54 then i get 3 of this in a row: https://kek.gg/i/xzZgW.png when after selecting a scatter loading file with the choose button Jun 24 08:03:00 and then the file is not set Jun 24 08:03:59 on the terminal is output following while flash_tool is running: http://paste.debian.net/hidden/e0b03e36/ Jun 24 08:04:05 any ideas on that? Jun 24 08:04:09 thanks in advance Jun 24 09:41:15 Why can't I update the kernel if Gemini boots Linux by default? Seems to stop booting if I do. Jun 24 09:41:39 No problems if Android is default boot target... Jun 24 09:42:14 is that with writing to the 'boot' partition from the flash tool or from 'dd' within linux? Jun 24 09:47:58 I've used the flash tool Jun 24 09:49:28 given that they offer a debian only option it seems odd that that would fail Jun 24 09:51:03 can you try some of the kernel debugging options: https://github.com/gemian/gemini-keyboard-apps/wiki/Debugging#kernel-debugging Jun 24 09:51:10 to see why it stopped Jun 24 10:15:42 ttim: i might be a permission problem. Generally the procedure works like this: Start the flashing tool, make the selections, click start download. Then plug in your powered off gemini. The flash tool should find it and start the flashing process. Jun 24 10:16:05 ttim: you're welcome ;) Jun 24 10:57:36 permission problem? Jun 24 10:58:17 i am running the flash_tool as root and i have entered the lines into /etc/udev/rules.d/20-mm-blacklist-mtk.rules and restarted udev Jun 24 10:58:36 so where else could be a permission problem? Jun 24 10:59:17 its debian testing amd64 where i try to run the flash tool... Jun 24 11:15:50 NotKit: I'd be extremely surprised if Android settings had any effect on Sailfish OS Jun 24 11:16:10 NotKit: do you have a source for this report? a URL? Jun 24 11:53:52 messages in chat in Russian Jun 24 12:03:57 hmm, cot anybody else an wrong crc on the rooted system.img? Jun 24 12:04:29 inflating: Gemini_x27_12062018/system.img bad CRC 46b85459 (should be a6be8492) Jun 24 12:04:31 *got Jun 24 12:08:07 from http://support.planetcom.co.uk/download/LinuxFirmware/gemini_x27_base.zip Jun 24 12:19:47 oh, shit - next download try, got another crc32! Jun 24 12:19:55 now I get a bit nervous Jun 24 12:20:16 the other files could be extracted flawless - even the 4.4GB linux file Jun 24 12:23:38 Can anybody else try to download & extract this file? Jun 24 12:27:30 NotKit: I stand corrected: setting the SIM as the default has made Sailfish OS work Jun 24 12:28:07 by which I mean, disabling SIM 2 and setting the default SIM for data, calls and sms to SIM 1 Jun 24 12:33:22 ok, 3 downloads, 3 different crcs Jun 24 12:35:11 can you check the checksums on the downloaded files? Jun 24 12:35:24 md5sum: 30c437b8d699bfd97d527fdca45d89c4 00e5d53d795ae34e6397e3a8b2b34ffe Jun 24 12:35:36 from the 2nd and 3th download Jun 24 12:35:43 (already killed the 1st) Jun 24 12:36:03 http://support.planetcom.co.uk/download/LinuxFirmware/MD5Sums.txt Jun 24 12:36:25 seems like your download is failing in odd ways Jun 24 12:36:27 I get the same crc while extracting (from the same file), so something must be wrong on the download side Jun 24 12:36:57 try using a different download tool, eg wget or some such Jun 24 12:37:34 ok, the debian_sailfishos.zip has a other md5 as well, but because unzip worked I assume that the file was updated Jun 24 12:37:43 tried already firefox and chrome Jun 24 12:38:29 maybe someone else who's downloaded it in the last day or so can tell us their checksums to see if that planet txt file is accurate Jun 24 12:38:31 I feel like I've run emacs for the first time Jun 24 12:38:39 how do I power off from Sailfish OS? Jun 24 12:38:54 trying downloading from my dedicated server... Jun 24 12:42:52 ok, from my server the debian-image is ok Jun 24 12:42:54 wtf... Jun 24 12:44:13 downloading the debian+sailfishos now Jun 24 12:44:56 who is playing nasty tricks to me?^^ Jun 24 12:51:13 rah, cool, now at least we have this confirmed, thank you Jun 24 12:51:28 I guess eSIM is implemented differently Jun 24 12:52:41 ok, the md5 from debian_sailfish is also correct if downloaded to my dedic server Jun 24 12:52:56 despite being able to extract in flawlessly Jun 24 12:53:09 /me is getting _very_nervous Jun 24 12:53:20 shows you not to trust downloads without checking checksums Jun 24 12:53:49 tcp _has_ checksums Jun 24 12:54:04 but without ssl... Jun 24 12:54:35 but why should somebody temper these files - and in a way which corrupt them? Jun 24 12:54:37 though they could just as easily have modified the checksum file if they were really trying to mess with you Jun 24 12:54:52 yes... Jun 24 12:55:09 that at least should be on https Jun 24 12:55:10 and all of the hw here is under my control (if not hacked) Jun 24 12:55:36 try downloading them from your dedicated server Jun 24 12:56:10 yes - but this server has only ssh and https *g* so if this gets wrong I've some severe problem on my laptop Jun 24 12:56:19 which would be very, very bad Jun 24 12:57:31 NotKit: kind of academic for me now though; having booted Sailfish OS, I've discovered that you need to sign up for an account with Jolla in order to receive OS updates Jun 24 12:57:34 fuck that Jun 24 12:57:54 that's wrong for community devices Jun 24 12:58:38 (you won't be able to update through UI even with Jolla account, but updating through OBS works manually) Jun 24 12:58:52 if I got into the settings menu and go to the OS-updating bit, it opens a screen telling me I need a Jolla account Jun 24 12:58:59 I don't know what you mean by OBS Jun 24 13:00:08 Open Build Service as used by openSUSE Jun 24 13:00:09 http://repo.merproject.org/obs/nemo:/testing:/hw:/planet:/geminipda/ Jun 24 13:03:07 there are a lot of things about sailfish that scare me Jun 24 13:03:23 when I first heard of it, I had high hopes, but I think I heard it's basically a fork of maemo with non-free bits on top Jun 24 13:03:38 so something like pureos (the thing purism made for the librem 5) would be preferable Jun 24 13:04:53 it's not really a fork of Maemo, but agree about proprietary bits Jun 24 13:05:21 well my point was more that they're putting non-free parts on top of a mostly free core, and I'd rather not compromise if possible Jun 24 13:05:36 tune: supposidly they are upstreaming all their stuff (purism) so we'll get it all in debian once they work through Jun 24 13:05:36 you could argue that Android already has a very similar issue, but I don't think making something just as bad as Android is the right thing to do Jun 24 13:05:51 adam_b_: ah that's good to know Jun 24 13:06:04 I like the sound of their out-of-box experience, with the matrix messaging and such Jun 24 13:06:12 and it seems to be a properly FSF-endorsed distro unlike plain debian Jun 24 13:06:21 we should be able to port PureOS once it materializes, especially if they continue to use wlroots for their Wayland compositor Jun 24 13:06:35 awesome Jun 24 13:06:47 well it sounded like pureos is just debian plus bits that they re putting back so its really not much of a port Jun 24 13:06:53 (sway with wlroots already runs on Gemini, but I tested on Arch) Jun 24 13:07:12 oh, you've used sway on the gemini? Jun 24 13:07:30 someone told me to try it recently but I didn't get around to it yet. I have i3 installed and the panel is screwed up, so I can't see the time and such Jun 24 13:07:44 not exactly used, more like started adding backend for wlroots to render through hwcomposer Jun 24 13:10:16 I'm not familiar with wlroots or hwcomposer, but I think I've got a vague idea of what you're on about... Jun 24 13:10:26 is sway useable right now then? perhaps I'll try it Jun 24 13:11:34 adam_b_: "error:1408F119:SSL routines:ssl3_get_record:decryption failed or bad record mac" while downloading the files from the dedicated server - ok, something is really doing nasty things on me... Jun 24 13:12:09 sigh... anyone else have fairly frequent problems with connman? the gemini didn't connect to wifi automatically, so I just launched cmst, went to wireless, chose my network, and hit connect. I'm getting an error now. "We received a DBUS reply message indicating an error. Error Name: net.connman.Error.InProgress Error Message: In progress". Jun 24 13:12:19 maybe it's saying that I was trying to connect while it was already trying to connect on its own Jun 24 13:12:31 tune: seems likely Jun 24 13:12:48 it's still not connected, though, so I'm not really sure what I'm supposed to do here Jun 24 13:13:21 sometimes the cmst ui dosn't update very quickly, try actually using the connection or close/reopen cmst Jun 24 13:13:35 I'm going into a terminal on the site and running ping to see if it's working Jun 24 13:13:43 s/site/side Jun 24 13:13:54 wtf - my network card on my router is going mad - which could explain everything (because of tcp offloading) Jun 24 13:14:24 oh weird... I tried to manually connect again and this time it didn't give an error, and the symbol under "connected" changed. Jun 24 13:14:34 still can't ping, and now going back to connman I'm not connected Jun 24 13:14:59 might be some useful errors in journalctl -f --unit=connman Jun 24 13:15:23 but sounds like your getting bounced on your wifi Jun 24 13:15:47 tune, not on stretch at least Jun 24 13:16:52 oh no... Jun 24 13:17:12 I have this recurring issue where I see "sudo: unable to resolve host iori: connection timed out" when trying to use sudo Jun 24 13:17:38 I saw the wifi symbol in the corner so I tried a ping again but was met with failure, and now the symbol is back to something with a red x on it Jun 24 13:17:56 journalctl appears to be actively outputting stuff, it hasn't returned my shell yet Jun 24 13:18:33 weird, the gemini just put up its lockscreen Jun 24 13:18:39 it definitely wasn't 5 minutes since last input Jun 24 13:19:15 from the journalctl I see in orange "Time request for server 192.168.1.1 failed (101/Network is unreachable) Jun 24 13:19:37 adam_b_: ethtool -K eth0 tx off rx off helped for me :-) Jun 24 13:19:39 the debian experience has been a mess the whole time, but usually I can get on my network after a little bit of struggling... Jun 24 13:20:44 tune: well its a TP, but for your host issue you did add your new host name to your /etc/hosts file? Jun 24 13:21:07 I don't think I did add it Jun 24 13:21:30 also, just managed to get on my 2.4GHz network after struggling with the 5GHz one. I had to enter the passphrase for the other one so I definitely hadn't used it before Jun 24 13:21:32 probably other times you've set a hostname its been through a setup tool that did that for you Jun 24 13:21:53 hm yeah probably Jun 24 13:22:02 I generally set a hostname during an OS install rather than changing it later Jun 24 13:22:05 thanks for the tip Jun 24 13:22:22 the -f on journalctl is to follow Jun 24 13:22:45 the idea being that you follow in one tab and then do the action in the other, then flip back to the first to see the output Jun 24 13:22:54 ah I see Jun 24 13:23:27 the unit just filters down to connman, might be worth trying without that too to see what else is going on Jun 24 13:25:01 weird... I'm getting network errors, but I'm connected to the gemini over ssh somehow Jun 24 13:25:30 like from ssh I can't ping gnu.org or 8.8.8.8 Jun 24 13:25:32 but I'm connected somehow Jun 24 13:25:42 usb ssh? Jun 24 13:25:46 no Jun 24 13:25:59 so connections in are working just not out? Jun 24 13:26:16 yeah, I guess. I was able to get in again Jun 24 13:26:32 ip route might tell you things Jun 24 13:28:06 maybe it's on my wifi but not connected to the internet. I can ping other local computers by IP (but not hostname) Jun 24 13:29:56 "apt search" goes through also Jun 24 13:31:09 but 'sudo apt update' fails to connect Jun 24 13:31:55 I'm just gonna turn it off and set it aside for now. I wasn't gonna do much with it right now anyway. Jun 24 14:18:35 NotKit: You mentioned Arch, do you have that running on the Gemini? Jun 24 14:21:13 also very interested in any progress on wayland, hopefully it will make things feel more hardware accelerated Jun 24 14:21:42 i've been trying to optimise various parts of the x11 pipeline (with the goal of making firefox smooth) but keep hitting dead ends Jun 24 14:29:16 Darkimmortal, I am pretty interested in your attepts though Jun 24 14:29:54 Darkimmortal, for testing - yes, but not for everyday use (lots of stuff Gemian currently has is not packaged) Jun 24 14:31:54 ah fair enough, very interested to use Arch if/when it's available, it would be my distro of choice Jun 24 14:32:59 the hot path for x11 scrolling seems to be glamor_upload_boxes>glTexSubImage2D, driver is probably swizzling or something, trying to get my head around changing this to use some sort of mapped texture Jun 24 14:33:37 so basically glamor isn't particulary optimized for mobile hardware? Jun 24 14:34:49 seems to be the case, at least for the fork of glamor that's in use Jun 24 14:35:11 we can probably rebase it on latest if this was fixed Jun 24 14:35:20 i tried building mainline glamor in es2 mode but didn't get very far beyond successful compilation Jun 24 14:35:25 probably needs hooked into libhybris in some way Jun 24 14:36:44 is mode without glamor more performant for e. g. Firefox or scrolling? Jun 24 14:37:29 i've not actually tried that, but i guess possibly Jun 24 14:38:45 dumb question: how can I use/activate the root rights on a rooted android? I find nothing in the settings, and superso (or similiar) isn't installed, too... Jun 24 14:39:16 mifritscher: install magisk manager from XDA Jun 24 14:39:48 ok Jun 24 14:44:58 Darkimmortal, but as for Wayland, the problem is that basically each compositor's code needs to be adapted, as we lack DRM/GBM Jun 24 14:45:21 kwin has hwcomposer backend already, it was fast, but not very stable last time I tried Jun 24 14:48:01 NotKit: ah didn't know kwin had a hwcomposer backend, interesting... Jun 24 14:48:26 how do i disable glamor? tried removing --enable-glamor-hybris in xf86-video-hwcomposer but i've just got a black screen + cursor Jun 24 14:49:01 sddm will be invisible, but should work Jun 24 14:49:35 Option "AccelMethod" "none" I think Jun 24 14:49:47 ah yes you're absolutely right, typed password blind and i'm in Jun 24 14:50:41 wow this is night and day faster! firefox is actually usable!! Jun 24 14:52:47 Darkimmortal are you using plasma mobile? Jun 24 14:53:09 no this is debian + lxqt/kwin Jun 24 14:53:15 ahhhh Jun 24 14:57:20 is there a way to set the global dpi in debian? I've found the text dpi, but this is not the culprit Jun 24 14:58:14 for my taste all the symbols are _huge_^^ Jun 24 15:01:18 NotKit: that's not really OBS, that's Mer's OBS Jun 24 15:01:44 NotKit: how does one update the OS from OBS? Jun 24 15:03:05 I'm aware of what Open Build Service is, I helped with the Debian packaging Jun 24 15:03:23 tune: FYI, PureOS was developed for Purism's laptops, long before the phone campaign Jun 24 15:03:50 hm okay Jun 24 15:04:04 almost mentioned purism instead of the librem 5 but it made more sense to me to compare the librem 5 and gemini Jun 24 15:04:11 and I guess I wasn't sure if pureos existed before Jun 24 15:04:32 also, PureOS will likely never run on the Gemini because of the binaries Jun 24 15:04:46 PureOS won't ship anything that isn't free software Jun 24 15:05:27 so in order for PureOS to run on it, you would need to reverse engineer all of the firmware blobs for the hardware Jun 24 15:05:42 rah, or fork PureOS and make it run with blobs Jun 24 15:05:55 even though first option is better Jun 24 15:06:07 NotKit: well you shouldn't need to, everything will be going upstream anyway Jun 24 15:06:11 or just build the pureos bits you like on on gemian Jun 24 15:06:12 so just use Debian :- Jun 24 15:06:13 :-) Jun 24 15:06:24 PureOS is a Debian derivative Jun 24 15:06:44 but with an obsolete kernel :/ Jun 24 15:07:01 fakefur: what do you mean? Jun 24 15:07:17 i mean the debian for gemini has an obsolete kernel not pure Jun 24 15:07:53 fakefur: what do you mean by "obsolete" and what do you mean by "not pure"? Jun 24 15:08:02 * fakefur sighs Jun 24 15:08:06 rah, like in https://wiki.merproject.org/wiki/Adaptations/libhybris/Install_SailfishOS_for_fp2#Over-the-Air_updates_.28OTA.29 Jun 24 15:08:14 pureos does not have an obsolete kernel Jun 24 15:08:31 debian for gemeni does have an obsolete (3.18) kernel Jun 24 15:08:53 fakefur: what do you mean by "obsolete" and what do you mean by "not pure"? Jun 24 15:09:11 rah can you read? Jun 24 15:09:22 fakefur: yes Jun 24 15:09:31 i answered your question Jun 24 15:09:38 I disagree Jun 24 15:09:45 ok lets try again... Jun 24 15:10:09 NotKit: So I think based on this, glamor-hybris is a net loss at the moment except for facilitating the XRender backend of KWin, might actually be best to remove/disable it from the distribution Jun 24 15:10:09 pureos does not have an obsolete kernel - i was not talking about it - your comments came as i was typing Jun 24 15:10:18 the usability with no compositor and no glamor is a huge step up Jun 24 15:10:32 the debian for gemini does have an obsolete (as in not supported and out of update cycle) kernel Jun 24 15:10:40 3.18 and not even the .48 Jun 24 15:10:48 Darkimmortal, but without it EGL rendering is mostly broken, which e.g. chromium benefits frome Jun 24 15:10:57 *from Jun 24 15:11:01 fakefur: I see Jun 24 15:11:27 rah sorry for the confusion Jun 24 15:11:41 NotKit: Ah yes it does completely break chromium :( Jun 24 15:12:36 fakefur: There's only one serious (read: remote) vuln in the gemini 3.18 kernel, patch here: https://github.com/lukefor/gemini-linux-kernel-3.18/commit/ec7eff2fc90538e08e9b6cbb902e75422e017d5d - if you keep bluetooth off it's irrelevant Jun 24 15:13:01 Darkimmortal but it's a *really* old kernel Jun 24 15:13:19 yeah no doubt, it would be nice to have a newer one Jun 24 15:13:30 that GKH says to avoid like the plague Jun 24 15:13:44 Darkimmortal, could you PR it? Jun 24 15:15:30 NotKit yeah will do Jun 24 15:20:32 Hi, loving the Gemini :) Jun 24 15:21:05 I was looking building some new graphical stack including a desktop of some sort Jun 24 15:21:37 Is there anything like DRM or FB working that I can use to draw? Jun 24 15:22:10 FB is working, but generally for accelerated rendering you would want to use HWComposer through libhybris Jun 24 15:22:19 ajwillia_ms, what kind of stack though? Jun 24 15:22:50 NotKit: using EFL libraries to do the render - building on top of that in Go Jun 24 15:23:56 will you able to launch regular apps in that stack though? Jun 24 15:24:34 I’ll be supporting wayland Jun 24 15:25:31 https://github.com/NotKit/wlroots/commits/master - wlroots is able to render with some changes Jun 24 15:25:40 10:05:27 < rah> so in order for PureOS to run on it, you would need to reverse engineer all of the firmware blobs for the hardware Jun 24 15:25:43 this is what I want Jun 24 15:26:10 to work on revese engineering blobs? Jun 24 15:26:16 not exactly those words Jun 24 15:26:17 I've got the time but not the skills to do it myself Jun 24 15:26:22 I want someone to do it Jun 24 15:26:23 I want it done Jun 24 15:26:52 not to say that I am telling anyone here to do it, just that if I were to say what I wanted between that option and the others, of course I want the first option Jun 24 15:26:57 a fully free experience would be ideal Jun 24 15:27:17 ah, sure, but this is really heavy task Jun 24 15:27:37 I know it can be difficult. I run GuixSD on my main laptop, but I can't libreboot it. Jun 24 15:27:45 so I'm not as free as I'd like to be Jun 24 15:27:48 tune: first is probably just getting mainline with blobs kernel going Jun 24 15:32:29 adam_b_ how can you use 3.18 blobs in, say, a 4.x kernel? Jun 24 15:33:22 I'm guessing you don't, but as mainline kernel has patches submitted by Mediatec they've presumably put in ones as appropriate Jun 24 15:34:49 adam_b_ would the blobs not need to be recompiled for the later kernel model though? asking as i genuinely don't know Jun 24 15:35:56 yea I don't know either, I presume that the blobs get loaded into the hardware as 'firmware' and then there is opensource stuff around that that links to the kernel API's that can then eovlve to some degree of flexibility Jun 24 15:36:58 I'm not actively involved in that project, but it dosn't sound like its something people are working on Jun 24 15:37:40 adam_b_ some blobs are firmware for sure but others are more at the driver level and they would have to be recompiled to my understanding Jun 24 15:38:06 presumably mainline kernel won't take patches unless they can keep them going for a bit Jun 24 15:40:32 adam_b_ i think the issue with mediatek is they don't give a f*** about that which is why the kernel for their soc's is stuck on 3.18 Jun 24 15:57:25 they do submit some code to mainline (https://patchwork.kernel.org/project/linux-mediatek/list/), but there are no support for lots of mt6797 parts Jun 24 15:58:23 essentially they are a crappy company to use for a linux device Jun 24 15:59:00 which is why some people (myself included) feel somewhat mislead by the "linux palmtop" advertising planet did Jun 24 16:04:44 tune: I'm guessing that will be highly unlikely Jun 24 16:04:55 tune: given how proprietary mediatek is Jun 24 16:06:39 adam_b_: I expect there will be all of: proprietary firmware, proprietary libraries called by free software Android drivers, and proprietary Android drivers Jun 24 16:07:54 and only in rare cases to have fully free software drivers Jun 24 16:09:38 there are some people upstreaming mediatek stuff, including mediatek employees Jun 24 16:10:40 but I epxect that'll be all for the SoC itself, as opposed to devices like GPS or audio codecs or the modem or whatever Jun 24 16:11:01 as NotKit says, the mt6797 support is still in very early days Jun 24 16:11:46 the 96boards X20 Development Board is apparently being upstreamed by some Linaro guy in his spare time.. Jun 24 16:11:47 they did upstream audio codecs for mt6797 pretty recently Jun 24 16:12:02 but modem, graphics, etc is going to be a problem Jun 24 16:12:06 NotKit: well that's good to hear, I stand corrected :-) Jun 24 16:12:08 aye Jun 24 16:12:14 there's just loads of stuff Jun 24 16:12:22 i think for the forseeable future the gemini is, in reality, an android device plus a linux tinkerer device Jun 24 16:12:36 hopefully sailfish 3 makes it a good half way between the two worlds Jun 24 16:13:03 the accelerometer, magnetometer, proximity detector, LEDs, etc. etc. Jun 24 16:13:34 i wish planet had used qualcomm and charged a bit more Jun 24 16:13:44 or arm Jun 24 16:13:50 but they didn't Jun 24 16:13:51 arm? Jun 24 16:14:09 well a decent processor that has decent linux support Jun 24 16:14:18 not from a company like mediatek Jun 24 16:14:20 but Qualcomm itself won't be much better unless it's same SoC as Dragonboard Jun 24 16:15:03 NotKit sure but there are soc suppliers who do have open policies Jun 24 16:15:09 fakefur: err.. the X27 is an Arm SoC Jun 24 16:15:20 tune: FYI, in case you're interested :-) https://puri.sm/posts/primer-to-reverse-engineering/ Jun 24 16:15:20 rah yes sure Jun 24 16:15:35 the blob bits Jun 24 16:15:51 Problem being that the "open" SoCs mostly don't have the interesting bits. Like LTE modems. Jun 24 16:15:58 but anyway, they used mostly ready board design, not builded a new one with custom SoC and rest separately Jun 24 16:16:09 yes Jun 24 16:16:11 It's unlikely you're ever going to get the source for the firmware for a commercial LTE modem Jun 24 16:16:14 thanks for the link, I'll take a look, although I can't even really program yet Jun 24 16:16:19 i guess this is the only way it could be made currently Jun 24 16:16:53 i would be happy with a properly working lineageos + microg Jun 24 16:17:05 then at least google becomes optional Jun 24 16:17:17 fakefur: There's the Librem approach I guess - discrete components for every major system, with as many of them as possible being open, and the units isolated from each other so the baseband controller doesn't have system memory access and such like. Jun 24 16:17:27 But your BoM becomes a *lot* more expensive when you do that. Jun 24 16:17:35 dapol2 yes Jun 24 16:18:25 mifritscher: this link https://bpaste.net/show/9867fac6f59b Jun 24 16:18:30 mifritscher: from here: https://github.com/gemian/gemini-keyboard-apps/wiki/LinuxMainlining#status-of-mainlining-x27 Jun 24 16:18:34 mifritscher: gives a 404 Jun 24 16:19:00 I'm not going to fix it because I don't know what the link was going to Jun 24 16:19:12 Maybe we should poke Planet and see if they'll work with Purism - Librem 5 wiht a Gemini Keyboard :) Jun 24 16:19:35 oooh Jun 24 16:19:36 yes Jun 24 16:19:38 I love that idea Jun 24 16:19:48 my biggest complaint with the librem 5 was the lack of a keyboard Jun 24 16:20:04 it's "just another phone". it REALLY nails it with the freedom and privacy aspects, but it's still just a touchscreen slate Jun 24 16:20:20 one step at a time :-) Jun 24 16:20:32 just having a free phone is a massive achievement Jun 24 16:20:39 :) Jun 24 16:21:01 of course, if the Librem 5 is a success, I'm sure a phone with a keyboard will be very high on the list of successors Jun 24 16:21:23 Of course, the followon for PureOS on a phone should be a version of QubesOS :) Jun 24 16:21:24 cause.. you know.. what geek doesn't want a phone with a keyboard? :-) Jun 24 16:21:50 rah: Me! I want the "Do what I meant, not what I said" precognisant neural interface :) Jun 24 16:22:39 Which would also do away with the problems I have with small displays - yay fading vision :) Jun 24 16:23:53 Hmm - modular cellphone - not Ara level modular, but a common bus for dropping in display, processor and I/O modules - you too can have a flagship class mobile with a Microwriter chording keyboard :) Jun 24 16:24:12 So, more Jolla OtherHalf than anything else Jun 24 16:24:34 i'd buy the librem 5 if it comes to market Jun 24 16:25:06 fakefur: I don't think there's an "if" - more a "when" Jun 24 16:25:33 dapol2 until it's a "is" it's a "if" for me Jun 24 16:25:34 :) Jun 24 16:25:49 * dapol2 remembers a friend who had his Zaurus C760 set up for "Maximum privacy" Jun 24 16:25:56 fakefur: what do you mean by "if it comes to market"? Jun 24 16:26:11 I can't help thinking that having a full featured mail server with remote polling written in BASH was taking the piss a little, but it worked. Jun 24 16:26:15 https://puri.sm/shop/librem-5/ Jun 24 16:26:28 rah sorry but you seem to be asking questions repeatedly for no apparent reason - why? Jun 24 16:27:19 fakefur: have you considered the fact that someone keeps asking you to clarify what you mean is an indicator that your communication could be improved, rather than an indicator that there is a problem with others? Jun 24 16:27:29 "estimated shipping jan 2019" isn't "available to buy now" Jun 24 16:27:42 rah no not really Jun 24 16:27:44 I see Jun 24 16:29:17 * dapol2 spots the Sailfish Watch. Hmm..... Jun 24 16:29:40 Sailfish Watch? Jun 24 16:29:42 *facepalm* Jun 24 16:30:34 https://blog.jolla.com/watch/ Jun 24 16:33:04 * dapol2 wants... well, many things. Jun 24 16:33:19 But honestly, the userspace is bigger for me than the backend these days Jun 24 16:34:06 If I could convince Blackberry Hub to actually provide me a timeline with all interactions from all media sources then I'd be happy - with my texts, IM's e-mails timelined with all of my social media feed in a single linear timeline. Jun 24 16:34:25 Not gonna happen, because Facebook are category A arseholes with an aversion to linear timelines. Jun 24 16:35:30 Hmm - BB10 on Gemini? Jun 24 16:35:39 Oh Blackberry!!!!!!!! Jun 24 16:41:14 rah: hmpf, then the nopaste got deleted :( Jun 24 16:55:26 yeah, blackberry an the Hub, that was (is) something. I think there is Android version of the Hub too, but I do not know how much functional that is. Jun 24 18:12:34 I updated the firmware on my Gemini and had the backups stored on SD but the SD was configured as extension storage - lesson learned the storage was inaccessible after the firmware upgrade Jun 24 18:18:05 I only wanted ext formatted storage instead of fat32 Jun 24 18:19:13 beiriannydd: so the SD card became an encrypted store that you've now lost the keys for in the upgrade? Jun 24 18:19:41 correct... Jun 24 18:20:18 it isnt too big a deal Jun 24 18:20:34 but good to learn early on in the device ownership Jun 24 18:20:53 precisely Jun 24 18:21:29 so now I am left wondering how I achieved some things Jun 24 18:22:06 apt doesnt seem to work under linuxdeploy any more Jun 24 18:22:15 problems with dns lookups Jun 24 18:22:32 ping and wget work just fine Jun 24 18:24:48 also I thought that I had managed to get shell -u gemini working before but now it doesn't have access to the ptty, so I have a workaround with starting as root but using su gemini as the shell Jun 24 18:25:47 I am guessing it is a privileges related thing Jun 24 18:32:01 I thought it might be that I needed to add _apt to the android groups Jun 24 19:01:45 hmm, where can I undo the "make everything huge" in geminis debian? Jun 24 19:04:21 you mean the scale factors: https://github.com/gemian/gemian-multistrap-config/blob/master/configs/etc/profile.d/hidpi.sh Jun 24 19:05:40 solved the apt problem, it wasnt enough to give _apt the 3003 group as an additional group, it had to be the default group Jun 24 19:08:32 ah here, thanks! Jun 24 19:13:58 ok, a bit better^^ Jun 24 19:14:01 but only a bit... Jun 24 19:16:04 now I need to find where the fonts got this big Jun 24 19:16:23 have a look around that repository, its got all the changes over standard debian Jun 24 19:16:35 ah, ok Jun 24 19:19:22 ok, the dpi setting is not honoured Jun 24 19:22:50 changing the screen size doesn't help either Jun 24 19:23:08 but it must be set somewhere *g* Jun 24 19:24:56 xterm is fine Jun 24 19:30:22 ok, QT_SCALE_FACTOR could help - but if I set it to 0.5 things go completely mad Jun 24 19:33:47 dpi could be force-hacked either in /usr/bin/Xorg or /etc/sddm.conf Jun 24 19:35:07 the system settings say 96 dpi - and does not help to change it. fiddling with the display size in xorg.conf doesn't help either. I'm afraid that qt gets the size of the display "somewhere" Jun 24 19:36:07 ok, in sddm.conf, EnableHiDPI=false Jun 24 19:36:08 well, I wrote above where 192 dpi gets set :) Jun 24 19:36:17 cat /usr/bin/Xorg Jun 24 19:37:18 if it's not the case in TP2, then it should be set in sddm Xorg cmdline Jun 24 19:37:30 yep, nothing in TP2 Jun 24 19:37:53 yep, it is set in the command line from Xorg Jun 24 19:39:14 ah, found it Jun 24 19:39:32 nice, thanks! Jun 24 19:42:07 just halfed it (to 96) Jun 24 19:42:14 now I can work with the machine *g* Jun 24 19:43:21 (somehow, alt+space does not work to open the "min/max/close" menu of the apps, a can't navigate to it from a open menu as well) Jun 24 19:53:16 hmm, the main SoC seems behind a black film. I'm wondering wether it could be worth to remove it and add a heat conductor to it Jun 24 20:06:32 anyone know what config file controls the screen timeout? Jun 24 20:22:20 Darkimmortal: on Debian? it uses xss-lock so you'd change the times with xset Jun 24 20:39:11 ok irssi is setup now :) Jun 24 20:59:26 adam_b_: hmm no joy with either 'xset s off' or 'xset -dpms', is there anything else it might be? Jun 24 21:05:50 you'd probably need to kill and restart xss-lock after making the change **** ENDING LOGGING AT Mon Jun 25 03:00:03 2018