| mntmn | glad to hear! | 00:00 |
|---|---|---|
| esad | thanks a lot for building this. so many details: the transparent bottom, the mechnical keyboard switches, the way screen lights through on the edges of the lid, the rubber latch :) | 00:01 |
| esad | it really made my day | 00:01 |
| esad | now, the questions :) | 00:01 |
| erlehmann | mntmn, the explore link on https//source.mntmn.com goes to https://source.mntmn.com/MNT – but is it possible to list all repos? | 00:01 |
| erlehmann | esad, as before, i suggest to comment out the fan control in reformd and use this: https://source.mntmn.com/erlehmann/reformfand | 00:02 |
| erlehmann | (mntmn knows if it is still necessary to comment something out) | 00:02 |
| esad | I just started reading a book titled "Feedback Control" :) | 00:03 |
| esad | is anything on default reform system reading the state of the hall sensor? when I close the lid nothing happens? | 00:04 |
| mntmn | erlehmann ahem | 00:04 |
| erlehmann | mntmn, maybe you improved reformd? | 00:04 |
| esad | what wifi card do you recommend? do I need to lay out my own antenna for it inside? | 00:04 |
| erlehmann | esad, the hall sensor should be read by reformd, but read its source, maybe the shutdown / standby is commented out. | 00:05 |
| esad | (maybe I'm not the best beta user - I've been taking computer for granted for too long - never even had to plug in my own wifi card into a laptop :) | 00:05 |
| erlehmann | esad, also the hall sensor randomly triggers when reform has low voltage AFAIK, so that means your computer shuts down then (which is appropriate, but infortunate) | 00:05 |
| erlehmann | unfortunate | 00:06 |
| mntmn | ok one after the other | 00:06 |
| erlehmann | mntmn knows the wifi card | 00:06 |
| mntmn | esad 1. i would recommend an ath9k or ath10k based minipcie card | 00:06 |
| erlehmann | i support this recommendation, but be aware that antenna placement matters a lot for low numbered channels. | 00:07 |
| specing | Doesen't ath10k require fimrware? | 00:07 |
| esad | any tips on antenna placement? do these things come with an antenna or do I have to order it extra? | 00:08 |
| mntmn | and yeah you have to figure the antenna out yourself, i’m using a pretty small one that has a sticky underside which you can put anywere in the case | 00:09 |
| mntmn | normally they don’t come with one | 00:09 |
| mntmn | i didn’t have the impression that placement is terribly important | 00:09 |
| mntmn | now to the lid | 00:10 |
| erlehmann | i have the impression that antenna placement is important because i could get more range with an ath5k (in a thinkpad T60) than with an ath9k (in REFORM) and the difference is antenna placement | 00:10 |
| mntmn | with the same antenna and the same watts and the same case material? | 00:11 |
| erlehmann | esad, even without an antenna you should be able to get 50% WLAN on channel 36 or so | 00:11 |
| esad | are there any that also have bluetooth? | 00:11 |
| esad | ok, so I can't really make it worse by bad antenna placement | 00:11 |
| erlehmann | mntmn, of course not, but if it is case material, shouldn't the antenna be outside the case? | 00:11 |
| erlehmann | esad, you will see WLAN on channel 5 though and not be able to connect | 00:12 |
| mntmn | wifi/bt: some intel ones i think, probably nonfree but i haven’t looked to closely | 00:12 |
| erlehmann | are low channels a/b and high channels g/n? | 00:12 |
| mntmn | about the lid: the feature is unfinished/experimental | 00:13 |
| mntmn | that’s why it is disabled by default | 00:13 |
| swivel | i've been very happy with my usb edimax ew-7811un for wifi, no binary blobs and it cost like $12 | 00:13 |
| erlehmann | can we collect such stuff on the hardware tips site? | 00:14 |
| mntmn | erlehmann yes we can | 00:14 |
| erlehmann | HARRY POTTER | 00:14 |
| erlehmann | O | 00:14 |
| erlehmann | B | 00:14 |
| erlehmann | A | 00:14 |
| erlehmann | M | 00:14 |
| esad | does anyone know if I can use the wifi card out of my thinkpad x230 for starters? these should be mpcie too? | 00:14 |
| erlehmann | A | 00:14 |
| erlehmann | :D | 00:14 |
| mntmn | erlehmann you will be banned | 00:14 |
| swivel | plus it's isolated on usb so i don't have to trust the firmware embedded in it to not be malicious like something on pcie... | 00:14 |
| erlehmann | mntmn, oh i'll stop ;_; | 00:14 |
| erlehmann | please don't | 00:14 |
| mntmn | last warning | 00:14 |
| erlehmann | ok | 00:15 |
| mntmn | thanks | 00:15 |
| esad | swivel: but that's one usb port less, no? :) | 00:15 |
| esad | (I'm coming from 12" Macbook, so glad to see usb ports - don't want to give them away :) | 00:15 |
| swivel | esad: yes, i'm on an old thinkpad x61s which has 3 | 00:15 |
| erlehmann | i have good experience with putting a 7 port USB hub on reform | 00:16 |
| mntmn | esad, maybe the wifi card has a model number? | 00:16 |
| esad | @mntmn so there's a flag in reformd somwhere to enable it? | 00:16 |
| erlehmann | it needs its own power supply to give power to ALL ports at once, though | 00:16 |
| esad | @mntmn yeah I'll get it out and see what it says on the label | 00:16 |
| mntmn | esad yeah it’s a commented out section | 00:17 |
| erlehmann | esad, look at the shutdown procedure in reformd, it is there | 00:17 |
| erlehmann | btw i have not been able to wake reform up from sleeping, maybe i am doing it wrong | 00:17 |
| mntmn | but beware, it has a suspend command in it... so i would replace that by logging something and test if it works correctly | 00:17 |
| erlehmann | i think the suspend command can be replaced by echoing or printfing something to /sys/power/state | 00:18 |
| erlehmann | there should be “freeze mem disk” at least | 00:18 |
| mntmn | the problem with the sensor is that it should be powered by 5v but it currently isn’t, so the values are not reliable enough | 00:18 |
| erlehmann | but disk does not work, because you need swap for it | 00:18 |
| erlehmann | i think the values are reliable enough if reform is powered externally or if the battery is full enough | 00:19 |
| mntmn | we can work on an update at 35c3 | 00:19 |
| erlehmann | we should, yes. i commit to it! | 00:19 |
| mntmn | true erlehmann | 00:19 |
| erlehmann | mntmn, since i removed the shutdown on hall sensor feature in my variant of reformd, several times reform just ran until the battery was empty | 00:19 |
| erlehmann | because i fell asleep or something | 00:20 |
| esad | regarding disk: are all ssds nowadays created equal? can I just get any and it will work? | 00:20 |
| erlehmann | with the sensor it would have shut down (albeit erroneously, but i would have liked it, undervolting is not fun or so i heard) | 00:20 |
| esad | do I need to enable TRIM or something or is this handled by the SSD? | 00:20 |
| mntmn | esad i tested samsung evo 850 and 860 so far | 00:20 |
| mntmn | this i don’t know | 00:20 |
| mntmn | in theory all msata ssds should work | 00:21 |
| mntmn | but i faintly remember a firmware bug or sth in the 1 TB evo | 00:21 |
| mntmn | erlehmann btw i think the 5v booster shuts down before it gets dangerous | 00:22 |
| mntmn | also it will be shutdown if attiny is below something considered 3.3v high logic level | 00:22 |
| esad | btw. I noticed today that the battery status was showing 100+% when I plugged the AC adapter | 00:23 |
| esad | it went up to 146% or so | 00:23 |
| mntmn | esad yes known bug | 00:23 |
| mntmn | it is very silly | 00:23 |
| esad | and the manual says that I need to connect the cable myself - but it was already connected? | 00:23 |
| mntmn | yes | 00:23 |
| esad | ok :) | 00:23 |
| esad | is it safe to leave it overnight charging? | 00:23 |
| mntmn | i figured that it would be actually more convenient to have the battery already connected ^^ | 00:23 |
| esad | haha, yes | 00:24 |
| mntmn | esad i don’t assume any responsibility for damage to your property ^^ | 00:24 |
| esad | otoh it would be kind of nice onboarding :) | 00:24 |
| esad | just asking for a prior probability distribution :) | 00:24 |
| mntmn | i left it charging overnight several times | 00:24 |
| mntmn | i had only 1 battery failure so far while developing the thing (and it was outside of the device) | 00:25 |
| mntmn | and the battery just bloated up | 00:25 |
| erlehmann | mntmn, good to know | 00:25 |
| erlehmann | i never had the high battery bug | 00:25 |
| mntmn | i left it then several days on the balcony in a metal bucket and nothing happened | 00:25 |
| erlehmann | i think the bloating is from hydrogen or something? | 00:26 |
| mntmn | hmm? | 00:26 |
| mntmn | source? | 00:26 |
| esad | any reason for not going with usb-c for charging? is it more expensive? | 00:26 |
| mntmn | esad, complex | 00:26 |
| mntmn | but you can probably hack a usb-c charger onto it somehow | 00:27 |
| mntmn | but afaik the power transfer stuff of usb-c also involves some sort of negotiation protocol | 00:27 |
| erlehmann | mntmn, what was the battery technology again? | 00:27 |
| mntmn | whereas now everything is pretty much analog | 00:27 |
| esad | yeah I just recently discovered that usb-c charging is not so simple - there's negotiation and cable sensing and weird stuff | 00:27 |
| mntmn | erlehmann, LiFePO4 | 00:28 |
| erlehmann | https://www.quora.com/What-would-cause-a-phone-battery-to-bulge-in-the-middle?share=1 | 00:28 |
| erlehmann | > Bulging batteries mean only one thing- buildup of gas inside. The gases are produced due to electrochemical oxidation of the electrolyte. Such oxidation occurs usually due to overcharging of the battery due to a faulty battery, or faulty charging electronics in the phone or battery charger. | 00:28 |
| mntmn | afaik it is much harder to set LiFePO4 on fire than LiPo or LiIon | 00:28 |
| erlehmann | yeah, this is about lithium ion batteries | 00:28 |
| esad | how much power is it drawing at most? | 00:28 |
| mntmn | esad, for charging? | 00:29 |
| mntmn | or the system power? | 00:29 |
| erlehmann | https://www.doityourself.com/stry/why-a-rechargeable-battery-becomes-bloated | 00:29 |
| mntmn | erlehmann LiFePO4 is very different from other battery tech | 00:29 |
| erlehmann | yeah, i am already impressed by you using it! | 00:30 |
| esad | I'm not sure how it works - when the external charger is connected, it's charging the battery and powering everything else? | 00:30 |
| mntmn | this video is very informative https://youtu.be/hB5ml_OouO4 | 00:30 |
| mntmn | esad yes | 00:30 |
| esad | so max load is dependend on the system usage too right? | 00:30 |
| mntmn | yes | 00:30 |
| esad | I'm just wondering what the maximum observed values are | 00:30 |
| mntmn | the charging is capped at 1.5A | 00:30 |
| esad | like screen full brightness, all cores 100% | 00:30 |
| esad | + charging | 00:30 |
| mntmn | system power maybe max 3A-ish | 00:31 |
| erlehmann | why are other manufactures not using them? | 00:31 |
| erlehmann | i mean, they seem to be the safest ones | 00:31 |
| erlehmann | PING 2973379141 | 00:31 |
| - erlehmann (QUIT: Remote host closed the connection) (~erlehmann@x59cc894d.dyn.telefonica.de) | 00:31 | |
| mntmn | (@5V) | 00:31 |
| mntmn | erlehmann because they are bigger | 00:31 |
| + erlehmann (~erlehmann@x59cc894d.dyn.telefonica.de) | 00:31 | |
| mntmn | erlehmann they’re not using them because lifepo4 is a bit bigger | 00:31 |
| mntmn | slightly less energy density | 00:31 |
| erlehmann | i see | 00:32 |
| mntmn | but all other parameters are better in lifepo4 | 00:32 |
| esad | I'm just wondering if it could be powered by the 30W apple usb-c charger (once it comes usb-c port :-) | 00:32 |
| mntmn | it can deliver massive current while not getting hot and it has more recharge cycles | 00:32 |
| erlehmann | btw very good irclog setup | 00:32 |
| erlehmann | mntmn, how many? li-ion has like 400? | 00:33 |
| mntmn | esad 30W should be enough | 00:33 |
| mntmn | erlehmann 2000 are advertised but maybe that’s hyperbole | 00:33 |
| erlehmann | wow | 00:33 |
| erlehmann | mntmn, surely more than the 400 | 00:33 |
| erlehmann | i am impressed | 00:33 |
| mntmn | so probably somewhere in the middle ;) | 00:33 |
| erlehmann | does anyone have any idea how X11 can hang through xgc? | 00:33 |
| esad | actually as a stop gap solution, maybe there is some kind of usb-c to (what is the name of the plug that reform is using?) adapter | 00:34 |
| mntmn | what is xgc? | 00:34 |
| mntmn | esad, “barrel jack” | 00:34 |
| erlehmann | mntmn, an X11 test program | 00:34 |
| esad | coming from apple land, we love adapters :) | 00:34 |
| mntmn | erlehmann, you could run X in gdb via ssh | 00:34 |
| mntmn | that’s how i debug these problems | 00:34 |
| erlehmann | mntmn, i can hang X11 on REFORM by starting xgc, then clicking “andReverse”, “DoubleDash” and “Run” i think | 00:35 |
| mntmn | esad true true | 00:35 |
| mntmn | erlehmann cool cool | 00:35 |
| erlehmann | someone here reproduced it, m2 maybe? can't remember right now | 00:35 |
| mntmn | no, you just asked him to when he was just trying to get familiar with his brand new system ^^ | 00:35 |
| esad | https://www.globalsources.com/gsol/I/DC-power/p/sm/1159257934.htm | 00:36 |
| erlehmann | esad, why USB-C, didn't you get an appropriate REFORM power adapter? | 00:36 |
| erlehmann | are you in a country with different power plug standard maybe? | 00:36 |
| erlehmann | (than germany) | 00:36 |
| mntmn | esad crazy stuff | 00:36 |
| esad | erlehmann I already have bunch of USB-C adapter everywhere | 00:36 |
| erlehmann | i see | 00:37 |
| esad | don't want to carry another brick around | 00:37 |
| mntmn | it could work! | 00:37 |
| mntmn | i hope it outputs something like 5V ;) | 00:37 |
| esad | I'd like to try using reform as my primary computer for a week :) | 00:37 |
| mntmn | ha | 00:37 |
| mntmn | i tried that as well! | 00:37 |
| erlehmann | it is certainly possible to use reform as a primary computer, i did that | 00:37 |
| esad | @mntmn and? :) | 00:38 |
| erlehmann | it worked pretty well, apart from hardware issues | 00:38 |
| mntmn | sadly the hw acceleration in kicad doesn’t work yet | 00:38 |
| mntmn | so i can’t use it for electronics yet | 00:38 |
| erlehmann | mntmn, have you tried phlipple btw? or leocad? both seem to have different 3D problems | 00:38 |
| erlehmann | LeoCAD works very slowly with LIBGL_ALWAYS_SOFTWARE=1, but does not show parts without it :( | 00:39 |
| erlehmann | which is unfortunate | 00:39 |
| esad | is the SoC soldered on the main (mother?) board? would it be easy to upgrade it in the future? | 00:39 |
| mntmn | i didn’t even know these tools but i can try them | 00:39 |
| mntmn | esad it’s not soldered, it’s on the tinyrex som module | 00:40 |
| mntmn | i plan an upgrade but it will probably involve swapping the motherboard as well | 00:40 |
| mntmn | because new i/o possibilities | 00:40 |
| mntmn | i think imx6qp 1.2ghz is kind of the pinnacle of that tech generation | 00:41 |
| mntmn | ah do you have 1.2ghz? i don’t remember right now | 00:41 |
| erlehmann | mntmn, phlipple is a game. but LeoCAD is a CAD program with a default parts library of LEGO bricks. | 00:41 |
| mntmn | if not then there is an upgrade option | 00:41 |
| erlehmann | i guess it is possible to not have it use LEGO bricks | 00:41 |
| mntmn | ok i will try both now | 00:42 |
| erlehmann | mntmn, the main purpose of LeoCAD is to make LEGO-like renderings of your model in different stages, so you can output assembly instructions. | 00:42 |
| mntmn | i have one success story: i compiled FreeCAD on reform. and it works very well | 00:42 |
| esad | 1.2 | 00:42 |
| mntmn | (i have more success stories of course. for example pcsx_rearmed) | 00:43 |
| mntmn | esad, not sure if there is a more powerful 32 bit arm system ;) | 00:43 |
| erlehmann | this is LeoCAD https://www.leocad.org/ | 00:43 |
| mntmn | erlehmann i also helped with an etnaviv patch that might come out soon, and you still have the mesa debug version that stops on asserts i think | 00:44 |
| esad | what is the state of ARM chromebooks? what do the they use? | 00:44 |
| erlehmann | mntmn, nice | 00:44 |
| mntmn | esad, current generation uses a53, a72 cores | 00:45 |
| mntmn | i plan an imx8 reform version once it's properly mainlined | 00:45 |
| mntmn | i have 2 imx8 systems and they're both highly experimental/unstable | 00:46 |
| erlehmann | what is the appeal of imx8 compared to imx6? | 00:46 |
| mntmn | if you want to make music on reform, LMMS works well btw (and even looks good with a proper theme installed) | 00:47 |
| mntmn | erlehmann, it is faster | 00:47 |
| esad | ahhh so reform is like iPad 2 cpu-wise :) | 00:47 |
| mntmn | esad, yes but with 4 cores and 4gb ram ;) | 00:47 |
| esad | true! | 00:48 |
| erlehmann | mntmn, i have not found any problems with REFORM being slow so far | 00:48 |
| mntmn | erlehmann, me only with JS heavy websites etc that you probably don't use | 00:48 |
| mntmn | and compiling stuff can always be faster | 00:48 |
| mntmn | or zooming into pdfs | 00:48 |
| esad | yeah with JS it starts to become bit sluggish | 00:48 |
| erlehmann | mntmn, JS is halting problem, even on a new macbook it can be sloooooooooooow | 00:48 |
| erlehmann | chrome with more than 5 tabs? slow | 00:48 |
| esad | also haven't tried editing large files in Sublime Editor | 00:48 |
| mntmn | erlehmann, i'm talking not about edge cases but about average cases | 00:48 |
| erlehmann | mntmn, zooming into PDFs in a browser or with mupdf / zathura? | 00:49 |
| mntmn | esad, you're using sublime on reform? how does it work? | 00:49 |
| esad | I was just wondering if 100 euro in unit cost more would bring 3x the performance | 00:49 |
| mntmn | erlehmann, evince i think | 00:49 |
| mntmn | esad, it's not only about cost | 00:49 |
| mntmn | imx8 drivers and ecosystem are mega immature | 00:49 |
| erlehmann | ah i see | 00:49 |
| esad | I mean at comparable thermal/power specs | 00:49 |
| esad | true | 00:49 |
| mntmn | it's not even in the kernel yet | 00:49 |
| mntmn | despite people claiming otherwise (but they just didn't take a look) | 00:50 |
| erlehmann | i think i tried to md5sum a REFORM boot image, but that took a long time, but i do not know if it is CPU bound. | 00:50 |
| mntmn | like, linux does not support imx8 yet | 00:50 |
| mntmn | there are some other beefy socs but the problem is always documentation | 00:50 |
| mntmn | i don't want anything where the users would have to sign an NDA or sth | 00:50 |
| esad | yeah | 00:50 |
| mntmn | erlehmann, you can check that with dstat for example | 00:51 |
| erlehmann | i can check what? | 00:51 |
| esad | and who knows maybe RISC-V catches up in the next years :) | 00:51 |
| mntmn | another candidate for me is layerscape (also freescape/nxp), but it's quite expensive and doesn't have a gpu | 00:51 |
| mntmn | yes | 00:51 |
| mntmn | i mean the 1.5ghz risc-v chip was launched on crowd supply as well | 00:52 |
| mntmn | so there will hopefully be synergies | 00:52 |
| erlehmann | i would like to have suspend-to-disk working on REFORM. anyone who could help me? | 00:52 |
| esad | that would be really great | 00:52 |
| mntmn | erlehmann, we can take a look at 35c3 | 00:52 |
| erlehmann | i already have a swap partition and “printf disk >/sys/power/state” saves an image to that apparently | 00:52 |
| erlehmann | but it does not get restored | 00:52 |
| erlehmann | on reboot i am where i was before | 00:52 |
| erlehmann | with a normal reboet | 00:53 |
| esad | what is the state of other cpu architectures? what about x86 socs? | 00:53 |
| erlehmann | but not where i was before saving | 00:53 |
| mntmn | erlehmann, you might have better luck on the arm kernel mailing list or sth | 00:53 |
| mntmn | esad, i find AMD V1000 interesting but it doesn't look like there are public docs | 00:54 |
| mntmn | and i find x86 overly complicated, it feels like legacy | 00:54 |
| erlehmann | mntmn, oh, suspend-to-disk is not specific to arm. it always works. it is more about the boot process. maybe i forgot to set a parameter in uboot or tell systemd? | 00:54 |
| erlehmann | x86 also has a management engine | 00:54 |
| mntmn | erlehmann, i don't know how it works. what does the documentation say? how does the process work? | 00:54 |
| mntmn | erlehmann, no, x86 does not | 00:55 |
| mntmn | some implementations have | 00:55 |
| erlehmann | mntmn, documentation says it needs documentation | 00:55 |
| mntmn | uh | 00:55 |
| erlehmann | mntmn, oh okay | 00:55 |
| erlehmann | i'll google it again | 00:55 |
| erlehmann | but the man page was not helpful | 00:55 |
| mntmn | intel or amd could totally decide to make a chip without remote control engines etc | 00:55 |
| mntmn | nothing in amd64 depends on it | 00:56 |
| erlehmann | well, as if they would | 00:56 |
| mntmn | erlehmann: > You will need to point the kernel to your swap using the resume= kernel parameter, which is configured via the boot loader | 00:56 |
| mntmn | https://wiki.archlinux.org/index.php/Power_management/Suspend_and_hibernate#Hibernation | 00:57 |
| mntmn | erlehmann, maybe they already did? | 00:57 |
| erlehmann | mntmn, i would not know about it if it happened in the last 10 years. | 00:57 |
| erlehmann | mntmn, good to know, but how do i update this on the fly? | 00:57 |
| mntmn | funny, AMD management engine is ARM | 00:58 |
| mntmn | > The PSP itself is an ARM core inserted on the main CPU die | 00:58 |
| erlehmann | or do i always put the resume parameter there and the kernel aborts if it does not find an image? | 00:58 |
| mntmn | erlehmann, tricky, but not impossible | 00:58 |
| mntmn | you would need to introduce a u-boot .scr | 00:58 |
| erlehmann | > You will also need to configure the initramfs. | 00:58 |
| mntmn | or wait... maybe you can script it in u-boot | 00:59 |
| erlehmann | i guess that is the difficult part, since REFORM does not have one | 00:59 |
| mntmn | because it can read ext4 | 00:59 |
| mntmn | nah | 00:59 |
| mntmn | if you don't have initramfs you don't need to configure it | 00:59 |
| mntmn | so you could write a script in u-boot's environment that checks for a file | 00:59 |
| mntmn | like, does some file exist that was placed there by the suspend script | 01:00 |
| mntmn | if yes, append kernel parameter and remove file | 01:00 |
| mntmn | or remove file during linux boot (easier) | 01:00 |
| mntmn | (probably better not letting u-boot mess with your fs) | 01:01 |
| erlehmann | mntmn, i'll try to see if i can just give the parameter and the kernel handles it properly if swap is garbage | 01:01 |
| mntmn | good plan | 01:01 |
| erlehmann | and then i'll tell you all | 01:01 |
| erlehmann | wait for it | 01:01 |
| erlehmann | brb | 01:01 |
| mntmn | but | 01:01 |
| erlehmann | REFORM activate | 01:01 |
| erlehmann | but? | 01:01 |
| mntmn | make sure you get the setenv commands etc right | 01:01 |
| - esad (QUIT: Quit: My MacBook has gone to sleep. ZZZzzz…) (~textual@84-113-242-212.cable.dynamic.surfer.at) | 01:01 | |
| mntmn | did you ever do that? | 01:01 |
| erlehmann | no | 01:02 |
| mntmn | ok, quick examples: | 01:02 |
| erlehmann | i thought i can just boot from the uboot command line? | 01:02 |
| erlehmann | by using the examples in the handbook | 01:02 |
| mntmn | yes but you need to know how to exactly do that | 01:02 |
| mntmn | before you do it | 01:02 |
| mntmn | ah | 01:02 |
| erlehmann | is that correct? | 01:03 |
| mntmn | there are no quotes in the handbook after the setenv bootargs thing | 01:03 |
| mntmn | let me double check if that's ok | 01:03 |
| mntmn | yes it's ko | 01:04 |
| mntmn | ok | 01:04 |
| + mrak (~textual@84-113-242-212.cable.dynamic.surfer.at) | 01:04 | |
| mntmn | if you want to use variables and they should stay variables (like in bash scripts) you have to use single quotes | 01:04 |
| mrak | sublime doesn't have an arm build :( | 01:04 |
| * mrak -> esad | 01:04 | |
| mntmn | esad, aha! | 01:04 |
| erlehmann | esad, so you have to recompile it? | 01:05 |
| mntmn | closed source i guess? | 01:05 |
| esad | it's closed source, yes :/ | 01:05 |
| mntmn | https://softwarerecs.stackexchange.com/questions/40456/sublime-text-3-alternative-to-use-on-arm-linux | 01:05 |
| esad | I'm just too used to it to switch to anything else at this late age :) | 01:05 |
| erlehmann | an advantage of REFORM: you can not be lured into proprietary offerings not compiled for ARM | 01:05 |
| mntmn | except if you use exagear | 01:05 |
| mntmn | :D | 01:05 |
| esad | weren't we talking before how JS brings down reform to a halt :) | 01:05 |
| erlehmann | esad, emacs probably has some keybinding mode for you? i mean, it tries to emulate wordstar | 01:06 |
| esad | now imagine a text editor in javascript :) | 01:06 |
| mntmn | esad, well it depends ;) | 01:06 |
| erlehmann | js is not the problem, badly written websites with js are | 01:06 |
| mntmn | yeah some js stuff can run well | 01:06 |
| erlehmann | mntmn, exagear is the x86 environment? | 01:06 |
| mntmn | so i'd give it a try | 01:06 |
| erlehmann | doesn't polkit include js? | 01:06 |
| mntmn | erlehmann, yeah the x86 emulator jit thing, also proprietary | 01:06 |
| mntmn | erlehmann, yes | 01:06 |
| esad | actually I could never switch to Atom/VScode on my Mac because it was laggy | 01:07 |
| mntmn | ah oh | 01:07 |
| erlehmann | sorry, i'm sleepy. i'll research suspend/wakeup from disk later, since the grub keyboard layout is not neo2. | 01:07 |
| mntmn | http://uvviewsoft.com/cudatext/ | 01:07 |
| erlehmann | so i need concentration for it to mentally remap it | 01:07 |
| mntmn | looks like a sublime clone | 01:07 |
| mntmn | erlehmann, sorry for that. you could use an usb keyboard maybe | 01:08 |
| mntmn | erlehmann, i'll finally test your 3d apps now | 01:09 |
| erlehmann | mntmn, i probably should. or be more awake | 01:10 |
| mntmn | yeah maximum alertness is required for fiddling with bootloaders | 01:10 |
| mntmn | erlehmann, leocad does not show any parts, correct? | 01:11 |
| mntmn | like, it renders only white? | 01:11 |
| erlehmann | mntmn, on my REFORM, yes | 01:13 |
| mntmn | bug confirmed | 01:13 |
| erlehmann | mntmn, compare to rendering with “LIBGL_ALWAYS_SOFTWARE=1 leocad” | 01:13 |
| erlehmann | i have no idea what could cause this | 01:13 |
| esad | btw my trackball isn't smooth when I apply even slightest pressure. i haven't use trackballs in a while, but I remember they could be bit better than this. any ideas how to improve it? | 01:13 |
| mntmn | probably shader related | 01:13 |
| mntmn | esad yes, see hardware tip #1, one moment | 01:13 |
| esad | would a rubbery ball like in old mice would make any difference? | 01:13 |
| mntmn | no the ball is perfect | 01:14 |
| erlehmann | which reminds me, i still need those ersatz screws | 01:14 |
| mntmn | and it's supposed to be smooth | 01:14 |
| mntmn | erlehmann: yes you'll get them | 01:14 |
| erlehmann | mntmn, do you have time before congress? | 01:14 |
| mntmn | not sure | 01:14 |
| mntmn | i'm in a lot of stress atm | 01:14 |
| erlehmann | then let's do it at congress | 01:14 |
| mntmn | yeah i'll have everything there | 01:15 |
| erlehmann | i also have the plastic proofs for you with the printed nyancats from the guy who told me he could print onto laptops | 01:15 |
| mntmn | our assembly will be like a service station ;) | 01:15 |
| mntmn | cool cool | 01:15 |
| erlehmann | he might or might not be at the congress, dunno | 01:15 |
| mntmn | right now i'm trying to ship all the beta reforms (i just completed number 9 of 11) | 01:15 |
| erlehmann | i have to ask someone what his name was | 01:15 |
| mntmn | and i have a backlog of todos also for the campaign etc | 01:15 |
| mntmn | esad https://source.mntmn.com/MNT/reform/wiki/HardwareTricks | 01:16 |
| mntmn | so, the ball is stiff if it is squeezed between the trackball assembly and the black top ring that prevents it from falling out | 01:17 |
| mntmn | you can try to carefully widen the ring, that should help | 01:17 |
| mntmn | erlehmann, what was the failing xgc sequence again | 01:17 |
| - esad (QUIT: Quit: My MacBook has gone to sleep. ZZZzzz…) (~textual@84-113-242-212.cable.dynamic.surfer.at) | 01:19 | |
| mntmn | esad btw epiphany-browser can yield better performance for some js-heavy sites (i use it for irccloud for example) | 01:19 |
| mntmn | erlehmann: phlipple renders completely black for me | 01:20 |
| + mrak (~textual@84-113-242-212.cable.dynamic.surfer.at) | 01:21 | |
| * mrak -> esad | 01:21 | |
| mntmn | ah no it seems to hang at 110% cpu | 01:21 |
| erlehmann | mntmn, for me too | 01:21 |
| esad | @mntmn yeah it seems bit better now :) | 01:21 |
| esad | I was wonder if the ball would fall out if I turn the reform upside down | 01:22 |
| mntmn | esad, cool. it should not fall out if the black roof prevents it ;) | 01:22 |
| esad | I'm now trying to switch to i3 and use "surf" | 01:22 |
| mntmn | erlehmann, which opengl version does phlipple require? | 01:22 |
| erlehmann | mntmn, no idea | 01:23 |
| mntmn | esad, did you try epiphany-browser yet? | 01:23 |
| esad | not yet, will do | 01:23 |
| erlehmann | dillo is very fast for me | 01:23 |
| esad | I tried midori or something | 01:23 |
| erlehmann | but can't do much | 01:23 |
| esad | I actually spent my afternoon surfing in netsurf | 01:23 |
| mntmn | qupzilla and qutebrowser will also work fine _soon_ when a certain etnaviv patch goes through | 01:23 |
| esad | it renders stuff wrong, but it's faster than regular browser on a modern pc :) | 01:23 |
| mntmn | yeah netsurf is really fast | 01:24 |
| erlehmann | mntmn, why does a browser need etnaviv patches? | 01:24 |
| mntmn | how fast the web could be ;) | 01:24 |
| mntmn | erlehmann, qtwebengine | 01:24 |
| mntmn | it uses gpu rendering | 01:24 |
| erlehmann | oh, so can it be turned off? | 01:24 |
| erlehmann | i have good experience with disabling multiprocess in firefox | 01:24 |
| mntmn | not sure maybe... it will probably be slower than a browser that's not designed for gpu rendering | 01:24 |
| esad | I'll give epihany a try | 01:25 |
| mntmn | disabling multiprocess? | 01:25 |
| erlehmann | set browser.tabs.remote.autostart to false in firfox | 01:25 |
| mntmn | does that make it faster? | 01:25 |
| erlehmann | then it does not start a new process for each tab | 01:25 |
| mntmn | i'll try | 01:25 |
| mntmn | hang on | 01:25 |
| erlehmann | i think it might make it slower | 01:25 |
| erlehmann | but it does not hog all cores on errors | 01:25 |
| mntmn | i also disabled that autocomplete/suggest stuff in the firefox address bar which causes a delay | 01:26 |
| mntmn | ah ok. | 01:26 |
| erlehmann | (errors meaning: cpu hungry javascript) | 01:26 |
| mntmn | i use noscript | 01:26 |
| mntmn | (in ff) | 01:26 |
| mntmn | and i run epiphany in parallel for js pages ;) | 01:26 |
| erlehmann | i also use noscript, but the multiprocess thing seems to be a double-edged sword (i hope i got the right metaphor) | 01:26 |
| erlehmann | for systems with high specs, ff got faster | 01:26 |
| erlehmann | for systems with low specs, ff got slower | 01:26 |
| specing | Firefox >= 60.0 is a real snail | 01:27 |
| mntmn | yeah? i noticed the opposite | 01:27 |
| specing | 23s for it to respond to input (as in letters to appear in box) | 01:27 |
| mntmn | i used pre-quantum FF on reform and it was incredibly slow | 01:27 |
| mntmn | since quantum it's really usable finally | 01:27 |
| specing | I also disable JIT for security | 01:28 |
| erlehmann | firefox 52 takes a long time to start | 01:28 |
| esad | would it be possible to emulate scrollwheel by holding some hotkey and then scrolling via trackball? | 01:28 |
| erlehmann | specing, how to disable JIT | 01:28 |
| specing | erlehmann: compile ff with --disable-ion | 01:28 |
| erlehmann | esad, i asked this already, the problem is that this is two usb devices | 01:28 |
| erlehmann | esad, so you'll have to do it in software probably? | 01:28 |
| erlehmann | i would love to have the keyboard be a mouse as well | 01:29 |
| erlehmann | for scrolling | 01:29 |
| mntmn | that's possible to do | 01:29 |
| mntmn | we can put it on the lists to prototype @ 35c3 | 01:29 |
| mntmn | esad there is wheelmode (documented in the manual) but it would be nicer with a hotkey, i agree | 01:30 |
| erlehmann | mntmn, also i was mistaken, ctrl being the most-left key in the bottom row conflicts with my muscle memory | 01:30 |
| mntmn | erlehmann, that was your choice ;) | 01:30 |
| erlehmann | mntmn, as i said, i was mistaken! | 01:30 |
| mntmn | for me it's the leftmost key in the middle row | 01:30 |
| mntmn | ah ok sorry | 01:30 |
| erlehmann | wheelmode makes it not-impossible but hard to use apps that need mouse3 | 01:30 |
| esad | my iritation with the layout was that arrow keys are not the rightmost | 01:31 |
| mntmn | yeah | 01:31 |
| erlehmann | without wheelmode, one could press both at the same time, as X11 handles it, but then you have no wheel! | 01:31 |
| erlehmann | :/ | 01:31 |
| esad | I think I will remap it so that it's like that | 01:31 |
| mntmn | true true. but fixable | 01:31 |
| erlehmann | esad, on my machine, arrow keys are the rightmost, so i got at least that right! | 01:31 |
| erlehmann | but it is easily fixable in the firmware | 01:31 |
| erlehmann | very easy i think | 01:31 |
| esad | and remap old "up" and right shfit to pg up/down | 01:31 |
| erlehmann | it is just an arrow | 01:31 |
| esad | ah, so layout changed in between revisions? | 01:32 |
| mntmn | esad, pgupdown is fn+up/down | 01:32 |
| mntmn | i think i forgot to document that in the manual | 01:32 |
| erlehmann | esad, no i use neo2 | 01:32 |
| erlehmann | so mntmn asked me to submit a keymap | 01:32 |
| erlehmann | which i did | 01:32 |
| esad | yeah but I'm using it so often that I really like it when it's single key (like on thinkpads) | 01:32 |
| mntmn | ok understood | 01:32 |
| erlehmann | i have pgup/pgdown on the top right, which is nice … but i often wish i had del/ins there. in reality, i am using it often, but i wish i had more keys … like on a thinkpad. | 01:33 |
| erlehmann | i will improve the fn thing, i think | 01:33 |
| esad | oh I installed a game called epiphany :) | 01:34 |
| mntmn | i think the next layout will have more keys, by splitting bigger keys | 01:34 |
| mntmn | esad, yeah i did that mistake as well :D | 01:34 |
| mntmn | the package is epiphany-browser | 01:34 |
| esad | it's boulder dash! :) | 01:34 |
| erlehmann | esad, epiphany-browser :D | 01:34 |
| mntmn | haha | 01:34 |
| erlehmann | ahaha | 01:34 |
| erlehmann | i remember chromium BSU being chromium | 01:34 |
| erlehmann | it is a shoot-em-up game | 01:34 |
| erlehmann | then the browser came | 01:35 |
| mntmn | haha true, same problem | 01:35 |
| erlehmann | mntmn, i think the idea with the split space bar is a really good one, because i certainly will remap two of the space keys to be modifiers | 01:35 |
| mntmn | current chromium works as well btw. but i think epiphany is a bit faster and less googly (if you mind that) | 01:35 |
| mntmn | btw i use claws-mail on reform because it's faster than the rendering of emacs-notmuch | 01:36 |
| mntmn | the irony... | 01:36 |
| esad | btw is it possible to get bitmap fonts in sakura? | 01:37 |
| mntmn | yes | 01:37 |
| mntmn | i use unifont | 01:37 |
| esad | I isntalled fixed from xfonts but it doesn't show up in the fonts picker | 01:37 |
| esad | but does it turn off antialiasing for bitmap fonts? | 01:38 |
| mntmn | ah wait... so i ship unifont-ttf i think, if you select unifont with size 12 you will have a nice pixely experience | 01:38 |
| erlehmann | mntmn, didn't you have a rendering tip for emacs? | 01:38 |
| mntmn | with no antialias | 01:38 |
| mntmn | erlehmann: yes, disable line spacing if it is active | 01:38 |
| erlehmann | mntmn, i mean like, not try to list *all* emails | 01:38 |
| esad | that's nice | 01:38 |
| esad | if it only could be little bit smaller :) | 01:38 |
| mntmn | erlehmann, yes but that is querying, not rendering | 01:39 |
| mntmn | esad, i believe there are other bitmap-like ttf fonts for this purpose, let me remember | 01:39 |
| mntmn | maybe this http://pelulamu.net/unscii/ | 01:40 |
| mntmn | (not sure if un-antialiased) | 01:40 |
| esad | epiphany looks nice. I can hear the fan after visiting the local newspaper site :) | 01:41 |
| erlehmann | esad, you can try to scale unifont smaller, some of it still works | 01:41 |
| erlehmann | but 12 or 16 are good font sizes for unifont | 01:41 |
| erlehmann | (i drew >500 emoji for it) | 01:41 |
| mntmn | erlehmann, but will it look crisp? | 01:41 |
| mntmn | the cool thing is that at 12pt it is pixel perfect | 01:41 |
| erlehmann | mntmn, it will not look smudged. i say try 10 or 8 and report back? | 01:41 |
| mntmn | esad, yeah! i heard with erlehmann's reformd you can have less harsh fan sequencing | 01:42 |
| mntmn | but it haven't integrated it myself yet | 01:42 |
| mntmn | sorry erlehmann's reform_fan_d | 01:42 |
| erlehmann | i think at 16pt it doubles some pixels | 01:42 |
| erlehmann | so not smudgy | 01:42 |
| erlehmann | reformfand! | 01:42 |
| mntmn | erlehmann: characters like v and y look bad | 01:42 |
| erlehmann | with 10pt or 16pt? | 01:42 |
| mntmn | erlehmann: with 10 and 8 | 01:43 |
| mntmn | i will integrate reformfand soon and publish a doc about how to update reformd + reformfand, but i need the time for that. | 01:43 |
| erlehmann | mntmn, then unifont is unfit for the purpose of being really small, but wasn't there something by leah neukirchen? | 01:43 |
| erlehmann | wait | 01:43 |
| erlehmann | i'll get it | 01:43 |
| mntmn | xgc crash confirmed | 01:44 |
| mntmn | not crash, X hang | 01:45 |
| erlehmann | so any idea what causes it? | 01:45 |
| erlehmann | i do not even know if it is specific to REFORM / etnaviv | 01:45 |
| mntmn | no because i did not run X in gdb | 01:45 |
| mntmn | sure, i don't really wanna try it on my debian-macbok right now | 01:45 |
| mntmn | because that's a hassle to reboot | 01:45 |
| mntmn | oh wait | 01:46 |
| erlehmann | esad, sorry those are still 15px and 13px high http://chneukirchen.org/fonts/ | 01:46 |
| mntmn | it does not actually hang | 01:46 |
| mntmn | erlehmann, you just have to be patient | 01:46 |
| mntmn | erlehmann, after a minute it will finish and X continues | 01:46 |
| mntmn | so they do something incredibly slow inside of X | 01:46 |
| erlehmann | > The goal is to create a well readable font that is only five pixels wide, such that you can fit four 80-chars wide shells next to each other on a 1600x1200 display. 5x13 is special in being pretty tall, which greatly increases readability. | 01:46 |
| erlehmann | mntmn, oh! | 01:46 |
| erlehmann | mntmn, so that means ANY application can hang X? | 01:46 |
| mntmn | it renders some geometrical thingie | 01:46 |
| mntmn | i don't know, how does the program work? | 01:47 |
| erlehmann | you give it parameters and it renders stuff | 01:47 |
| erlehmann | to demonstrate X11 primitives | 01:47 |
| erlehmann | but if this is only very slow and not a hang, false positive | 01:47 |
| mntmn | yes but how does it work internally | 01:47 |
| mntmn | what does the code do etc | 01:47 |
| mntmn | maybe it just kicks off some pathological case in X or in glamor on etnaviv | 01:48 |
| erlehmann | yeah that's what i thought | 01:48 |
| mntmn | (glamor is the hw accelerated 2d drawing in X) | 01:48 |
| erlehmann | wanna try if it is faster on the macbook? | 01:48 |
| erlehmann | the only other machine i have is even slower than reform | 01:48 |
| mntmn | sure i will, hang on | 01:48 |
| erlehmann | Percentage of Test - scrollbar permits specifying only a percentage of | 01:49 |
| erlehmann | the test to be run. The number at the left indicates the current set‐ | 01:49 |
| erlehmann | ting, which defaults to 100%. | 01:49 |
| mntmn | no hang on the i7 | 01:49 |
| erlehmann | (from “man xgc”) | 01:49 |
| erlehmann | oh | 01:49 |
| mntmn | but it also looks a bit different, the result | 01:49 |
| erlehmann | so it is a pathological case | 01:49 |
| mntmn | yes probably | 01:49 |
| mntmn | i will try now with glamor turned off | 01:49 |
| erlehmann | i will try now on 01:00.0 VGA compatible controller: Advanced Micro Devices, Inc. [AMD/ATI] RV515/M54 [Mobility Radeon X1400] | 01:49 |
| erlehmann | okay, it renders *immediately* | 01:50 |
| erlehmann | 0,006 seconds | 01:50 |
| erlehmann | mntmn, i'll open a ticket on source.mntmn.com, ok? | 01:50 |
| mntmn | tickets are always welcome yeah | 01:51 |
| mntmn | oh wow | 01:52 |
| mntmn | with glamor turned off, it renders immediately on reform | 01:52 |
| mntmn | the sad story is (and it's really strange to me, i have to ask some x people) -- turning off glamor also turns of GL | 01:52 |
| mntmn | like, hardware GL | 01:52 |
| mntmn | the software drawing is actually fast enough, but it prevents the use of HW GL | 01:53 |
| erlehmann | i see | 01:53 |
| mntmn | maybe X can be patched somehow to use more software rendering pathss | 01:54 |
| mntmn | also funny (or sad) how blender is more responsive with software / llvmpipe GL then with "full redraw" hardware GL | 01:57 |
| erlehmann | mntmn, can i enable llvmpipe for arbitrary programs? | 02:02 |
| mntmn | erlehmann, yeah i think you already did it | 02:03 |
| erlehmann | how? | 02:03 |
| mntmn | with the GL_ALWAYS_SOFTWARE flag | 02:03 |
| erlehmann | oh | 02:03 |
| erlehmann | interesting | 02:03 |
| mntmn | there are also tons of mesa/GL debugging capabilities | 02:03 |
| mntmn | but i don't remember them atm | 02:03 |
| erlehmann | here https://www.mesa3d.org/envvars.html | 02:04 |
| erlehmann | > MESA_EXTENSION_OVERRIDE - can be used to enable/disable extensions. A value such as "GL_EXT_foo -GL_EXT_bar" will enable the GL_EXT_foo extension and disable the GL_EXT_bar extension. | 02:05 |
| erlehmann | this is nice | 02:05 |
| mntmn | ah cool | 02:05 |
| mntmn | there is also a gdb init file that makes gdb much more interesting | 02:06 |
| mntmn | with live disassembly etc | 02:06 |
| mntmn | erlehmann, what is the diablo clone called again? and other games that work well except minetest? | 02:08 |
| erlehmann | flare | 02:08 |
| erlehmann | garden-of-coloured-lights | 02:09 |
| erlehmann | liberation ciruit worked very well | 02:09 |
| erlehmann | i remember you asking if it was immediate mode GUI | 02:09 |
| erlehmann | and i was like “this is what allegro looks like” | 02:09 |
| mntmn | thanks yeah1 | 02:13 |
| mntmn | i'm writing the recommended/tested software lisst | 02:13 |
| erlehmann | mntmn, btw the reform man page does not belong into section 1 | 02:16 |
| erlehmann | or am i mistaken? | 02:16 |
| mntmn | that is true | 02:16 |
| erlehmann | maybe section 7 | 02:17 |
| erlehmann | Section 7: conventions and miscellany - Linux man pages | 02:17 |
| erlehmann | see, for example, https://linux.die.net/man/7/boot | 02:18 |
| mntmn | btw, for gameboy fun: gngb -o -s --sample_rate=48000 supermarioland.rom | 02:18 |
| erlehmann | where do i get supermarioland.rom? | 02:18 |
| mntmn | (disclaimer: i own an original copy of this game) | 02:18 |
| erlehmann | haha | 02:19 |
| mntmn | i just wanted to give you well working options for gngb ;) | 02:19 |
| mntmn | also there is an amiga emulator with arm jit | 02:19 |
| erlehmann | didn't you say everything pandora worked well on REFORM? | 02:20 |
| erlehmann | i am sure they have some kind of software repo | 02:20 |
| mntmn | yes it should, good question | 02:20 |
| mntmn | uae4arm is the one. but i couldn't get the mouse to work yet :/ | 02:21 |
| esad | ok I had to remove /etc/fonts/conf.d/70-no-bitmaps.conf | 02:22 |
| esad | then bitmaps font show up in font picker | 02:22 |
| mntmn | ahh! good hint | 02:22 |
| esad | anyway "Fixed Medium Semi-Condensed 10" is perfect for terminal/sakura | 02:22 |
| mntmn | i will put it in the wiki later | 02:22 |
| esad | you can try in sakura.conf | 02:22 |
| esad | ah yes apt install xfonts-traditional | 02:23 |
| esad | could it be that the last column of pixels of the display is somehow cut off? | 02:28 |
| erlehmann | esad, because of i3bar, right? | 02:29 |
| esad | yeah | 02:29 |
| esad | but it could be i3 | 02:29 |
| esad | the bottom right celsius C is cut off | 02:29 |
| esad | but also when I open a new terminal window it has blue borders everywhere except on the right side | 02:30 |
| erlehmann | esad, i say open an issue on https://source.mntmn.com | 02:32 |
| erlehmann | i am sure mntmn will give you credentials (or has already, maybe?) | 02:32 |
| mntmn | i will on request sure | 02:34 |
| mntmn | 2 pixels are missing yeah | 02:35 |
| mntmn | because there was a driver problem with the 1366 physical resolution | 02:36 |
| mntmn | (non multiple of 8) | 02:36 |
| mntmn | so i set the resolution to 1368 | 02:36 |
| mntmn | i’m sure it can be worked around though | 02:36 |
| esad | @mntmn how do I send the request? mail? | 02:37 |
| mntmn | yeah or i can just invite you to your email that you ordered with | 02:38 |
| esad | btw I just installed dillo browser. it's ancient and nearly unusable but wow it's fast. | 02:38 |
| esad | yeah that works | 02:38 |
| mntmn | ok | 02:38 |
| erlehmann | esad, nearly unusable, yes, but FAST | 02:40 |
| erlehmann | esad, another thing would be elinks, which is really nice. or links2 -g (which is worse than dillo for most stuff). | 02:40 |
| esad | it'd be interesting to make browser that kind of approximates what css3 layout wants to do | 02:40 |
| esad | or supports only the most used part of it | 02:41 |
| esad | but it's this fast | 02:41 |
| esad | I don't care about perfect layout if it would mean this speed | 02:41 |
| esad | but yeah | 02:41 |
| mntmn | yeah that would be nice | 02:41 |
| mntmn | there can never be enough browser engines ;) | 02:41 |
| mntmn | have you heard of browsh? | 02:42 |
| erlehmann | no? | 02:42 |
| esad | in safari you can turn on reader mode per default for all websites :) | 02:42 |
| esad | browsh? | 02:42 |
| mntmn | https://www.brow.sh/ | 02:42 |
| erlehmann | haha https://www.dillo.org/FAQ.html#q28 | 02:42 |
| erlehmann | > If Dillo requests an image and receives a response containing a redirection (i.e., pointing to a different URL), the redirection is not followed. These have a strong tendency to be advertisements. | 02:43 |
| mntmn | haha | 02:43 |
| erlehmann | > For privacy reasons, Dillo runs with cookies disabled by default. | 02:43 |
| esad | yeah it's from a different era | 02:43 |
| erlehmann | > Dillo does not have a file containing browsing history. | 02:44 |
| esad | so brow.sh runs "real" engine on the server and transmits asciified output or? | 02:44 |
| mntmn | yes | 02:44 |
| erlehmann | complex | 02:45 |
| mntmn | so, now we also need an open hardware server ;) | 02:45 |
| erlehmann | harr harr | 02:45 |
| esad | haha | 02:45 |
| mntmn | btw very off topic but do you know if there are schematics for talos secure workstation? i think not? | 02:46 |
| mntmn | esad, did you get the invite? | 02:47 |
| erlehmann | mntmn, why is the resolution 1366? also is it possible to make it 1360? | 02:48 |
| esad | @mntmn yes, thanks | 02:48 |
| esad | it's 6 pixels more than 1360 :) | 02:48 |
| esad | I guess 1360 would look even more "off" | 02:49 |
| erlehmann | hmmm | 02:50 |
| mntmn | erlehmann, the physical resolution is 1366x768 | 02:51 |
| mntmn | but maybe with xrandr it's possible to constrain the screen to 1366 again and not use the full virtual 1368 | 02:51 |
| erlehmann | oh, i can try that | 02:51 |
| mntmn | you can try to make a custom resolution | 02:52 |
| mntmn | i'm sure X has some kind of facility for this kind of stuff | 02:52 |
| erlehmann | https://a3nm.net/blog/xrandr.html | 02:53 |
| erlehmann | > It turns out that, when you want to use a mode that is not proposed in xrandr -q, you may be able to add one yourself. To create a mode for 1024x768 at 60 Hz, you first run: | 02:53 |
| erlehmann | gtf 1024 768 60 | grep Modeline | cut -d' ' -f4- | 02:53 |
| mntmn | but be careful with timings if that comes up | 02:53 |
| erlehmann | > Now you can run xrandr --newmode LINE, where LINE is the output of the above, and run xrandr --addmode DISPLAY NAME, where DISPLAY is the name of the display to which you want to add the mode, and NAME is the first field of LINE (the one in quotes). The new mode should now appear in xrandr -q, and you can try switching to it with --mode. This may fail in creative ways if your screen doesn't like the mode. | 02:54 |
| mntmn | like, if you do very extreme things it can have ghosting etc effects on the LCD that stays for minutes | 02:54 |
| erlehmann | i suspect you have tried various “very extreme things” to get that knowledge? | 02:55 |
| mntmn | yes | 02:55 |
| erlehmann | if so, i would like to know what not to do | 02:55 |
| mntmn | for example don't do a resolution that is smaller than the physical one | 02:55 |
| mntmn | like, smaller than 1366x768 | 02:56 |
| mntmn | i'll get you the modline, sec | 02:56 |
| mntmn | erlehmann: https://github.com/mntmn/reform-linux/blob/master/imx6qdl-mntreform.dtsi#L466 | 02:57 |
| mntmn | ah i also wrote a comment about the bug there | 02:57 |
| mntmn | so with 1366 apparently the /dev/fb0 device fails and there is maybe no kernel console... maybe it also has been fixed in the meantime | 02:57 |
| mntmn | this is one of the first problems i ever encountered during development of reform, more than a year ago | 02:57 |
| erlehmann | i see | 02:57 |
| erlehmann | mntmn, i thought maybe i could make it smaller :/ | 02:58 |
| mntmn | if the pixels are not refreshed they will light up in pathological ways IIRC | 02:58 |
| mntmn | but try adding 1366x768 with the timings from that dtsi | 02:59 |
| esad | oh, I just realized that reform has 3 spacebar keys :) | 02:59 |
| mntmn | esad yeah :) the original idea was to have them customizable, for more mouse buttons etc | 02:59 |
| mntmn | i actually mostly use the small space bars as spaces and almost never the middle | 02:59 |
| esad | yeah would be great for scrollwheel hotkey | 02:59 |
| esad | with the right key one could even scroll onehanded | 03:00 |
| mntmn | true | 03:00 |
| mntmn | or left key if you are left-handed ^^ | 03:00 |
| erlehmann | overscan | 03:02 |
| erlehmann | the word for cut-off borders on a screen is “overscan” | 03:02 |
| mntmn | trrue | 03:02 |
| erlehmann | i mention it just in case some of you might want to look for a solution | 03:02 |
| erlehmann | so you know what to search for | 03:02 |
| esad | true :) | 03:03 |
| erlehmann | https://github.com/rcuyegkeng/fix_overscan_with_xrandr | 03:03 |
| erlehmann | seems xrandr can do it? | 03:03 |
| erlehmann | let me try and report back! | 03:03 |
| erlehmann | ah, screen must support it :( | 03:04 |
| mntmn | funny https://html.brow.sh/https://en.wikipedia.org/wiki/Internet | 03:04 |
| esad | for some reason it looks very bad in net surf | 03:06 |
| esad | it looks very nice in safari | 03:06 |
| mntmn | meh | 03:06 |
| esad | could one click on a cookie consent popup via brow | 03:06 |
| mntmn | haha maybe yes | 03:07 |
| erlehmann | images do not load in dillo | 03:07 |
| mntmn | you can even watch youtube in browsh | 03:07 |
| mntmn | so what you would want is xrandr --output LVDS-1 --panning 1366x768 | 03:07 |
| mntmn | but it doesn't work | 03:07 |
| mntmn | with a larger number it exposes a funky etnaviv bug btw | 03:07 |
| mntmn | example: xrandr --output LVDS-1 --panning 1500x768 | 03:08 |
| erlehmann | funky? | 03:08 |
| mntmn | and then move the mouse out to the right | 03:08 |
| mntmn | you will end up in a more and more glitchy display | 03:08 |
| mntmn | moving to the left border will decrease the glitch | 03:08 |
| mntmn | i know this from developing my gfx card | 03:08 |
| esad | there's a newer version of netsurf (3.8) that got released this year - reform ships with 3.6 | 03:08 |
| mntmn | it's when the graphics scanout memory and the framebuffer memory layout don't match | 03:09 |
| mntmn | esad, so they don't have that new one in debian yet | 03:09 |
| erlehmann | esad, you can complain to debian probably, via „reportbug $(which netsurf)“ or something like that | 03:10 |
| mntmn | ok gonna get some sleep | 03:20 |
| mntmn | esad i always read logs in case you have more questions later | 03:20 |
| mntmn | good night! | 03:20 |
| erlehmann | good night | 03:21 |
| esad | also going to sleep. talk to you later! | 03:21 |
| - esad (QUIT: Quit: My MacBook has gone to sleep. ZZZzzz…) (~textual@84-113-242-212.cable.dynamic.surfer.at) | 03:21 | |
| - erlehmann (QUIT: Remote host closed the connection) (~erlehmann@x59cc894d.dyn.telefonica.de) | 03:24 | |
| + erlehmann (~erlehmann@x59cc894d.dyn.telefonica.de) | 03:24 | |
| + mrak (~textual@84-113-242-212.cable.dynamic.surfer.at) | 03:30 | |
| - mrak (QUIT: Quit: My MacBook has gone to sleep. ZZZzzz…) (~textual@84-113-242-212.cable.dynamic.surfer.at) | 03:39 | |
| - erlehmann (QUIT: Remote host closed the connection) (~erlehmann@x59cc894d.dyn.telefonica.de) | 03:46 | |
| + Jookia (~Jookia@gateway/tor-sasl/jookia) | 04:08 | |
| + erlehmann (~erlehmann@x59cc8959.dyn.telefonica.de) | 04:21 | |
| - Jookia (QUIT: Quit: Jookia) (~Jookia@gateway/tor-sasl/jookia) | 05:11 | |
| - erlehmann (QUIT: Quit: Leaving) (~erlehmann@x59cc8959.dyn.telefonica.de) | 05:53 | |
| mntmn | https://www.nytimes.com/2018/12/15/business/apple-california-manufacturing-history.html | 11:56 |
| + mrak (~textual@84-113-242-212.cable.dynamic.surfer.at) | 12:52 | |
| - mrak (QUIT: Client Quit) (~textual@84-113-242-212.cable.dynamic.surfer.at) | 12:55 | |
| + mrak (~textual@84-113-242-212.cable.dynamic.surfer.at) | 12:59 | |
| mntmn | the article is not very good but lead me to watch the “macintosh factory” videos where stuff looks hilariously overengineered for the time | 13:16 |
| - mrak (QUIT: Quit: Textual IRC Client: www.textualapp.com) (~textual@84-113-242-212.cable.dynamic.surfer.at) | 13:21 | |
| + mrak (~textual@84-113-242-212.cable.dynamic.surfer.at) | 13:21 | |
| - mrak (QUIT: Quit: Textual IRC Client: www.textualapp.com) (~textual@84-113-242-212.cable.dynamic.surfer.at) | 14:23 | |
| + mrak (~textual@84-113-242-212.cable.dynamic.surfer.at) | 14:24 | |
| * mrak -> esad | 14:25 | |
| - esad (QUIT: Quit: My MacBook has gone to sleep. ZZZzzz…) (~textual@84-113-242-212.cable.dynamic.surfer.at) | 14:36 | |
| + esad (~textual@84-113-242-212.cable.dynamic.surfer.at) | 14:53 | |
| andrej235 | and now for the first time in years I am a bit sad that I am not going to Xc3 anymore. | 14:58 |
| - esad (QUIT: Quit: My MacBook has gone to sleep. ZZZzzz…) (~textual@84-113-242-212.cable.dynamic.surfer.at) | 15:16 | |
| mntmn | andrej235, yeah, you'll miss out on the MNT assembly ;)) | 15:32 |
| mntmn | wow, x2go is incredibly fast | 15:32 |
| mntmn | i'm using it on reform now with my linux-macbook as a x2go server | 15:32 |
| mntmn | it's mind-blowingly fast | 15:33 |
| + esad (~textual@84-113-242-212.cable.dynamic.surfer.at) | 15:34 | |
| mntmn | esad, if you want a nice remote desktop experience on reform, x2go is the way to go | 15:36 |
| mntmn | for example, if you have a beefy linux server somewhere and want to use its cpu power on reform | 15:36 |
| esad | hm, how is latency for everyday usage? | 15:37 |
| esad | I mean stuff like editing, surfing.. | 15:37 |
| mntmn | almost zero | 15:37 |
| mntmn | that's why i'm so hyped atm... | 15:37 |
| mntmn | my faster computer is on the same LAN though | 15:38 |
| mntmn | i'll try it later with an AWS server | 15:38 |
| esad | what is the nearest AWS availability zone to germany/austria? | 15:38 |
| mntmn | germany ;) | 15:39 |
| esad | oh, that's nice | 15:39 |
| mntmn | frankfurt i think | 15:39 |
| mntmn | i recently spun up a GPU (nvidia tesla) machine there and tried to use it over vnc | 15:39 |
| mntmn | it was disappointing | 15:39 |
| mntmn | but i'll try again with x2go | 15:39 |
| mntmn | so i just have to put a talos secure workstation or sth similar on the internet to have a pretty libre mainframe-like experience ;) | 15:40 |
| mntmn | i'm typing this from firefox running on an AWS machine via x2go on reform | 16:01 |
| mntmn | also funny to be able to download stuff with 88MB/s on this machine | 16:08 |
| - esad (QUIT: Quit: My MacBook has gone to sleep. ZZZzzz…) (~textual@84-113-242-212.cable.dynamic.surfer.at) | 16:36 | |
| + esad (~textual@84-113-242-212.cable.dynamic.surfer.at) | 17:08 | |
| - esad (QUIT: Quit: My MacBook has gone to sleep. ZZZzzz…) (~textual@84-113-242-212.cable.dynamic.surfer.at) | 17:42 | |
| + esad (~textual@84-113-242-212.cable.dynamic.surfer.at) | 18:23 | |
| - esad (QUIT: Quit: My MacBook has gone to sleep. ZZZzzz…) (~textual@84-113-242-212.cable.dynamic.surfer.at) | 18:40 | |
| + esad (~textual@84-113-242-212.cable.dynamic.surfer.at) | 18:45 | |
| - esad (QUIT: Quit: My MacBook has gone to sleep. ZZZzzz…) (~textual@84-113-242-212.cable.dynamic.surfer.at) | 19:09 | |
| + esad (~textual@84-113-242-212.cable.dynamic.surfer.at) | 19:10 | |
| - esad (QUIT: Read error: Connection reset by peer) (~textual@84-113-242-212.cable.dynamic.surfer.at) | 19:29 | |
| + esad (~textual@p238-095.vps.tuwien.ac.at) | 19:29 | |
| esad | @mntmn and how is the latency? and that's probably the easiest way to get sublime on reform :) | 19:30 |
| mntmn | so the latency on AWS is bigger than to a computer in the LAN of course | 19:32 |
| mntmn | but it's still good... it seems to be optimized for text stuff | 19:32 |
| mntmn | and x2go is really easy to set up, it doesn't require any networking component except ssh | 19:33 |
| esad | ohhh https://www.youtube.com/watch?v=Dk306ZkNOuc is gold | 19:33 |
| esad | black and white tiled floor so robots can move around with precision :) | 19:34 |
| mntmn | yeah... part 2 is also very fun | 19:36 |
| esad | speaking of manufacturing - what is the unit cost for the 1.2 ghz reform device? i mean bom only | 19:36 |
| mntmn | i'm... not sure atm | 19:36 |
| mntmn | maybe i'll do a postmortem when all are built | 19:36 |
| mntmn | in the end it was probably more than the sale price | 19:38 |
| esad | but just as an estimate, do you think the 400-500 euro target I've read somewhere maintainable? | 19:38 |
| esad | is* | 19:38 |
| mntmn | i won't build the campaign reforms like this | 19:39 |
| esad | yeah you're probably also not counting hours spend on sanding the keycaps etc. :) | 19:39 |
| mntmn | it will have to be simplified | 19:39 |
| mntmn | yeah counting everything each reform cost maybe 2000 euro | 19:39 |
| esad | well it's all function of scale, right? it'd all be very different if you'd be building 100, 1000, 10.000 etc units | 19:40 |
| esad | I guess 10 devices is the worst scale | 19:40 |
| mntmn | yes of course | 19:40 |
| esad | because it's still theoretically doable by hand :) | 19:40 |
| esad | i mean single person's/small team hand | 19:41 |
| mntmn | yeah | 19:41 |
| mntmn | the case construction will be simplified, the board maybe simplified a bit as well | 19:41 |
| esad | but the idea is still that you can buy a preassembled reform from mnt? | 19:41 |
| mntmn | right now it has a few more high-end components | 19:41 |
| mntmn | yeah, i want to offer an "ikea" kind of kit as well | 19:42 |
| mntmn | i might also switch from cherry to kailh | 19:42 |
| mntmn | reduce the number of key shapes | 19:42 |
| mntmn | etc etc | 19:42 |
| esad | btw does ESC key have a different switch? | 19:42 |
| mntmn | no | 19:43 |
| esad | it seems to have different response | 19:43 |
| esad | maybe it's the backplane | 19:43 |
| esad | it has bit deeper travel and springier response | 19:43 |
| mntmn | hmm maybe keycap irregularity or because it's in the corner | 19:43 |
| mntmn | ah yeah maybe the pcb is a bit springy there :D | 19:43 |
| esad | what kind of switches did laptops in the early 90ies use? | 19:44 |
| esad | I remember those keyboards weren't that bad | 19:44 |
| esad | probably also cherrly ml :/ | 19:44 |
| mntmn | alps maybe? | 19:44 |
| mntmn | https://deskthority.net/keyboards-f2/http-appletothecore-me-files-mac-portable-disassembly-php-t8392.html | 19:45 |
| mntmn | https://deskthority.net/wiki/Alps_SKCM_Orange | 19:46 |
| mntmn | pretty similar to cherry ml internally | 19:47 |
| esad | yeah just found https://deskthority.net/wiki/Alps_SKFL_series | 19:47 |
| mntmn | but bigger | 19:47 |
| mntmn | i sampled these kind of switches and like them https://novelkeys.xyz/products/kailh-low-profile-switches | 19:48 |
| mntmn | also would halve the price of the keyboard possibly | 19:48 |
| esad | do you know if there's an external keyboard with those? | 19:49 |
| esad | oh there are plenty | 19:49 |
| mntmn | yeah there are some | 19:49 |
| esad | very interesting | 19:49 |
| mntmn | so that's also a factor, there are not many ML keyboards | 19:50 |
| mntmn | ML might even be discontinued in favor of cherry's low profile mx | 19:50 |
| esad | boahh I really can't stand the rgb led backlight | 19:50 |
| mntmn | yeah design choices nowadays... | 19:50 |
| esad | I understand the utility and it's really useful if done nicely but most of them look like tuned Golf GTI with neon light underneath or something | 19:50 |
| mntmn | yeah like most PC stuff sadly | 19:51 |
| esad | there's also https://deskthority.net/wiki/Topre_short-throw_keyboards :-) | 19:52 |
| mntmn | ah yeah i believe my HHKB lite has thos | 19:53 |
| mntmn | those | 19:53 |
| mntmn | ah no > membrane rubber domes switches | 19:53 |
| mntmn | my amiga 2000 keyboard also has similar switches like topre | 19:54 |
| mntmn | feels pretty nice | 19:55 |
| - esad (QUIT: Quit: My MacBook has gone to sleep. ZZZzzz…) (~textual@p238-095.vps.tuwien.ac.at) | 20:20 | |
| + esad (~textual@84-113-242-212.cable.dynamic.surfer.at) | 20:27 | |
| - esad (QUIT: Quit: My MacBook has gone to sleep. ZZZzzz…) (~textual@84-113-242-212.cable.dynamic.surfer.at) | 20:59 | |
| + esad (~textual@84-113-242-212.cable.dynamic.surfer.at) | 21:16 | |
| esad | I'm trying to get my intel wireless card running (it's a centrino N6205, should be supported by iwlwifi) | 21:16 |
| esad | btw. is it normal for non-full size pcie cards to hang diagonally | 21:16 |
| esad | after I install firmware-iwlwifi, modprobe says fatal module not found in /lib/modules/4.2.0-... | 21:18 |
| esad | am I missing something from the kernel | 21:19 |
| swivel | sounds like it | 21:36 |
| esad | are the dep files from reform kernel available somewhere? | 21:38 |
| swivel | i dunno, but if you have /proc/config.gz you can see if your driver was included in the running kernel config | 21:39 |
| esad | it wasnt | 21:41 |
| swivel | esad: what all do you see in /lib/modules/$(uname -r) ? | 21:43 |
| esad | I don't have /lib/modules at all | 21:43 |
| swivel | some distros include the kernel build machinery there, making it quite easy to build modules after the fact | 21:43 |
| swivel | oh | 21:44 |
| swivel | what does zgrep CONFIG_MODULES /proc/config.gz say? | 21:44 |
| esad | CONFIG_MODULES=y | 21:45 |
| esad | _TREE_LOOKUP=y | 21:45 |
| esad | and something about elf rel | 21:46 |
| esad | that's it | 21:46 |
| swivel | well at least the kernel supports loadable modules | 21:46 |
| swivel | i wonder where they are if not in /lib/modules | 21:46 |
| - esad (QUIT: Quit: My MacBook has gone to sleep. ZZZzzz…) (~textual@84-113-242-212.cable.dynamic.surfer.at) | 21:56 | |
| swivel | ah esad left, but for posterity sake the arch wiki has good info re: building kernel modules independent of the kernel https://wiki.archlinux.org/index.php/Compile_kernel_module | 22:05 |
| + esad (~textual@84-113-242-212.cable.dynamic.surfer.at) | 22:12 | |
| mntmn | esad sorry i didn’t ship any modules infrastructure yet, but we can piece it together later. or you clone reform-bootstrap and run mkkernel.sh | 22:52 |
| mntmn | like, in a few hours i could take a look. currently i’m not loading any modules and just put everything in the kernel | 22:52 |
| esad | so I should be able to load the intel module if I had the dep files? | 22:53 |
| esad | maybe you can put a tarball somewhere | 22:53 |
| mntmn | let me see | 22:53 |
| swivel | esad: you'll need the .ko | 23:01 |
| swivel | the dep file can be generated using depmod, it's not the critical piece | 23:02 |
| - esad (QUIT: Quit: My MacBook has gone to sleep. ZZZzzz…) (~textual@84-113-242-212.cable.dynamic.surfer.at) | 23:02 | |
| swivel | mntmn: if you wanted to continue shipping things that way, it would be convenient to provide an Arch style /lib/modules/$(uname -r)/build setup so it's easy to build arbitrary modules... or just use Arch instead of Debian ;) | 23:03 |
| + esad (~textual@84-113-242-212.cable.dynamic.surfer.at) | 23:04 | |
| mntmn | i'm very inexperienced with modules and always just compile a new kernel with the stuff i need, so i'll have to read up on how all these things work together | 23:09 |
| - esad (QUIT: Quit: My MacBook has gone to sleep. ZZZzzz…) (~textual@84-113-242-212.cable.dynamic.surfer.at) | 23:13 | |
| + esad (~textual@84-113-242-212.cable.dynamic.surfer.at) | 23:16 | |
| mntmn | esad: http://dump.mntmn.com/reform-kernel-modules-20181216.tar.gz | 23:17 |
| mntmn | i guess you can extract this to /lib/modules/<kernel-version> | 23:17 |
| mntmn | there is iwlwifi and some other stuff there, like coda | 23:17 |
| mntmn | (coda is the hardware h264 decoder that has to load a blob to work) | 23:18 |
| mntmn | oh interesting https://github.com/pH5/coda-bits | 23:19 |
| mntmn | this has some very interesting and new to me info about the coda unit in i.MX6 | 23:20 |
| - esad (QUIT: Quit: My MacBook has gone to sleep. ZZZzzz…) (~textual@84-113-242-212.cable.dynamic.surfer.at) | 23:22 | |
| + esad (~textual@84-113-242-212.cable.dynamic.surfer.at) | 23:23 | |
| mntmn | haha > The following information has been obtained by staring at the CODA firmware files distributed by Freescale, for a really long time, until patterns emerged | 23:23 |
| esad | :) | 23:23 |
| esad | yay I can see my wifi card | 23:23 |
| esad | which is hanging diagonally from the pcie port :) | 23:24 |
| mntmn | esad, after loading the module from my archive? | 23:24 |
| esad | yes just modprobe iwlwifi | 23:24 |
| mntmn | argh, better use some tape or get one of those cheap half-to-fullsize adapters :D | 23:24 |
| mntmn | esad, cool | 23:25 |
| esad | hm, wicd network manager doesn't see it, but it's there in ifconfig | 23:25 |
| mntmn | i guess i should just ship all kind of modules then | 23:25 |
| mntmn | yeah i don't know, people had a bunch of disappointments with wicd and stuff worked better with networkmanager | 23:25 |
| - esad (QUIT: Quit: My MacBook has gone to sleep. ZZZzzz…) (~textual@84-113-242-212.cable.dynamic.surfer.at) | 23:28 | |
| + esad (~textual@84-113-242-212.cable.dynamic.surfer.at) | 23:30 | |
| esad | ok it just needs interface name in preferences | 23:31 |
| esad | and rfkill tool | 23:32 |
| esad | yay and i'm online | 23:33 |
| esad | strength: 42% sitting right next to router - no antenna :) | 23:33 |
| - esad (QUIT: Quit: My MacBook has gone to sleep. ZZZzzz…) (~textual@84-113-242-212.cable.dynamic.surfer.at) | 23:36 | |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!