<Monni>
any attempt to enter boot menu won't work...
<Monni>
first start keyboard doesn't work in boot loader, second start, keyboard input crash boot loader ;)
<waddlesplash>
wait, so the new bootloader works on a cold boot?
<waddlesplash>
I don't remember you saying that before
bjorkint0sh has joined #haiku
<Monni>
waddlesplash: If I power down, the new bootloader tries to load kernel, but I can't access boot menu to change any boot options...
bjorkintosh has quit [Ping timeout: 480 seconds]
<waddlesplash>
I'm still confused
<waddlesplash>
does the new bootloader get to the desktop under any circumstances?
B2IA has quit [Quit: Vision[]: i've been blurred!]
<Monni>
waddlesplash: It does it I disconnect the keyboard...
B2IA has joined #haiku
<Monni>
I can still use mouse to open the about dialog to see which revision it booted...
<waddlesplash>
aha
<Monni>
so something goes wrong right after it tries to initialize keyboard
<waddlesplash>
what if you connect the keyboard after boot?
<waddlesplash>
and where does it crash when the keyboard is in? at the boot splash?
<waddlesplash>
the message you posted earlier
<Monni>
keyboard works in Terminal, but not in bootloader if I try to restart from Deskbar menu...
<waddlesplash>
okay
<waddlesplash>
so to summarize: if the keyboard is plugged in it crashes during boot?
<waddlesplash>
and if it's not plugged in it boots to desktop?
<Monni>
exactly...
<waddlesplash>
OK
<waddlesplash>
well that's important information
B2IA has quit [Quit: Vision[]: i've been blurred!]
B2IA has joined #haiku
<sen>
just to confirm since I'm still learning the Gerrit workflow, if I submitted something for review, it had some comments, I've fixed those and git added them, is the correct next step to "git commit --amend" then push again, rather than as a separate commit? I _think_ that's right based on the docs, but wanted to confirm before I mess something up :)
xet7 has joined #haiku
flag has joined #haiku
janoz has joined #haiku
flag_ has quit [Ping timeout: 480 seconds]
<nephele_xmpp>
Yes that is correct
<nephele_xmpp>
Aslong as the correct commit-id line is in the commit it will update the review
janus has quit [Ping timeout: 480 seconds]
janoz is now known as janus
<waddlesplash>
Monni: does the older bootloader let the keyboard work?
<waddlesplash>
sen: nephele_xmpp is correct. Also I forgot to mention, please hard wrap your commit message at 80 cols
<sen>
thanks! will do :)
erysdren has joined #haiku
chilledfrogs has joined #haiku
OscarL has quit [Quit: Gone with the wind]
kinkinkijkin has quit [Remote host closed the connection]
kinkinkijkin has joined #haiku
erysdren has quit [Quit: Konversation terminated!]
<waddlesplash>
it is a little nuts that vm_allocate_early_physical_page went from 50 source lines in r1beta5, to now (after pending patches are taken int account) it's 137
<waddlesplash>
tbh some of this init code should be shuffled into another file
<Monni>
waddlesplash : No problems with keyboard when using old bootloader...
HaikuUser2 has joined #haiku
HaikuUser2 has quit []
HaikuUser2 has joined #haiku
HaikuUser2 has quit [Read error: Connection reset by peer]
<Monni>
waddlesplash: I tried bootloader from pretty much every hrev between 58215 and 58228, so something around there broke keyboard input in boot loader... With all the different changes overlapped, it's hard to pinpoint which is the exact hrev that breaks keyboard initialization or input handling...
mmu_man has joined #haiku
flag_ has joined #haiku
flag has quit [Ping timeout: 480 seconds]
mmu_man has quit [Ping timeout: 480 seconds]
mmu_man has joined #haiku
janking has quit [Quit: Vision[]: i've been blurred!]
janking has joined #haiku
MisthaLu has joined #haiku
HaikuUser2 has joined #haiku
HaikuUser2 has quit [Quit: Vision[]: i've been blurred!]
Hannah has joined #haiku
Kernel86 has quit [Ping timeout: 480 seconds]
janking has quit [Quit: Vision[]: i've been blurred!]
mmu_man has joined #haiku
mmu_man is now known as Guest6428
Guest6428 has quit [Ping timeout: 480 seconds]
janking has joined #haiku
_-Caleb-_ has left #haiku [#haiku]
_-Caleb-_ has joined #haiku
DKnoto_W has quit [Ping timeout: 480 seconds]
DKnoto_W has joined #haiku
dalme has joined #haiku
MisthaLu has quit [Quit: Leaving]
mmu_man has quit [Ping timeout: 480 seconds]
janking has quit [Quit: Vision[]: i've been blurred!]
mmu_man has joined #haiku
<waddlesplash>
Monni: what's the first hrev that doesn't work?
<waddlesplash>
Monni: does 58215 work? or is it broken?
<Monni>
waddlesplash: I remember I got a crash with 58215, but not sure if it was in bootloader or kernel...
illwieckz has quit [Ping timeout: 480 seconds]
DKnoto_W has quit [Ping timeout: 480 seconds]
_-Caleb-_ has left #haiku [#haiku]
_-Caleb-_ has joined #haiku
illwieckz has joined #haiku
<Monni>
waddlesplash: 58215 crash both with own bootloader and with old bootloader...
<Monni>
KDL thinks it's NULL pointer deference in packagefs, which isn't surprising as packagefs didn't work in two earlier hrevs either, just didn't crash...
DKnoto_W has joined #haiku
<waddlesplash>
Monni: are you sure your system RAM has no problems?
<waddlesplash>
does the 58215 bootloader boot the new kernel?
<waddlesplash>
the current hrev kernel I mean
<Monni>
58215 crashes before it lets me select the kernel to boot...
<waddlesplash>
the bootloader does? ok
<Monni>
waddlesplash: selecting kernel needs packagefs and packagefs is definitely broken in 58215...
<waddlesplash>
Monni: kernel packagefs isn't initialized until the "disks" icon.
<waddlesplash>
so crashes in _lookup are much later and from totally different code than in the bootloader
<Monni>
waddlesplash: I get packagefs initialized during bootloader already...
<waddlesplash>
no?
<waddlesplash>
bootloader packagefs is very different
<Monni>
For me, if it fails to boot a kernel, it asks packagefs for another kernel...
<waddlesplash>
it does not have any methods with the signature you listed
<waddlesplash>
so if it's crashing in that packagefs_lookup it's not the bootloader
<Monni>
waddlesplash: bootloader has no serial output... first thing I see in serial output is "reboot" when it fails to load bootloader...
<waddlesplash>
the beta5 bootloader has no serial output, but master bootloader does
<Monni>
I'm not using beta5 bootloader, because that just works...
<waddlesplash>
okay
<waddlesplash>
does 58213 bootloader work?
<Monni>
waddlesplash: I can get to boot menu with 58213 bootloader...
ebruce613 has left #haiku [Error from remote client]
<waddlesplash>
OK
janking has joined #haiku
<waddlesplash>
I will have a few changes for you to try later then to see if we can narrow down what in 14 or 15 broke this
<Monni>
waddlesplash: It takes less than 5 minutes for me to build a new ISO to test...
_-Caleb-_ has left #haiku [#haiku]
novaphoenix has quit [Quit: gone fishing]
zard has joined #haiku
ebruce613 has joined #haiku
novaphoenix has joined #haiku
_-Caleb-_ has joined #haiku
<Monni>
4 minutes 54 seconds to build an ISO image ;)
<nephele_xmpp>
another day another day banned from oftc because their blacklist does not know what a dynamic ip is
<nephele_xmpp>
we really live in the future
<Monni>
hrev58214 doesn't let me go to boot menu...
<Monni>
would be really hilarious if the bootloader just runs out of heap space ;)
chilledfrogs has quit [Quit: connection reset by purr]
dalme has joined #haiku
chilledfrogs has joined #haiku
<nephele_xmpp>
nipos: librewolf keeps trying to connect to push.services.mozilla.com ;) don't have any notifications allowed i think. Interesting.
<Monni>
waddlesplash: bootloader tries to allocate 4327472 bytes too much... it basically runs out of memory...
jagtalon has quit [Quit: leaving]
jagtalon has joined #haiku
<nipos>
nephele_xmpp: Yes,that's the first thing I turn off at every new install.I think it should be off by default,but others seem to find push notifications more important than privacy 🙄 https://lemmy.ml/post/14437221
<nipos>
It's really useless in my opinion.Notifications still work perfectly fine while webpages are open.And when they're closed,I don't want them to annoy me anyways.
<nephele_xmpp>
what do you mean when they are closed?
dovsienko has joined #haiku
<nipos>
It works somewhat like notifications on smartphones: Servers send the notification to a centralized server operated by Mozilla and that sends it to your Firefox/LibreWolf install which then activates a Service Worker the website installed before to handle the notification and show a popup even while the page is currently closed.
<nipos>
If the site is currently open,that centralized Mozilla server isn't needed at all,the website can simply open a websocket to its own server.
nephele has joined #haiku
<nephele>
wtf nipos
<nephele>
that sounds super wierd
<nipos>
Yes,the modern web sucks,but that's how things are
<nephele>
nah. i don't have to implement this.
<nephele>
hi phschafft
<nipos>
I doubt anyone has a need for that.If you want a webpage to constantly distract you,just keep the tab open.
<nipos>
Hell no,please don't implement that.I only wanted to describe why LibreWolf does it and how it can be disabled.
* phschafft
waves to nephele.
<nephele>
opensnitch is a bit user unfriendly, i clicked deny on that request so it saved "Okay! until reboot librewolf is banned!"
<nephele>
which is not what i ment
<nephele>
phschafft: will this be a new coference url?
<nipos>
No need to ban it,simply disable it in about:config
<nephele>
nipos: i'm not going to configure every of those things manually. i'll just block it in the firewall
<nephele>
the OS also pings to ubuntu constantly to see if i have internet .-.
<nipos>
Set dom.push.enabled and dom.push.connection.enabled to false and to go sure,you can also set dom.push.serverURL to an empty string
<nipos>
Caring about privacy and using Ubuntu,funny xD
<nephele>
*shrug* what am i to do nipos? I can *and have* maintained linux distros other than it in the past. But really the more you veer of the "common ground" the harder it gets, and the more time investment the system sucks out of me
<nephele>
and i really don't want to get sucked into making my own linux distro
<nephele>
manjoro also pinged home constantly
<nephele>
phschafft: i think that is fine :)
<nipos>
There's so much shit going on in the background on modern operating systems.What do these developers think?!Oh,maybe 0.01% of users may need to visit a webpage before the Wifi works,let's just cause a background connection for 100% of them,who cares?
<nephele>
tried to copy the url out of the discourse edit log, this is what it came up with
<nipos>
I don't use Linux anymore,it just sucks.If Haiku doesn't work,I use some BSD.FreeBSD is good in most cases,sometimes OpenBSD has better Wifi drivers.But if I were *forced* to use Linux,I'd at least pick one that is community-maintained and systemd-free and free of container nonsense.
<phschafft>
nephele: when do you need me to open the door?
<nephele>
systemd-free doesn't mean "proper s6 based service supervisor" sadly
<nephele>
phschafft: depends, i am ready now, but 10 minutes before 16 should be plenty of time to prepare
<nipos>
That's true.systemd-free means better than others,but it's still Linux.And sucks.
<nephele>
if you have configured your services properly systemd is somewhat competent at supervising them, and that is nice
<phschafft>
nephele: ok.
<nephele>
if i go systemd-free i often get sysvinit scripts maintained *somehow* without much sense
<nipos>
FreeBSD rc.d is rather competent,too
<nephele>
if i want proper service supervision with for example s6 i have to provide the stuff myself
<nephele>
nipos: doesn't that also use pidfiles...?
<nipos>
I've written rc.d scripts for many things on my servers because often there aren't premade packages and it works fine so far
<nipos>
Yes,it uses pidfiles
<nephele>
see, pidfiles are already a major design flaw
<waddlesplash>
Monni: where do you see that? what log message?
<nipos>
Oh, that looks really interesting. Not sure if Haiku qualifies with the criteria, but seeing that, with FreeBSD, another operating system is already supported... You have nothing to lose by trying it, right?
<waddlesplash>
nephele: I don't think we qualify as "critical open source technology"
carlomonte has joined #haiku
carlomonte has quit []
JakeSays has joined #haiku
bbjimmy has quit [Quit: Vision[]: i've been blurred!]
bbjimmy has joined #haiku
_-Caleb-_ has left #haiku [#haiku]
_-Caleb-_ has joined #haiku
Coldfirex has joined #haiku
_-Caleb-_ has left #haiku [#haiku]
<nipos>
nephele: Does Jitsi work on Haiku?
<nipos>
I'd prefer using a desktop computer this time as my laptop got very hot last time,but on the OpenIndiana one I don't get audio... :/
_-Caleb-_ has joined #haiku
HaikuUser2 has joined #haiku
<waddlesplash>
it worked on Falkon last time I tried actually
<nipos>
I'm now connected using GNOME Web,which also seems to work.I'm positively surprised
<nipos>
Nothing to see and hear yet,but that's probably because the stream is supposed to start in 5 minutes
<nephele>
i already have video running
HaikuUser2 has quit [Quit: Vision[]: i've been blurred!]
<nipos>
Ok,will retry in Falkon
* phschafft
nods.
<phschafft>
I see now media streams from you.
<nipos>
In Falkon,the CPU goes up a lot more and I still have no video.Nice.Will boot the FreeBSD laptop then
ebruce613 has left #haiku [Error from remote client]
<Begasus>
finaly packaged :) grabbing r2ghidra-5.9.4-1-x86_64.hpkg and moving it to /Share/haikuports/packages/r2ghidra-5.9.4-1-x86_64.hpkg
<phschafft>
zard: ping?
<zard>
pong :)
<Begasus>
using a git clone, so not ready for haikuports
<phschafft>
zard: you'll join?
<zard>
Hmm, let's see...
ebruce613 has joined #haiku
<zard>
got to do something first, though
<phschafft>
the door is open, feel free to come in any time :)
<waddlesplash>
Monni: do you actually get the "platform_allocate_heap_region: region end is beyond identity map"?
<Monni>
waddlesplash: Yup...
<waddlesplash>
OK. so then what's the deal with the KDLs you said happened?
<waddlesplash>
i.e. the ones in the bootsplash methods
<waddlesplash>
if we get "beyond identity map", that's long before the KDL backtrace you posted in the ticket
<Monni>
waddlesplash: "beyond identity map" only happens if I try to enter boot menu...
<waddlesplash>
really? now that's odd
<Coldfirex>
@waddlesplash: some ya
<Coldfirex>
The KDLs were from the hrevs prior. No more today though
<waddlesplash>
the log you attached has a KDL in it
<Monni>
waddlesplash: I haven't looked yet into what it actually does when entering the boot menu... I got only as far as checking how much memory it tries to allocate when it crashes...
<waddlesplash>
OK
<waddlesplash>
I wonder if this is related to framebuffer sizing somehow
<zard>
phschafft: It barely works out right now, but I'll take a peek. 16:30 UTC is the closest time that would work well for me
_-Caleb-_ has left #haiku [#haiku]
<nephele>
don't shoot the messenger ;)
_-Caleb-_ has joined #haiku
tuaris has joined #haiku
HaikuUser2 has joined #haiku
<phschafft>
zard: no worries from my side. just thought it might be of interest.
HaikuUser2 has quit []
<zard>
Aye, it is somewhat interesting :)
<waddlesplash>
ok I managed to reproduce Coldfirex's crash at least by creating a massive swap file and then trying to boot in QEMU with 100G RAM
<waddlesplash>
I still haven't seen Monni's crash
mmu_man has quit [Ping timeout: 480 seconds]
linuxmaster has quit [Ping timeout: 480 seconds]
<Coldfirex>
Sorry, I mean the white kdl that pops up on screen
<Coldfirex>
Thats what I no longer get
<nephele>
zard: what time generally works for you?
<nephele>
could make the next stream later in the day also :)
<zard>
around 16:30-17:00, or 19:15-20:00
<nephele>
time zone?
<zard>
UTC
<nephele>
somewhat short intervals anyhow :)
<nephele>
i think the latter would be too late for here
<Monni>
waddlesplash: It takes precise timing to hit the sweet spot for the crash, especially with too fast machine... There is only fraction of a second before it's already loading the kernel...
<waddlesplash>
you don't need to spam spacebar
<waddlesplash>
with the BIOS loader, you can hold SHIFT pretty much as soon as the system BIOS splash appears
<waddlesplash>
and it will open the menu
<waddlesplash>
and I thought you said the kernel crashes while booting anyway?
<Monni>
waddlesplash: I select reboot from Deskbar and hold shift while it shuts down... I release shift key when it either crash or show boot menu...
<waddlesplash>
I would hold SHIFT only at the machine bootsplash, before the OS load starts
<waddlesplash>
BIOS may not detect SHIFT held before that
mmu_man has joined #haiku
<Monni>
waddlesplash: My machine is too fast to display any BIOS screen...
<waddlesplash>
hm, ok
<Monni>
6 core i5.9600k with 3 cores disabled is quite fast to boot Haiku ;)
<waddlesplash>
well the early boot process has no SMP anyway
<waddlesplash>
we only have SMP around the time the first icon lights in the bootsplash
_-Caleb-_ has left #haiku [#haiku]
<Monni>
waddlesplash: This still clocks at 3.66 GHz per core...
_-Caleb-_ has joined #haiku
freddietilley has quit [Quit: WeeChat 4.4.2]
talos has quit [Quit: talos]
<nipos>
I should really try to get one of my desktops to work for the next stream.I don't like my laptop getting that hot.It's one of those flat fanless things
<nephele>
can send you a macbook in the mail :P
BrunoSpr has joined #haiku
<nephele>
hmm, but i don't think that build would finish anytime soon. not sure why it wants to rebuild everything again ugh. Maybe I'll let it build "natively" later instead
<nipos>
If you don't need it anymore,why not? :P I collect computers and laptops of all types
<nephele>
i want to port Haiku to this one
<nipos>
Some day,when I finally got into driver development,they should all run Haiku.I think nowadays about 60-70% already are,the rest has major driver issues
<nipos>
Major driver issues = There aren't enough USB ports to replace all the things that don't work internally lol
<nephele>
use usb hub
<nipos>
That technically works,but it's rather inconvenient for a laptop.And seriously,I want at least internal storage to always work.Touchpad/touchscreen would be cool,too,but can be replaced with a mouse.And network,well,if it's Broadcom,I just use a USB dongle
janking has quit [Quit: Vision[]: i've been blurred!]
<nipos>
Jitsi is a great solution for the stream btw.I wouldn't have joined if it were on some proprietary garbage.My issues are mostly a result of that fact that I nearly never need audio on my computers and therefore don't have a properly working setup.
<nephele>
I don't like "web tools" too much
BrunoSpr has quit [Quit: Vision[]: Ich wurde eingeweicht!]
<nephele>
but yes, I wouldn't use some random prop stuff of course, after all it has to run somewhere i can use it too
<nephele>
come to think of it, ideally, something like Renga would be what we want for such a straem. Make it be able to send and receive jingle streams. Implement stuff like "voice channels" etc
BrunoSpr has joined #haiku
<nephele>
I've ended the stream for today, can't really do much with those build times
<nipos>
Video streaming over XMPP,you mean?Never seen that anywhere before.And what Matrix does is just embedding Jitsi.
<nephele>
XMPP has severall functions for streaming stuff, one "jingle" is mostly compatible with webrtc
<nephele>
but nothing is stopping you from implementing that natively properly in an application
<nipos>
Something that works natively today would be hosting a MP4 livestream on some server and load that using a native video player application
<nipos>
Would be more work on your site,however,for hosting that.
_-Caleb-_ has left #haiku [#haiku]
<nephele>
probably, but then at the same time i want something like a chat too
<nephele>
and bidirectional comms
<nipos>
That wouldn't work so easy in a native application.Implementing this would be an effort of multiple weeks,if not months,I think
heisarun[m] has joined #haiku
<nephele>
Sure, it would take quite some time :)
ebruce613 has left #haiku [Error from remote client]
_-Caleb-_ has joined #haiku
<BrunoSpr>
Where was your stream?
<BrunoSpr>
Hello all,
<BrunoSpr>
sorry fiddling around
ebruce613 has joined #haiku
OscarL has joined #haiku
<Begasus>
hi BrunoSpr, stream ended
<Begasus>
hi OscarL! :)
<Begasus>
sleeping in are we?
<BrunoSpr>
Begasus, ah ok next time then, I will join... just came from work
<OscarL>
Hello Begasus. Doing some stuff around the house, to make up for lost time spent when being in too bad of a mood to do much :-)
<Begasus>
yay! build/packaging fine now! :D grabbing r2ghidra-5.9.4-1-x86_64.hpkg and moving it to /Share/haikuports/packages/r2ghidra-5.9.4-1-x86_64.hpkg
<OscarL>
Even so... got an app_server crash earlier today: https://bpa.st/5VBA (app_server crashes are quite rare for me. Can't even remember the last time I saw one)
<Begasus>
BrunoSpr, it's announced in the forum, so you can follow nephele's progress there :)
<BrunoSpr>
ah ok thank you
<Begasus>
np, have been tackling radare2/iaito/r2ghidra today :)
<OscarL>
Begasus: is that the full ghidra or the "plugin" version for radare? (only remember some datails from those from a while back)
Monni has quit [Remote host closed the connection]
<Begasus>
to have or not to have ... :)
<Begasus>
"It is solely based on the decompiler part of Ghidra, which is written entirely in C++, so Ghidra itself is not required at all and the plugin can be built self-contained."
<OscarL>
yeah, I remember suggesting to try to use that instead of attempting the full Ghidra (in the context of radare/iaito).
<phschafft>
nephele: all good?
<nephele>
ok so wayland does just not work *shrug*
<nephele>
phschafft: i think ubuntu needs to be reinstalled soon...
<Begasus>
can't remember the details, but seeing they release 5.9.6 for radare2/iaito I though I'd gave it a spin :)
<OscarL>
Begasus: nice work :-). (now... wish I remember for which task I thought it might be of use to me, lol! :-D)
<Begasus>
heh
<Begasus>
well, without the buildmaster up and running it could take some time to merge these PR's :)
<nephele>
phschafft: do you have something more on your mind you want to talk about now? otherwise i'd suggest tommorow :)
<OscarL>
"_IO_fputs" maybe it is trying to write to stdout and when launched from Tracker it doesn't has that FD ?
<Begasus>
no idea :D
<OscarL>
(to anyone that actually understand this stuff... sorry for my uneducated stabs in the dark :-D)
dby has joined #haiku
<OscarL>
"ConsoleWidget::processQueuedOutput()" before that... yeah... to me it sounds like it is just trying to write to a stream/fd that it is just not there when launched from Tracker.
ebruce613 has left #haiku [Error from remote client]
BrunoSpr has quit [Quit: Vision[]: Ich wurde eingeweicht!]
<OscarL>
Begasus[m]: as I see it, it crashes here: https://github.com/radareorg/iaito/blob/master/src/widgets/ConsoleWidget.cpp#L461 . The crash is a segment violation, so either "origStderr" or "output" variables is poiting to a "Bad address" (and output just got created the line above so... "origStderr" must be just pointing to the wrong place when launching from Tracker)
<OscarL>
why the difference between Tracker/Terminal there... no clue (it is not the usual: "we need to use chdir()" case we see often :-))
<Coldfirex>
@waddlesplash: when should the change for 58240 show up under software updater? Its not showing anything yet
<waddlesplash>
nightlies only build once a day
<waddlesplash>
however there is a change on Gerrit that should pick this up soon
<Coldfirex>
oh, I thought software updater was a separate process. ok
nephele_xmpp has left #haiku [Error from remote client]
ebruce613 has joined #haiku
<OscarL>
regarding the app_server crash I got... (segment violation on "ServerWindow::GetInfo(window_info& info)")... might need an "if (fWindow != NULL)" in there?
<waddlesplash>
I think there's a report for this
<Begasus[m]>
"we need to use chdir()" iirc tried that approach on a previous version OscarL
<nipos>
I just noticed that depending on the background image,the labels of the desktop icons can be very hard to read.Is there an easy fix?
<botifico>
[haikuports/haikuports] Begasus 85f1383 - radare2, bump version (#11214)
<OscarL>
nipos: set "Icon label outline" in Backgrounds preflet maybe?
<nipos>
Just found it myself
Coldfirex has quit [Quit: Page closed]
<nipos>
The German translation "Icon-Textkontur" wasn't that obvious to me
arjen_ has joined #haiku
jmairboeck has joined #haiku
nephele_xmpp has joined #haiku
nephele_xmpp has left #haiku [Error from remote client]
nephele_xmpp has joined #haiku
<Begasus[m]>
OK, 32bit buildmaster stuck again I think
<OscarL>
Seems so. "moving package ponpokodiff*.hpkg to cache" at 17:15... is not that big of a package to require 15 minutes to continue with the next recipe :-(
<Begasus>
right
<waddlesplash>
workaround was lost
<waddlesplash>
applied it again, rebooting gcc2 builder
<Begasus>
putting version string to radare2 for iaito so it doesn't take the wrong version
<phschafft>
(now only a milion more ports before I have a functional editor! ;)
<nephele>
now to find a ssd or something to install debian
<phschafft>
I still have some unformatted paper tape around.
<nephele>
is that paper tape fast enough for 4k videoplayback?
<phschafft>
considering how much better the codes got the last few years it soon will be!
ebruce613 has joined #haiku
<phschafft>
(I mean stereo waveform music on paper tapes is clearly possible. (MIDI style music is something paper tapes have been used for way too many years to count))
gouchi has joined #haiku
higen has quit [Ping timeout: 480 seconds]
mmu_man has quit [Ping timeout: 480 seconds]
nephele has quit [Quit: Vision[]: i've been blurred!]
<phschafft>
I'm wondering about small string handles again. hm.
HaikuUser has joined #haiku
HaikuUser has quit []
<B2IA>
(ashtons) hello
Atomozero has joined #haiku
janking has joined #haiku
arjen_ has quit [Quit: Vision[]: i've been blurred!]
gouchi has quit [Quit: Quitte]
Atomozero has quit [Quit: Vision[]: i've been blurred!]
diver1 has quit [Ping timeout: 480 seconds]
mmu_man has quit [Ping timeout: 480 seconds]
_-Caleb-_ has left #haiku [#haiku]
_-Caleb-_ has joined #haiku
mmu_man has joined #haiku
diver has joined #haiku
nephele has joined #haiku
<nephele>
hi :)
AlienSoldier has joined #haiku
<phschafft>
mau.
nephele has quit [Read error: No route to host]
nephele has joined #haiku
<AlienSoldier>
waddlesplash i updated 228 to 239. Laptop boot fine, perhaps it was nstal corruption all along. I don't know. I don't see much memory saving on my 32bit laptop, from 201, only 3M lower.
<nephele>
phschafft: oh man. Bin gerade echt happy. Debian netinstall gebootet. Und die haben ein "dark contrast accesible installer menu"
<phschafft>
haha. :))
<nephele>
ich bin so froh, ich habe das noch nie gesehen bei anderen installern :)
<phschafft>
dann hat es ja einen pluspunkt schonmal.
deneel has quit [Remote host closed the connection]
nephele_xmpp has left #haiku [Error from remote client]
<nephele>
sie lassen mich aber kein leeres passwort setzen :/
<phschafft>
hm.
<nephele>
phschafft: debian has debootstrap, would be very funny to make a Haiku based debian installer ;)
<phschafft>
if you get dpkg working, why not?
diver has quit [Ping timeout: 480 seconds]
<nephele>
hmm, debootstrap install seems to tie in with the idea of chrooting into the new system, that might be a bit troublesome
<phschafft>
no chroot support?
<nephele>
sure haiku has chroot
<nephele>
but it doesn't do linux ELF files
<nephele>
and syscalls :)
janking has quit [Quit: Vision[]: i've been blurred!]
<phschafft>
hm.
<nephele>
not sure i'm onboard with debian wanting to create a 1GB swap partition
<nephele>
seems kind of pointless. if you want swap atleast make it a bit bigger...
<phschafft>
you can skip it if you like.
<phschafft>
nephele: debootstrap supports installing different archs. so maybe that is the starting point.
<nephele>
Dunno. I'm not searching for a haiku compatible arch ;). It would be cool to be able to directly install to different disks without having to reboot.
<nephele>
Maybe qemu can help, but maybe i can also look at how debians install does stuff internally and what can be implemented "generically"
<nephele>
also maybe for a potential OpenBSD install in a similar vain
<nephele>
probably some basics of Haiku supporting whatever FS is used properly, and then writing files to it :D
<phschafft>
what I mean: debootstrap can install for archs that are not supported by the host CPU.
<phschafft>
meaning it cannot run binaries.
dby has quit [Quit: Vision[]: i've been blurred!]
<phschafft>
running on a different host OS might be a similar case.
<nephele>
Yes. It can. But the debian installation instructions suggest to use chroot to install, among other things, a linux kernel image *after* using debootstrap
<nephele>
implying that debootstrap on it's own does not create a bootable system
<phschafft>
yes. that is exactly the point of it.
<phschafft>
it installs Debian, not any boot system.
<nephele>
yep, so for a booting debian we need atleast a linux kernel too :)
<phschafft>
for systems like Haiku that is all the same.
<nephele>
and the debian instructions suggest installing it via apt in a chroot
<phschafft>
but in the GNU world the userland is the OS and the kernel is the kernel and they are completly independent. and the bootloader is basically another system.
<phschafft>
yes, but that is similar to 'if you like to use some office software go and install it'.
<phschafft>
it is just outside of the scope of debootstrap.
<nephele>
yes, i don't anticipate many problems installing a bootloader however, if i can properly obtain it
<phschafft>
debootstrap is to install Debian, not to make a working machine.
<nephele>
I get what you are saying, but this feels like arguing semantics. I do want to have a working partition with debian as the end goal
<phschafft>
also keep in mind that you can boot a Debian perfectly fine on a kernel that is not managed at all by said system.
<phschafft>
so there is totally no problem to install a kernel from Haiku.
<nephele>
Sure, just the question which tools to use :)
<phschafft>
IF that is your goal than debootstrap is the wrong tool.
<nephele>
chroot seems badly suited from haiku. Perhaps porting dpkg
<phschafft>
semantics are important here as they are the very reason for debootstrap to exist.
<nephele>
Okay. what tool would you suggest instead of debootstrap?
<phschafft>
if you want a ready made system, use the installer.
<phschafft>
that is the tool for that.
<nephele>
That doesn't satisfy the requirement of not booting into another system. hmm
<phschafft>
but it's the tool for *ready made system*.
<nephele>
It is one tool
<phschafft>
which includes the OS, the kernel, the bootloader, but also config, tasksel and all those things.
<phschafft>
you could think about debootstrap as one of the tiny(!) little gears inside the installer.
<nephele>
yep. I guess the "easy" solution is running whatever installer image in qemu with exposing the disk device
<phschafft>
yes.
<nephele>
I could try how that works, but i expect this to be quite slow :D
<phschafft>
or you just install the kernel and the bootloader via Haikus package system.
<phschafft>
just as Debian does it.
<nephele>
what do you mean by via haikus package system?
<nephele>
I don't think debian can read .hpkg
<phschafft>
it doesn't need to!
<phschafft>
/boot is not part of the system.
<phschafft>
on Debian.
<phschafft>
there is no problem in having your /boot managed by Haiku and not even mounting it on Debian.
<nephele>
Assuming EFI boot, is it expected to have a seperate /boot partition at all?
<nephele>
I guess that could be done anyhow, but then how would updating a kernel image from debians side look?
<phschafft>
I mean you just need to put the kernel somewhere in your BFS and chainload it from the Haiku boot process at whatever stage you like and Debian will be happy with that.
<phschafft>
(doing something like that is actually very common in Linux world :)
<nephele>
that sounds a bit crazy, but in a good way :P
<phschafft>
the point is that the kernel is not part of the system. it's an independent component.
<nephele>
I doubt this is easy to maintain though, one nice thing about efi is that it makes co-installing OSes (on a desktop) relatively painless
<phschafft>
so if you like to manage it from within the system. fine. if you like to manage it from outside. also fine.
<nephele>
It is only partially independent, I expect to be able to update it from Debian
<nephele>
Having a "rescue linux" would still be cool though, to have.
<phschafft>
but that is just your thought model. it is technically independent. end of story.
<nephele>
Not sure what your argument is? one part is hooked up with apt, the other is not. I agree both can be done and are viable. But I don't want to maintain it from the outside, since now my debian installation depends on the correct operation of my haiku installation too
<phschafft>
kind of. maybe. maybe not.
<nephele>
linux atleast supports bfs only (oficially) in a RO capacity
<phschafft>
the point is: you wanted an alternative to the installer. here is one. if you don't like it, go for yet another one! there surely are many if we spend a bit time thinking about it.
<nephele>
:)
<phschafft>
plus, just because you installed a kernel from Haiku doesn't mean you cannot install another one from within Debian.
<phschafft>
Debian is totally fine with installing stuff in /boot with a /boot that is shared.
<phschafft>
(it has all those nice safeguards to not break things :)
<nephele>
I like the idea of a rescue linux. I once build myself one that was a couple of megabytes big, with an included initramfs that included the toybox set
<phschafft>
so if you install a kernel from haiku and boot debian with /boot mounted than you can after you booted just do the very same steps as outlined in that debootstrap+chroot+apt example.
<nephele>
nice thing is, with an efistub boot, you can stuff this into efi system partitions ;)
<phschafft>
after that you have one kernel managed by Haiku and one managed by Debian.
<phschafft>
and they are all friends.
<nephele>
By the way phschafft, we were talking about OS compat before. I was thinking about OS compat in a different way too. For the "<FOSS> on phone" projects, it seems there are no nice protocols for a desktop OS to manage this
<phschafft>
hm?
<nephele>
For example with an egg phone and a mac you can connect those together, and the computer can browse the phone, backup it, reinstall it. etc.
<nephele>
For browsing there are somewhat working protocols, but with manual setupo
<nephele>
but there seems to be nothing easy for those phone OSes to distribute firmware images so i could just reinstall the phone, or protocols to back them up in a lowlevel fashion and such
<phschafft>
little disclamer: I don't own a cell phone myself. so my perspective is limited to more dev tool world and I may not be aware of user software.
<phschafft>
not fully sure what you're asking for.
<phschafft>
only knowing the Android side. and there you have a number of options of doing different kind of things.
<nephele>
I want to figure out what software exists, or can exist, for the usecase of owning a haiku desktop, and using that to backup, restore, etc a phone
<nephele>
the fosdem linux on phone stands didn't have an answer either other than "uhh, just use ssh?"
<nephele>
so don't worry too much if you don't know either :)
<phschafft>
on Android you can interact with the filesystem (basically mount it if you like), you can run commands on it (including interactive PTYs), you can interact with what is installed, ....
<phschafft>
most system software is also a package.
<phschafft>
and you can download them.
<nephele>
i mostly know how the android side of things works, but other than lineageos I am not aware of many "freeish" distros that don't force google down your throat
<phschafft>
there are a number of command line tools to do all kind of stuff like working with the phone's IO and stuff.
<nephele>
just that the linux on phone stuff doesn't seem to have much of an equivalent
<nephele>
hmm, i wonder when some desktops on linux will figure out that setting my native refresh rate might be a cool idea...
<phschafft>
hu?
<phschafft>
is it not reported correctly?
<nephele>
they keep setting 60hz or 59.99999hz per default or something
<phschafft>
does xrandr list the setting you like?
<nephele>
sure it does
<nephele>
i just wish it would be picked by default, and i would not have to configure this
<phschafft>
is it marked with an '+'?
mmlr has quit [Quit: Vision[]: i've been blurred!]
<nephele>
the refresh rate? not sure. There is a + between 60.00 and 240.00
diver has joined #haiku
<phschafft>
the marking is *after* the mode it applies to.
<phschafft>
that means the monitor sends 60Hz as the preferred value.
<nephele>
I don't prefer it :P
<phschafft>
'+' is what the monitor sends as preferred. '*' what is currently selected.
<nephele>
with adaptive sync this would not matter as much....
<nephele>
Ideally, anyhow. Not sure how that is taken into account by systems currently
<phschafft>
I mean the system asked your monitor for it's preferred mode. if you like something different you need to talk to your system and tell it.
<nephele>
The problem, to me, is that the way the api works none of my applications can decide to use the higher refresh rate if the system configures it for 60hz
<nephele>
Isn't adaptive sync ment to solve exactly this problem? keep a low resresh rate when nothing is updating, and have a higher one if stuff happens :g?
diver1 has joined #haiku
diver has quit [Read error: Connection reset by peer]
coolcoder613_ is now known as coolcoder613
<nephele>
phschafft: okay, got the backports kernel, got backports mesa... let's hope the system boots :D
<phschafft>
the point is that the system depends on what the monitor announces.
<phschafft>
if it announces something strange the OS can't do much about it.
<nephele>
it announces all supported refresh rates
<phschafft>
yes, but the OS selects the one that the monitor announces as preferred.
<nephele>
hmmm, lxqt is quite slow :(
<phschafft>
my personal guess is that it annouces something not full data rate as windows seems to have huge problems if the monitor announces something the cable doesn't support.
<nephele>
I doubt you can use the full 240hz without display stream compression for this monitor
<phschafft>
(it will just ignore that the cable doesn't support it and you go with no picture and no way to fix it ;)
<nephele>
uhh, isn't link training ment to figure out what the cable supports?
<phschafft>
;)
<nephele>
I'm still curious how adaptive sync plays into this
Anarchos has quit [Quit: Vision[]: i've been blurred!]
<phschafft>
but windows seems to be 'nahhhh! go for it! full power murdochs!
<nephele>
haven't found good documentation for that :(
<phschafft>
-s
<nephele>
waaah, debian also doesn't package noto emoji :(
<nephele>
I ment of course: 😭
nephele_xmpp has joined #haiku
<phschafft>
haha.
<nephele>
there is a package for the inferior version though, noto color emoji
xet7 has quit [Remote host closed the connection]
xet7 has joined #haiku
diver1 has quit [Ping timeout: 480 seconds]
<nephele>
phschafft: anyway, thanks for talking about debian with me. so far I am happy with the setup. Now hoping i can also start the games ;)