IRC channel logs

***Server sets mode: +nt

***bpsecret- is now known as bpsecret

<p4r4D0xum>Hello, is there any git client on Debian GNU/Hurd? Tried to install breezy but lacks python3-breezy if I'm not mistaken,

***rekado_ is now known as rekado

<civodul>janneke: so it seems to be our pc_kbd driver that's somehow not reading keyboard input

<civodul>(speaking about the console client)

<janneke>civodul: hmm

<janneke>civodul: did you see our: fsysopts /servers/socket/2 => Translator died

<janneke>?

<janneke>i wonder what we're missing...

<civodul>what does showtrans say?

<civodul>could be just a wrong file name or something, no?

<janneke>/hurd/pfinet --interface eth0 --address 10.0.2.77 --netmask 255.255.255.0 --gateway 10.0.2.2 --ipv6 /servers/socket/16

<janneke>hmm, ls -l says: (only for socket/2): ???????? ? ? ? ? 2

<civodul>dunno

<damo22>why ipv6?

*janneke thought civodul knew all these hurd'y things

<civodul>heh, don't overestimate me :-)

<janneke>damo22: puzzles me too

<civodul>try running it explicitly

<civodul>with "settrans -a /servers/socket/2 /hurd/pfinet ..."

<civodul>and try without IPv6, too :-)

<gnu_srs1>On a working qemu image: /hurd/pfinet --interface=/dev/eth0 --address=10.0.2.15 --netmask=255.255.255.0 --gateway=10.0.2.2 --address6=fec0::5054:ff:fe12:3456/64 --address6=fe80::5254:12:3456/10 --address6=fe80::5054:ff:fe12:3456/10 --gateway6=fe80::2

<gnu_srs1>Note /dev/eth0

<civodul>oh indeed

<damo22>i would drop the ipv6 option

<janneke>gnu_srs1: right!

<janneke>damo22: okay... debian hurd has: showtrans /servers/socket/2

<janneke>/hurd/pfinet -6 /servers/socket/26

<janneke>which also puzzles me (but i probably should re-read the docs)

<gnu_srs1>fsysopts /servers/socket/26 gives the same result: for more info see: debian/local/setup-translators

<civodul>26? 16?

<civodul>janneke, rekado: http://web.fdn.fr/~lcourtes/tmp/hurd.png !

<janneke>civodul: \o/

<janneke>hmm, with /dev/eth0 the translator still dies

<janneke>hmm https://www.gnu.org/software/hurd/hurd/translator/pfinet.html mentions -i eth0

<civodul>then also note that Debian uses netdde (user-space drivers) while we don't

<civodul>so there's a translator on /dev/eth0 on Debian

<civodul>but not in our case

<civodul>i suppose in our case we're using Mach devices

<civodul>so "eth0" makes sense: it may be the name of the Mach device

<civodul>if you rpctrace it, perhaps you see a device_open("eth0")

<gnu_srs1>Have you enabled the build of mach's ethernet drivers? Otherwise add nettdde to the build.

<civodul>good point, we should try --enable-device-drivers=qemu

<civodul>dunno how that differs from "default"

<janneke>civodul: afaics debian/rules only has --enable-kbd on gnumach

<civodul>janneke: but they prolly don't use the in-kernel drivers anymore

<civodul>well, --enable-device-drivers=defaults prolly includes some drivers, i haven't checked

<janneke>ah!

<civodul> https://guix.gnu.org/blog/2020/a-hello-world-virtual-machine-running-the-hurd/

<civodul>woohoo! :-)

<gnu_srs1>civodul: janneke, et al congrats :)

<almuhs>hi

<almuhs>I'm thinking about how refactor my SMP code in a better design. Talking with Richard, some months ago, he told me that my SMP implementation is very dirty and fragile, and not follows the Mach design philosophy

<almuhs>this is my code https://github.com/AlmuHS/GNUMach_SMP

<almuhs>I'm thinking in the idea about start the work after the Hurd starting, not before. And take advantage of ACPI translator to find the MADT table with ACPI structures (Samuel told me this idea a year ago)

<almuhs>but I don't know how can I do this, and how to give the APIC pointer to the microkernel (and another related structures)

<almuhs>I'm asking about ideas. I'm searching opinions and any detail explaination

<rekado>civodul, janneke Woo! This is beautiful. Love the “Why bother?” section!

<user_oreloznog>+1

<almuhs>what is this?

<rekado> https://guix.gnu.org/blog/2020/a-hello-world-virtual-machine-running-the-hurd/

<rekado>so, civodul how did you record that gif?

<civodul>ah ah!

<civodul>i used byzanz

<rekado>(+ 2 54… and then what? Where’s the closing paren?

<rekado>I need closure!

<civodul>which actually records the whole screen

<rekado>ah

<civodul>yeah!

<civodul>and then i cropped the image in the GIMP

<civodul>but byzanz is kinda boring because you have to tell it upfront how long the recording lasts

<almuhs>hurd from scratch?

<rekado>oh, I see.

<civodul>almuhs: in a way!

<rekado>I was looking for a GIF recording program recently (I think prompted by your request for a Guile Studio blog post) and found this: https://github.com/phw/peek

<almuhs>jbgg: maybe you like this

<civodul>i used "guix search" :-)

<civodul>i know there's an Emacs GIF recording thingie that looks nice

<civodul>but it's not applicable here

<civodul>almuhs: BTW, i don't know the answers to the questions you asked above, but thumbs up for your work in this area!

<civodul>it's crucial

<almuhs>what is the meaning og "thumbs"?

<almuhs>*of

<rekado>civodul: what’s up with the chown error?

<rekado>“thumbs up”

<rekado>means approval and encouragement

<almuhs>oh, thanks

<almuhs>now I need inspiration to continue the work in a better way

<civodul>rekado: no idea, i did spend time on it but eventually gave up

<civodul>future work :-)

<rekado>almuhs: you are probably familiar with the Mach documentation, aren’t you?

<almuhs>It's my first operating system project. Never before I developed anything in Hurd

<almuhs>rekado: a few, but there are very little information about design or implementation

<rekado>civodul: with a Guixified Hurd VM image I’m very tempted to hack on it again

<rekado>almuhs: the wiki has copies of “Mach 3 Kernel Principles”, “Mach 3 Server Writer’s Interfaces”, and other documents in Postscript format.

<almuhs>ok

<rekado>they are full of details about how Mach works and how data are passed to Mach

<almuhs>I need a document which explain the function name and the flow of the calls

<rekado>I must admit that I don’t understand anything about ACPI or SMP, so I don’t know what exactly your project requires.

<rekado>“function name” of what?

<almuhs>by example "what function does x?"

<almuhs>or the inverse "what is the purpose of y function?"

<rekado>these documents contain both explanatory prose and procedure references.

<almuhs>ok, I'll try to read It

<almuhs>but I want to find any updated document about gnumach, not only Mach 3?

<almuhs>**not only Mach 3

<rekado>there is also a document called “The GNU Mach Reference Manual”, which explains the differences (and more)

<almuhs>is these updated?

<rekado>I found the Mach 3 documents to be easier to read, honestly.

<almuhs>I'll check both, them

<almuhs>*then

<rekado>I don’t know if it’s up-to-date. But Mach hasn’t really changed all *that* much over the years.

<rekado>The core ideas are still the same.

<almuhs>in the SMP code I take original Mach code as base, but gnumach has many own code in this area

<rekado>I see.

<rekado>The documents are all available here: https://www.gnu.org/software/hurd/microkernel/mach/documentation.html

<almuhs>yes, I know this

<almuhs>I also need info about how to create a Hurd server, and how to share data with the microkernel

<rekado>that’s covered in “Mach 3 Server Writer’s Guide” and “Mach 3 Server Writer’s Interfaces”

<almuhs>ok

<almuhs>ty

<civodul>rekado: sure, you should resume Guix+Hurd hacking :-)

<rekado>but I *just* got started with Guile Emacs…! :)

<civodul>ah well, one thing at a time

<janneke>and maybe one while compiling :-)

<janneke>rekado: oh, i wondered about a commit you made guile-for-guile-emacs

<janneke>nice

IRC channel logs

2020-04-08.log