<rburton>
preferred provider has no meaning in an image recipe
<linex[m]>
I also tried in local conf
Estrella has joined #yocto
<linex[m]>
but the distro seems to take precedence over that.
<linex[m]>
I just set it temporarily in the image recipe , it does not belong there indeed
<rburton>
getvar will show why the distro won over the local
<linex[m]>
yep :)
<linex[m]>
I found that machine/conf will win over that !
rich1234 has joined #yocto
amitk_ has quit [Ping timeout: 268 seconds]
<RP>
rburton: I'm running tests with the libinput upgrade reverted
<RP>
rburton: if I get these three qemuppc builds passing with that, I'm going to blame libinput
tooqnfxca has quit [Ping timeout: 252 seconds]
nemik has quit [Ping timeout: 240 seconds]
olani- has quit [Ping timeout: 260 seconds]
nemik has joined #yocto
<mckoan>
linex[m]: you can't set PREFERRED_PROVIDER in an image file
nemik has quit [Ping timeout: 276 seconds]
nemik has joined #yocto
<kanavin>
paulg, you probably saw do_compile of rust-llvm-native, not rust-native
<kanavin>
rust doesn't do separate build/install steps
<paulg>
kanavin, I'm thinking you could be right. I'll watch more closely next time.
<rburton>
paulg: try buildstats-summary :)
<paulg>
rburton, the slow arse POS still hasn't finished yet. :)
GillesMM has quit [Read error: Connection reset by peer]
<rburton>
it will show you data for in-progress builds as the files are written incrementally
GillesMM has joined #yocto
<paulg>
...or it will remind me of one of my pet peeves of fragmented python crap.
<paulg>
File "/home/paul/poky/scripts/buildstats-summary", line 9, in <module>
<paulg>
import dataclasses
<paulg>
ModuleNotFoundError: No module named 'dataclasses'
PobodysNerfect has joined #yocto
<rburton>
do you have a very old python?
<paulg>
18.04 ubu turd using the python from buildtools
<paulg>
Speaking of... Can we put rust into buildtools? :-P
florian_kc has quit [Ping timeout: 240 seconds]
<paulg>
oh well. Doesn't matter. I'm not building anything I need. Just to see what is broken after a pull, and since the home office is a bit cool on this damp rainy day.
<rburton>
hm i wonder why that didn't include dataclasses
<rburton>
buildtools includes the full bundle, in theory
olani has joined #yocto
olani has quit [Remote host closed the connection]
<d-s-e>
dataclasses where introduced in python 3.7, maybe ubuntu 18.04 has an older version?
<kanavin>
paulg, an option to shortcut rust-native madness would be welcome, so... patches welcome :)
* paulg
volunteers vmeson
Schlumpf has joined #yocto
<kanavin>
ubuntu 18.04 goes EOL at the end of this month
<paulg>
d-s-e, python3 from the buildtools/sysroots/x86_64-pokysdk-linux/usr/bin/python3 claims it is 3.10
olani has joined #yocto
<paulg>
kanavin, actually I think it went EOL two days ago.
olani has quit [Remote host closed the connection]
<kanavin>
paulg, I think it's been extended by one month :) but give or take that, upgrade in any case please
<JaMa>
I have some 18.04 builders with buildtools for newer python and haven't seen this issue yet
kscherer has joined #yocto
<d-s-e>
ok, seems it got a python update somewhere in the past. but 18.04. is really outdated anyway.
olani has joined #yocto
<paulg>
I know, but old machines with "low" memory suffer every time you upgrade.
<tgamblin>
paulg: sounds like it's time to roll your own distro with Yocto :)
<paulg>
tgamblin, ha - been there, done that.
<RP>
paulg: I can't help thinking himem.sys would help with that low memory
sakoman has joined #yocto
<paulg>
RP, I loaded EMM386.EXE already, isn't that good enough?
<RP>
paulg: clearly it isn't helping with rust! Trying changing the ordering
* RP
is suffering flashbacks :/
Thorn has quit [Ping timeout: 268 seconds]
zpfvo has quit [Ping timeout: 268 seconds]
<paulg>
outside of downloading freedos to make a BIOS update USB turd, I'm trying to think when I would have last fought with DOS crap...
<RP>
paulg: definitely been a while for me, I just remember it all too well
<paulg>
Probably Win 3.1/DOS-6.2 computers of relatives from 20+ years ago?
<paulg>
The sad part is, that if I somehow was sat in front of one of those ancient crappers, I'm sure a lot of that useless information would come flooding back.
Thorn has joined #yocto
<paulg>
did windows95 still have DOS and highmem.sys lurking under the surface?
<paulg>
I just seem to remember Win-3.1 being the last release where DOS and windows were truly separate products.
<paulg>
anyway, enough off-topic rambling. I blame RP for the distraction.
<RP>
paulg: it is scary to think about those times. Windows 3.11 for workgroups
camus has joined #yocto
<RP>
rburton: those test builds are doing fine, it is looking like libinput
<rburton>
i wonder what libinput did
<JaMa>
everything you need on just 8 floppy disks :)
<RP>
JaMa: wasn't that slackware? :)
camus has quit [Client Quit]
<paulg>
RP, awesome, wasn't it? You had to pay extra to get a networking stack.
<RP>
paulg: it was great at the time. You didn't always have a network anyway
* RP
is still distracting paulg
<paulg>
instead you had a parallel port dot matrix printer!
<RP>
paulg: er, have. Or perhaps I'm not admitting to that
<paulg>
I wonder if anyone still makes tractor feed paper?
<rburton>
i saw some recently!
* RP
did buy a USB parallel port adaptor so I could work with some old stuff (microcontroller programmer)
<paulg>
probably costs a fortune if they do. Selling to 1980s nuclear reactor sites and the like.
rfuentess has quit [Ping timeout: 240 seconds]
<kanavin>
the reason I chose red hat over debian in 1998 is that I was a cash-starved student back then, and a computer shop offered me red hat on 1 CD or Debian on 6 CDs, which would be 6x more pricey
<kanavin>
so debian's tendency to package everything ever backfired in this case :D
rfuentess has joined #yocto
<JaMa>
I've started on Corel Linux just because I got it for free
<JaMa>
but with 860MB disk and no access to internet it didn't last long
<kanavin>
oh, their attempt to make a windows competitor
<kanavin>
I barely remember, it lasted a couple of years maybe?
<rfs613>
it eventually morphed into Xandros which is apparently still a thing today
<rfs613>
actually no, it looks like it is no longer active
PobodysNerfect has quit [Ping timeout: 276 seconds]
<rburton>
DvorkinDmitry: broken upstream configure, you'll need to figure out how to stop it looking in /usr
<JaMa>
DvorkinDmitry: even dunfell uses newer nginx, so try it with some supported release and backport whatever is needed if you have to stay on old unsupported release
tunahan has quit [Quit: tunahan]
tunahan has joined #yocto
tunahan has quit [Client Quit]
<DvorkinDmitry>
Jama, I see ./meta-openembedded/meta-webserver/recipes-httpd/nginx/nginx_1.17.8.bb ./meta-openembedded/meta-webserver/recipes-httpd/nginx/nginx_1.16.1.bb in my dunfell. I'm building 1.16...
<JaMa>
does it work with 1.17.8?
<DvorkinDmitry>
rburton, I replaced /opt/local/... case it was not able to find libxslt/parser.h with default settings in Docker. On my host system it is OK
bps has quit [Ping timeout: 240 seconds]
<DvorkinDmitry>
JaMa, nginx 1.17 - result is the same
seninha has quit [Quit: Leaving]
Schlumpf has quit [Quit: Client closed]
mckoan is now known as mckoan|away
starblue1 has quit [Quit: WeeChat 3.0]
starblue has joined #yocto
Haxxa has quit [Ping timeout: 240 seconds]
GNUmoon2 has quit [Ping timeout: 240 seconds]
amitk_ has joined #yocto
GNUmoon2 has joined #yocto
d-s-e has joined #yocto
d-s-e has quit [Client Quit]
Haxxa has joined #yocto
rfuentess has quit [Remote host closed the connection]
<RP>
paulg: the issue is the From address gets mangled by the list due to the DMARC policy of the WR servers
<RP>
or was it the other thing to DMARC, anyway, that kind of thing
<rburton>
$ git config --get sendemail.from
<rburton>
Ross Burton <ross.burton@arm.com>
<paulg>
yeah, DMARC poo or whatever the abbreviation -- was turned on like 10d ago.
<RP>
rburton: when it writes out the patch, does it add a From: at the top?
<RP>
if not, the FAQ is wrong :/
<rburton>
From 0ef68f5c068ecec8506849329d595f73cbb09a6d Mon Sep 17 00:00:00 2001
<rburton>
From: Ross Burton <ross.burton@arm.com>
<RP>
rburton: in the body of the message rather than as a header?
<rburton>
that was the first two lines of the git-format-patch output
<RP>
The idea is to inject From: at the start of the body so when git am pulls it, it gets the address right and then we don't care about what the mailing list does
<rburton>
right
<rburton>
my mails on the list don't have a from in the body
<RP>
abelloni: happen to be around and able to remember how to do this?
frieder has quit [Remote host closed the connection]
<rburton>
git adds a From: to the _body_ if the from doesn't match the sender
<rburton>
can't see a way to force it to do that always
<RP>
rburton: someone did have a trick for this :/
<paulg>
never ceases to amaze me how we've managed to complicate 1980s technology e-mail into a crippling hideous frustrating morass...
<RP>
paulg: it is a shame we can't just telnet to port 25 and type the mail...
<paulg>
Make E-mail Great Again. :-P
* paulg
is guilty of telnet ... 25 and typing HELO
* RP
used to test email servers by sending complete emails via telnet
<RP>
paulg: would be fun to crosswire that with chatGPT. HELO might get a "Good afternoon" reply
<abelloni>
rburton: just put ross.burton@arm.com in sendemail.from
<rburton>
ok i'll try that
<abelloni>
it will figure out the name
<abelloni>
the display-name
<paulg>
wonder if chatGPT can parse sendmail.cf files. I'm sure I lost days of my life trying to do so around 1993.
<rburton>
aha
<RP>
paulg: I still run my system through sendmail
<rburton>
so i think git is sufficiently dumb that that works
<rburton>
the list says via still, but the body now has a From: header
<rburton>
obviously git is just doing a strcmp and not parsing the address
<RP>
rburton: can you please update the wikis to make this clear please?
florian has quit [Quit: Ex-Chat]
<rburton>
done
<khem>
nerdboy: We have patches under test for gcc 13
<RP>
rburton: thanks
<RP>
abelloni: thanks, mystery solved
<RP>
rburton: btw, your emails were displaying correctly for me. There is something odd where the list corrupts your own emails always
<RP>
rburton: so I see my own as corrupted but others are fine, others see mine fine
* paulg
follows the discussion an dutifully puts ross.burton@arm.com in his sendemail.from field.
<rburton>
RP: huh
<RP>
paulg: it would be funny if you did sent that
<RP>
rburton: michael did try and explain it to me but in the end we agreed I'd just live with it
<rburton>
RP: ah so maybe i should have ignored the list mail i was seeing
<RP>
rburton: that is what I do, yes
<RP>
rburton: I'd not seen any issue with your patches
sudip_ has joined #yocto
eLmankku has quit [*.net *.split]
bryanb has quit [*.net *.split]
ramacassis[m] has quit [*.net *.split]
nohit has quit [*.net *.split]
eLmankku_ has joined #yocto
bryanb_ has joined #yocto
bryanb_ is now known as bryanb
nohit_ has joined #yocto
ramacassis[m] has joined #yocto
whuang0389 has joined #yocto
<whuang0389>
Hi has anyone seen this before during the unpack step? "No up to date source found: clone directory not available or not up to date ... shallow clone not enabled"
PobodysNerfect_ has quit [Quit: Gone to sleep. ZZZzzz…]
otavio has quit [Ping timeout: 268 seconds]
<RP>
paulg: I hate to ask you this. Do you know if the G4 powerpc cpu we use in qemu should have the mffscrni instruction? Should gcc be generating that instruction for the 7400 tuning we're using?
<RP>
paulg: basically libinput is segfaulting due to mffscrni being present in sincosl in libm.so
<RP>
paulg: I'm wondering how to blame.
<RP>
khem: ^^^ if you've any ideas?
PobodysNerfect has joined #yocto
<khem>
that seems like a ppc10 ISA instruction
<khem>
is it coming from some assembly in libm ?
<RP>
khem: good question
<RP>
khem: if it is I can't spot it easily
<khem>
no I am wrong. its ISA 3.0 so should work
<khem>
wait to its ok on qemu but not on real h/w
<khem>
I see some patches to qemu from Jun 2022 which implemented it
<khem>
it certainly is possible that for 7400 machine we are emulating its not enabling ISA 3.0 for sure
florian has joined #yocto
alessioigor has quit [Quit: alessioigor]
alessioigor has joined #yocto
otavio has joined #yocto
Haxxa has quit [Quit: Haxxa flies away.]
Haxxa has joined #yocto
alessioigor has quit [Quit: alessioigor]
mvlad has quit [Remote host closed the connection]
PobodysNerfect has quit [Quit: Gone to sleep. ZZZzzz…]
* paulg
has nothing useful to add, other than we shouldn't still be building/using 7400.
kscherer has quit [Quit: Konversation terminated!]
<RP>
paulg: what should we be using?
<RP>
As far as I can tell. glibc is emitting a mffscrni instruction from C code for s_sincos.c regardless of which -mcpu I pass it
<RP>
even for power4 and other old stuff
<paulg>
maybe an abacus or a slide rule?
* RP
has one of the latter, perhaps if I find that I'll understand why this is breaking
<RP>
hmm, the assembler is intersting. .machine push; .machine "power9"; mffscrni 0,0; .machine pop
Nostromo8 has joined #yocto
<Nostromo8>
I've been using the Yocto Kirkstone branch and I noticed the default GCC Version is set to 11.3.0. Is the recommended Ubuntu version for Kirkstone 22.0.4 or can it support older versions?
seninha has joined #yocto
Thorn has quit [Quit: Copywight 2016 Elmer Fudd. All wights wesewved.]