<Begasus>
bah ... parse error on 32bit (gcc2) for libxdmcp :/
novaphoenix has quit [Quit: gone fishing]
novaphoenix has joined #haiku
Begasus_32 has quit [Ping timeout: 480 seconds]
vdamewood has quit [Quit: Life beckons.]
x10z has joined #haiku
x10z has quit [Ping timeout: 480 seconds]
BiPolar has joined #haiku
<Begasus>
Hi OscarL :)
<BiPolar>
Hello Begasus :-)
HaikuUser has joined #haiku
HaikuUser has quit []
<BiPolar>
Today on bare metal, to test some things on the Python 3.12 recipe / test-suite. Cleaning the work tree is just TOO slow on the VM (still slow on bare metal, but less so :-D).
<Begasus>
installed 32bit bare today on the other laptop (had 64bit there before), should speed some things up too compared to running from external USB drive
<BiPolar>
hope it goes well :-)
<Begasus>
still has issues with heating, so not hoping for the best (but lucky at best) :)
<BiPolar>
yesterday I made recipes for tzdata (python package) and also for "tzdb", then... after comparing with Gentoo's naming, I notice we already had "timezone_data" (which is exactly what I was doing with "tzdb" :-P)
<BiPolar>
So I just updated timezone_data now (and split it in pure data, _devel, and _tools.
<BiPolar>
Now about to see if "--with-tzpath=$dataDir/zoneinfo" makes Python finally find that data withouth needing to change path at runtime.
<BiPolar>
(also want to see if it helps with some errors on the test-suite)
<Begasus>
nice! (knew about the naming and Gentoo) :)
<BiPolar>
Web+ back/forward buttons do not seem to work/show properly for me. (stop button never goes "dim" either). Known issue?
<BiPolar>
("back" from context menu works)
<BiPolar>
forward too.
<Begasus>
if you switch tab it should show up
<BiPolar>
Yup!
<BiPolar>
Can't find a ticket for this on Trac.
<Begasus>
I tend to just disable gcc2 for libxau and libxdmcp
<Begasus>
thought it was known, have seen it for a long time
<BiPolar>
this one at least it is clear about NOT having releases.
<Begasus>
hence the mess at repology :/
<Begasus>
don't mind taking a look at updating, but this is scary ;)
<BiPolar>
"grabbing python3.12-3.12.0-1-x86_64.hpkg", 28.4 minutes. Not thaaat bad (considering I'm using "make -j 1").
<Begasus>
In Terminal or with hp?
<BiPolar>
with hp. (I know I can just use "make" there, just made it ULTRA explicit on the recipe, because you get issues otherwise).
<BiPolar>
"import zoneinfo; zi = zoneinfo.ZoneInfo("America/Catamarca")" <<< now this works on Py 3.12 out of the box. Noice!
<BiPolar>
bbiab
nerdistmonk has quit [Quit: Leaving]
Diver has quit [Read error: No route to host]
Diver has joined #haiku
Diver is now known as Diver
nerdistmonk has joined #haiku
<zdykstra>
LuaJIT is such a packaging nightmare these days
<zdykstra>
I'm still baffled by the choices that the new maintainer is making
<kallisti5[m]>
lol.. I tried to compile it once. It was a trip
<kallisti5[m]>
Then again... it's Lua 😀
<BiPolar>
https://luajit.org/status.html <<< seems to suggest our recipe should do a "git clone" (otherwise we'll miss the "POSIX timestamp" of the newest commit, that's supposed to be part of your version number.
<BiPolar>
("git clone" vs getting a tarball from GH, as we do for some recipes with $srcGitVer, for example)
x10z has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
x10z has joined #haiku
x10z has quit []
<zdykstra>
exactly - it actually needs metadata in the .git directory ... which is absolutely insane.
<Begasus>
yeah, that was my thought too BiPolar
<BiPolar>
zdykstra: to be fare, some odd python packages do too, but you can usually just override it with an ENV var.
<Begasus>
BiPolar, python3.12 build failed here :)
<BiPolar>
s/fare/fair/
<BiPolar>
Begasus: works for me (TM)! :-P
<Begasus>
or you got changes not pushed to your branch :P
<BiPolar>
lemme check.
<Begasus>
runtime_loader: Cannot open file libpython3.12.so.1.0 (needed by /sources/Python-3.12.0rc2/python): No such file or directory
<BiPolar>
Got that one 2 times only, in about 20 builds attempts since I'm testing 3.12, Begasus.
<Begasus>
not reproducable?
<BiPolar>
hp -c python3.12, and trying again solves it, AFAICT.
<BiPolar>
(I've tested with both -G and without it, just in case too)
<Begasus>
could be tricky on buildmaster :)
<BiPolar>
Didn't manage to properly reproduce the error, yet, no.
<BiPolar>
As it works after "hp -c", shouldn't have too much problems in buildmasters, I think.
mmu_man has quit [Ping timeout: 480 seconds]
<BiPolar>
regarding python packages "needed" .git metadata, that's why I had to patch that ugly "constantly" one needed by twisted.
<BiPolar>
(the one that jmairboek suggested I should do on PATCH() instead of on INSTALL())
<Begasus>
yeah, saw that
<Begasus>
progressing ...
<Begasus>
biab
freddietilley has quit [Quit: WeeChat 4.0.2]
xet7 has quit [Remote host closed the connection]
<Begasus>
same thing (second run)
<BiPolar>
:-/
<BiPolar>
VM?
xet7 has joined #haiku
mmu_man has joined #haiku
<Begasus>
BM
<BiPolar>
Weird. One more reason to keep doing repeated builds here, then.
<BiPolar>
(to see if I can find why it fails)
<Begasus>
third run ...
<BiPolar>
after "hp -c" or just repeated runs?
<Begasus>
always after hp -c
<yann64[m]>
Haikuporter is not generating patch files for source sub-directories :(
<yann64[m]>
haikuporter -e "portname" should do it (even for additional sources) right?
<Begasus>
do you link/copy them to $sourceDir?
<BiPolar>
yann64[m]: did you "git commit" your changes inside the work dir before using "haikuporter -e" ?
<Begasus>
if you commit the changes it should
walkingdisaster has joined #haiku
<yann64[m]>
will try again, I probably have missed something obvious
walkingdisaster has quit []
walkingdisaster has joined #haiku
walkingdisaster has quit []
<Begasus>
running the tests now BiPolar, didn't do that with the other attempts
walkingdisaster has joined #haiku
<BiPolar>
Begasus: attempting a build again (after uninstalling my test py3.12, `query -a libpython3.12.so.1.0` return nothing), this one with -G. Lets see how it goes.
<yann64[m]>
indeed I got mixed up with commit, thx
<Begasus>
np, I gather the SOUCE_FILENAME worked earlier yann64[m] ;)
<yann64[m]>
Indeed :D
<BiPolar>
Begasus: Nice. I can't see much (uncommited) changes, besides adding "--with-tzpath=" (just today). Wonder what the issue is.
<Begasus>
yeah, and it worked now on the third run, not something you want to see happening at buildmasters
<BiPolar>
bad timing bewteen linking the libs and running the "bootstrap" python, I guess, but no idea how to debug it.
<Begasus>
no $jobArgs, so that shouldn't interfere with it
<BiPolar>
exactly.
<BiPolar>
Running the test part of the build here, so I guess it will work OK.
<Begasus>
yeah, didn't get that far on the earlier attempts
<BiPolar>
(as we're using profile-guided optimizations, it runs parts of the test suite to gather the needed info, before rebuilding all using that)
<BiPolar>
at least we can cut down the search for the bug up to that part of the makefile then :-)
bjorkint0sh has joined #haiku
<Begasus>
gentoo's ebuild is a bit complex too :)
<BiPolar>
Still running the build... `find . -name libpython3.12.so.1.0` returns nothing on the work dir :-/ (guess I should try finding that file before the test part ends).
<BiPolar>
now it appears (right in the root of the work dir)
bjorkintosh has quit [Ping timeout: 480 seconds]
<BiPolar>
Wonder if it is just a search path issue.
<Begasus>
saw the fedora recipe? me happy ours is not that complicated! :)
<BiPolar>
I think I read it (and the gentoo one) when trying to filter out tests that hang. Got a trick from the gentoo one, and bailed out of them :-D
<Begasus>
heh
<BiPolar>
Build OK: real 28m37,284s
<Begasus>
I think on the other laptop it would take several hours to builld here ;)
<BiPolar>
Takes about 45/50 on VirtualBox for me (I/O seems to be the bottleneck there).
<BiPolar>
k, "--with-tzpath=" works nicely. Trying to enable the test_zoneinfo didn't worked. Sigh...me goes to read again how to enable test resources for the nth time.
<Begasus>
good luck hunting ;)
xet7 has quit [Remote host closed the connection]
<BiPolar>
Web+ scrolling stopped for a solid 2 minutes (some high activity in some WebSocket threads). Notice that It has *a whole lot* of CryptoQueue (inactive?) ones.
<Begasus>
why am I getting "fatal: not in a git directory" when checking up on the recipe for git?
<Begasus>
yeah, mentioned that one here also BiPolar
<Begasus>
annoying :/
<BiPolar>
Other than that.. feels quire snappier than older versions :-)
<BiPolar>
*quite
walkingdisaster has quit [Quit: Vision[]: i've been blurred!]
walkingdisaster has joined #haiku
<Begasus>
lol ... DESCRIPTION="Stupid content tracker: distributed VCS designed for speed and efficiency"
ClaudioM has joined #haiku
tuaris has quit [Read error: Connection reset by peer]
mmu_man has joined #haiku
aa72 has joined #haiku
aa72 has left #haiku [#haiku]
tuaris has joined #haiku
<erysdren>
Begasus, i have seen you come online every day for the last 4 months and do nothing except work on Haiku ports
<erysdren>
what do you do for a living? :P
<zdykstra>
he's a professional hand model
<BiPolar>
:-)
<Begasus>
on sickleave erysdren :)
<Begasus>
working up to retirement :P
<erysdren>
nice :D
<Begasus>
situation not that good (it isn't as bad as it was), otherwise I would still be at work during the day ;)
aa72[m] has joined #haiku
<Begasus>
k, don't know what's up with this git recipe, keep hitting "fatal: not in a git directory" ...
<BiPolar>
try with -G? /me ducks.
<Begasus>
buildmaster doesn't work that way :P
<Begasus>
but could try just to check out ...
<Begasus>
works as expected ...
erysdren has quit [Quit: Konversation terminated!]
<Begasus>
grabbing git-2.42.0-1-x86_64.hpkg and moving it to /Opslag/haikuports/packages/git-2.42.0-1-x86_64.hpkg
aa72[m] has left #haiku [#haiku]
<Begasus>
previous update there was last year, so shouldn't interfere with this (checked with previous tar version also)
<hightower4>
Do I see it right that when I run Haiku under Xen it doesn't recognize the disk?
<hightower4>
(the virtual hard disk I mean, for installation)
<hightower4>
is that possible or I'm doing something wrong?
<hightower4>
<hightower4>
other than that question, I did play with what was visible in the haiku that runs from the install image, and was quite pleased, nice work
<zdykstra>
it's pretty darn nice, isn't it?
<hightower4>
yes, pretty charming/captivating, while being unassuming :)
x10z has joined #haiku
x10z has quit []
x10z has joined #haiku
mfny has joined #haiku
<mfny>
Hello,
<mfny>
last time i tried Haiku the ability to change refresh rate was not working, has this been fixed as i belive it was a known bug at the time
<mfny>
if it has been fixed would that be in nightlys or via a updated post install normal image
<augiedoggie>
what graphics chip? it's unlikely to have changed if you're using AMD or Nvidia
<mfny>
Intel
<augiedoggie>
probably best to try a nightly but i don't recall very many changes to the graphics drivers since beta 4
<mfny>
i see
<mfny>
i have a high refresh rate display, up to 144hz and i use that all the time, 60hz just does not cut it any more
<mfny>
so not being able to switch is a bit of a problem for me and i had the issue on nvidia and on intel as i recall
<mfny>
but this was around release of beta 4
<mfny>
so was hoping something had changed in the mean time
<augiedoggie>
yeah, i don't pay attention to those types of changes so it's possible that it's been fixed
<zdykstra>
it depends on which driver you're using - the intel_extreme driver supports changing the refresh rate, but the VESA driver will just use whatever your initial efifb is set to
<zdykstra>
but take that with a grain of salt, 144hz might not be possible on Haiku due to <some limitation>
<mfny>
as i recall i did see at least 120hz in the menu but when clicking apply it was not being applyed
<mfny>
that was the bug, refresh rate not being applyed regardless of what it was set to
<mfny>
will try a nightly image tho and see what happens
walkingdisaster has quit [Quit: Vision[]: i've been blurred!]
jmairboeck has joined #haiku
<mfny>
no luck, still same issue
<mfny>
with nightly
<waddlesplash>
i don't think anyone has changed anything here recently
<waddlesplash>
so yes refresh rate hasn't been worked on
<jmairboeck>
win8linux[m]: you asked yesterday about the HVIF format on other platforms. I wrote an HVIF renderer for WPF, just for fun (and to see if VB.NET is good for anything): https://github.com/jmairboeck/HVIFViewer - so Windows is covered as well. Beware that it is not very polished and not feature complete though.
<BiPolar>
timezone handling in our Python's time module is kaput :-/ (we also seem to not define the TZ envar as apparently other *nixes do?)
<augiedoggie>
i don't have TZ in my env on linux
<BiPolar>
good to know, augiedoggie, thanks! Maybe it is just my poor reading comprehension again (while reading Python's time docs).
<BiPolar>
yeah.. that might be it... "On many Unix systems (including *BSD, Linux, Solaris, and Darwin), it is more convenient ..." and then talks about setting that TZ var.
mr_lou has quit [Quit: Leaving]
<Begasus>
closing down here, cu peeps!
Begasus has quit [Quit: Vision[]: i've been blurred!]
gouchi has joined #haiku
ClaudioM has quit [Quit: leaving]
tqh has quit [Quit: Leaving]
Maturi0n_ has quit [Remote host closed the connection]
Maturi0n has joined #haiku
BiPolar has quit [Quit: Vision[]: i've been blurred!]
Maturi0n_ has joined #haiku
Maturi0n has quit [Ping timeout: 480 seconds]
<kallisti5[m]>
I've come to the conclusion that Thunderbolt 3 + external GPU's suck
<kallisti5[m]>
pretty much, if CPU is high (80%+), GPU rendering lags hard
<kallisti5[m]>
guessing the CPU (or something) has a hand in making calls to the GPU over Thunderbolt resulting in lovely GpuWatchdog segfaults.
<bitigchi[m]>
kallisti5: how does your framework fare with Haiku lately
<bitigchi[m]>
Thinking of getting one
jmairboeck has quit [Quit: Konversation terminated!]
Anarchos has joined #haiku
Anarchos has quit [Quit: Vision[]: i've been blurred!]
nerdistmonk has quit [Quit: Leaving]
x10z has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
x10z has joined #haiku
<hightower4>
Zach: btw, the haiku look somehow reminded me of Qnx's GUI as well