<nekobot>
[haiku/haiku] 16c6cc4418ea - BTextView: Fixup AdoptSystemColors() and HasSystemColors()
nephele_xmpp has left #haiku [Error from remote client]
Nasina has quit [Read error: Connection reset by peer]
nephele_xmpp has joined #haiku
<nephele_xmpp>
Skipp_OSX: what is the context for your question?
<Skipp_OSX>
8878
<nephele_xmpp>
It’s not related to the button color directly, this gradient is also used by many other controls that now use the panel color
<nephele_xmpp>
These controls have to use the control color also
<nephele_xmpp>
Ugh, I guess I really can’t make small changes without them beeing -2 because they aren’t perfect alone. This really kills my motiivation.
<Skipp_OSX>
ok but the color of all those are going to be from the now untinted B_PANEL_BACKGROUND_COLOR
<Skipp_OSX>
welcome to my world!
<nephele_xmpp>
no, they are partially panel colors and partially control colors
<Skipp_OSX>
assuming you're using the base color for all these gradients then my change should behave identically to before.
<Skipp_OSX>
right, which is why them matching is going to make it all work.
<Skipp_OSX>
they match either way, but this way without having to tint explicitly to get that result.
<Skipp_OSX>
well, unless you change your colors of course, but you have to deal with that anyway.
<nephele_xmpp>
a) they aren’t matching now, it’s not possible to match arbitrary color A * tint = arbitrary color b
<nephele_xmpp>
b) The change is not needed
<nephele_xmpp>
and c) with the changed gradients with the panel color these controls look bad
<nephele_xmpp>
so the next change would be moving more controls to control color and adjusting the control color
<Skipp_OSX>
but of course part of the work is the make them match, meaning to replace uses of B_PANEL_BACKGROUND_COLOR with B_CONTROL_BACKGROUND_COLOR.
<Skipp_OSX>
If we make the change I suggested that will all be made much easier, if we do it now before you start replacing B_PANEL_BACKGROUND_COLOR that is.
<nephele_xmpp>
No, I don’t think it is. We want to match the output, roughly that is “how does the button look now"
<nephele_xmpp>
but not the input, which is the color constants
<Skipp_OSX>
right, but we also want buttons and other controls to match panel.
<nephele_xmpp>
from a visual standpoint, because of the gradient liberal use of the passed in color, it really is much brighter than the panel. So the control color as a result, which we want should be brighter, and the controllook less liberal with it’s interpretation of the color
<Skipp_OSX>
but that's already how it works, and will continue to work.
<nephele_xmpp>
huh? no. the controls are brighter, despite the panel color matching. not because of it :)
<Skipp_OSX>
at least in the case of BButton...
<Skipp_OSX>
oh but that's YOUR problem (as the HaikuControlLook tinter)
<Skipp_OSX>
I'm trying to make the input you receive correct, you are responsible for giving me the appropriate output from it.
<nephele_xmpp>
Well, see, i made the gradient match the color. And now it doesn’t look acceptable at all. And I think you will find the same thing with any other controlLook that matches the colors properly
<Skipp_OSX>
I mean, you wrote #19357 look at what you wrote...
<Skipp_OSX>
Yes but that's irrelevant, assuming we're talking about buttons here.
<nephele_xmpp>
yes, I did not mean we should replace it with the panel color. Because i am sure that it is not the right color
<Skipp_OSX>
Really for anything, even if it's not a button it's using panel.
<Skipp_OSX>
And yet here I am and am 100% sure that it is the right color
x10z has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
<nephele_xmpp>
so we have two places modifying colors right now: BButton tint and HaikuControlLook gradient. We have to remove *both* and then set the Control Color to the color that matchess roughly the previouss rendering
<Skipp_OSX>
except for BeControlLook, that will need a B_LIGHTEN_1_TINT applied to go from 216 => 245
Hannah has quit []
<Skipp_OSX>
but you realize that the BButton tint that I removed yields exactly the same color as B_PANEL_BACKGROUND_COLOR does right?
<nephele_xmpp>
No, it doesn’t, as i explained before
<Skipp_OSX>
once you apply the tint it does
<nephele_xmpp>
Again, no.
<Skipp_OSX>
how could it not?
<nephele_xmpp>
It does this only when you have *exactly* the default colors. otherwise the calculation is wrong
<Skipp_OSX>
I mean I tested the change, the color is exactly the same after as before.
<Skipp_OSX>
no because it's always tinted the same, even switching tints for dark mode currently.
<Skipp_OSX>
it should be 100% identical to before only without the tint.
<Skipp_OSX>
The difference is that once you start expanding to other controls besides BButton either you'd need to apply the same tint the BButton does, or live with the color difference, or fix it up in HaikuControlLook.
<nephele_xmpp>
I think you misunderstand. If you adjust the control color in any way, as a user, which is perfectly valid, the calculation with the Tint messes up and you don’t get the equivalent color for the panel anymore
mmu_man has quit [Ping timeout: 480 seconds]
<nephele_xmpp>
Apart from saying that you think your change has no consequence, what is the *reason* for it? the controls clearly don’t have the same visual color as the panels.
<nephele_xmpp>
> The difference is that once you start expanding to other controls besides BButton either you'd need to apply the same tint the BButton does, or live with the color difference, or fix it up in HaikuControlLook.
<nephele_xmpp>
let me fix HaikuControlLook (my patch 1), remove the button tint (your patch), adjust other controls (scrollbar -> nipos patch), (and more patches from me). We can then have the color be set properly
<Skipp_OSX>
The reason for it is so that we continue to get the same effective behavior as before but be freed to use B_PANEL_BACKGROUND_COLOR outside of BButton without causing colors not to match with panel colors where not intended. Basically B_CONTROL_BACKGROUND_COLOR is a lighter version of B_PANEL_BACKGROUND_COLOR and that has caused us to avoid using it up unto this point out of fear of colors not matching.
<Skipp_OSX>
I'm sorry, be freed to use B_CONTROL_BACKGROUND_COLOR outside of BButton...
<nephele_xmpp>
I’m confused. If you want panel colors just use panel colors and delete control
<Skipp_OSX>
no, do not fix in HaikuControlLook...
<Skipp_OSX>
no, because you can't CHANGE the Control Color separately from the panel color in Appearance.
<nephele_xmpp>
Sorry, but panel colors == control colors is obviously wrong. Buttons do not have the same color ass panelss
<Skipp_OSX>
no it's not, it's only == by default.
<Skipp_OSX>
It's not obviously wrong.... arg
<nephele_xmpp>
Yes it is. They don’t look the same, and controls looking differentt than panels is not a rendering artifact of the control look
<Skipp_OSX>
What you're looking at is a tinted version of B_CONTROL_BACKGROUND_COLOR that makes it identical to B_PANEL_BACKGROUND_COLOR though.
<Skipp_OSX>
so the differences you're seeing are NOT caused by the value being different and wouldn't be altered either.
Hannah has joined #haiku
<nephele_xmpp>
still no. You would be right if HaikuControlLook would not modify the color.
<Skipp_OSX>
where does that happen?
<nephele_xmpp>
and I am strongly against the controlLook shifting the color upwards this much, it makes it much harder to properly set the color as a user
<Skipp_OSX>
buttonBgColor = tint_color(base, B_LIGHTEN_1_TINT); changed to buttonBgColor = base; is undoing the tint huh?
<Skipp_OSX>
So if I make it 245 again, or undo the change I mean, can you make your patch work with that value?
<Skipp_OSX>
undo the 245 => 216 change, but continue to remove the tint from BButton I mean
<nephele_xmpp>
Yeah, if you remove the the 216 part, and only have the tint deleted, and 8878 appplied the button will look mostly similar to before. Likely have to adjust the control color somewhat
Halian has quit []
<Skipp_OSX>
ok that works for me, that's all I'm concerned about.
<nephele_xmpp>
but all other controls using the glossy gradient and the panel now look bad, so those need to be moved to control colors aftterwards
<nephele_xmpp>
to match their prior appearence
<Skipp_OSX>
right we'll have to wait to apply both patches at the same time
<nephele_xmpp>
anyhow, that is what i ment primarily, it’s a bit hard to figure out how controlLook shifting + tint influeces the color value. So we should remove both and then pick the then matching control color
<Skipp_OSX>
I still think it should match but whatever, I'll live with it.
<nephele_xmpp>
but once the gradient shifting is removed all panel color using controls look broken. If you want apply 8878 and see if you are happy with the rendering… i think you won’t be
<nephele_xmpp>
check the comboboxes in appearence for example
<nephele_xmpp>
or the BSpinner butttons
<Skipp_OSX>
ok I'll look
<nephele_xmpp>
hmm, i think we are atttacking the same problem just from different angles. but it is complex
<Skipp_OSX>
Seems like if the button being too light is your problem, this will exacerbate the problem, not lessen it.
<Skipp_OSX>
or the would-be B_CONTROL_BACKGROUND_COLOR elements I mean
hightower4 has joined #haiku
<Skipp_OSX>
245 is even lighter than 216, even less margin between it and 255... unless your trying to go the other way I suppose
<nephele_xmpp>
I think the button color beeing brighter than the panel, in light mod, is semantically correct
<nephele_xmpp>
the tint does make it more like panel, and for a flat button this matches, but the gradient undoes the tint again
<nephele_xmpp>
(not exactly, but it makes it substanttially brighter, which is my point)
OscarL has joined #haiku
hightower3 has quit [Ping timeout: 480 seconds]
<OscarL>
bah... if I use VMware, I have internet access to my Haiku VM... (via USB tethering), until it makes my cellphone auto-reboot after a bit over 1/2 an hour. When using VirtualBox... can't get internet access at all for some reason :-(
<Skipp_OSX>
ok then
* OscarL
waves to everyone on the room.
<nephele_xmpp>
hello OscarL
<OscarL>
Hi nephele_xmpp. Good evening Skipp_OSX.
<nephele_xmpp>
Skipp_OSX: I guess i will tommorow start a more comprehensive patch to change most controls to the control color, remove the tint, and fix the controllook all in one go
<nephele_xmpp>
for now, have a good night :) going to sleep
nephele_xmpp has left #haiku [Error from remote client]
Halian has joined #haiku
thesynthax has joined #haiku
mmu_man has joined #haiku
phcoder has joined #haiku
<phcoder>
Hello, all. I try to run Haiku on System76 laptop. It starts if I disable SMP but hangs whenever it waits for an interrupt eben if the wait has a timeout. My guess is that no interrupt happens, not even timer interrupts. I suspect a firmware bug. Fortunately firmware is open-source. Any ideas where I should look? Linux, Windows and *BSD don't have this problem
mmu_man has quit [Ping timeout: 480 seconds]
mmu_man has joined #haiku
HaikuUser has joined #haiku
HaikuUser has quit []
oevl has quit [Ping timeout: 480 seconds]
HaikuUser has joined #haiku
HaikuUser has quit []
<PulkoMandy>
phcoder: hello :) I don't have much more ideas about it, it can be either a problem with MSI or MSI-X, or maybe just with the timer used for timeouts? We had problems in early 2000s with some chipsets not implementing the old PIT correctly, but I would hope this is not the case anymore
<PulkoMandy>
and I don't even know what we use for these timeouts in modern machines
<phcoder>
I have disabled MSI but it made no difference. Also EC problem traced to timed wait on GPE interrupt. It's as if entire APIC was non-functional
<phcoder>
I don't even know if this board has PIT
<phcoder>
My guess would be HPET or LAPIC timer
nephele_xmpp has joined #haiku
<PulkoMandy>
some APIC things can be disabled from the boot menu as well, but I don't know if that would help. So at this point it will be adding debug logs to all timer related code I guess
illwieckz__ has joined #haiku
nephele_xmpp has left #haiku [Error from remote client]
<phcoder>
I tried stuff in boot menu but to no avail. I think it's the case of not enough APIC support rather than too much of it. I'll see if there are any interrupt counters I could check to prove my theory
<PulkoMandy>
yes, sometimes disabling the new and a bit unfinished code can help by getting a more legacy but well-tested path, but maybe not in this specific case. So we'll have to actually fix the problems :)
AD_Anime2 has joined #haiku
AD_Haiku_onPC_ has quit [Read error: Connection reset by peer]
nephele_xmpp has joined #haiku
<nekobot>
• korli (084e0d71): noto_serif_cjk: bump version
<nekobot>
• korli (54a99ae8): xerces_c: bump version
<Begasus[m]>
Korli is back at work :)
nephele_xmpp has left #haiku [Error from remote client]
JakeSays has joined #haiku
JakeSays1 has quit [Ping timeout: 480 seconds]
nephele_xmpp has joined #haiku
nephele_xmpp has left #haiku [#haiku]
nephele has joined #haiku
thesynthax2 has joined #haiku
thesynthax has quit [Ping timeout: 480 seconds]
OscarL has joined #haiku
freddietilley has quit [Read error: Connection reset by peer]
freddietilley has joined #haiku
<Begasus[m]>
Hi OscarL and nephele :)
<OscarL>
Good day Begasus[m]. Answered your question on libsigrokdecode's PR. Let me know if you want me to remove cmd:cmp/cmd:diff (before I disconnect to save data :-D)
<nephele>
hi Begasus[m], you better be at the coding sprint! :P
jpelczar[m] is now known as jpelczar_[m]
<Begasus[m]>
need to check the PR OscarL doing some other thing first :)
<nephele>
OscarL: is irc really that heavy? I'd imagine it is very converservative?
<Begasus[m]>
I'll get back to it on the PR :)
<Begasus[m]>
nephele: at home doing sprints :)
<OscarL>
nephele: the problem is wrangling Windows/Android and stupid apps from eating my data :-(
<nephele>
OscarL: ouch. I see :(
<OscarL>
Begasus[m]: no worries. If you have changes suggestions... I might do them later at night when I have some "free" MBs to spare :-D
<nephele>
OscarL: i know in android 4 one could prevent apps from using background data
<nephele>
but no clue how to do this for newer aside from turning on metered / low data mode
<nephele>
but that may not tune it down enough
<Begasus[m]>
will do OscarL
<OscarL>
Not sure about Android 4, but Android 6 has that setting, still notice things slipping through thou :-/
<OscarL>
got my Win10 pretty locked down, still there are surprices.
<nephele>
Application level firewall for everything. But windows 10 is really just malware
<OscarL>
Also... I haven't been on metered (and expensive) internet in a couple of years, so I need to re-adjust my usage patterns :-D
<nipos>
On Android 10 it's still possible to disable background data for individual apps
<nephele>
on my phone no apps may use background data. I only get notification pushes, and there only for the 1-2 applications i actually use for communication
<OscarL>
(no more Steam or YT. Need to ask my brother to download a full dump of Wikipedia in .zim format :-D)
<nipos>
Go to the app in the settings,then mobile data and wifi,there you'll find a switch for background data
<Begasus[m]>
checking libsigrokdecode
<nephele>
OscarL: for steam, use offline mode, find people in a different LAN that have the game and then do in-home transfer
<nephele>
or copy the app manifest plus the folder in common directly
<OscarL>
nipos: newest device I have is android 6.0.1, with a couple of old batteries soldered in parallel, and hooked up in the back of the phone case :-D
<nipos>
On my phone all communication apps (Conversations,Session) have background data (as well as disabled battery saving) so I get notifications while away
<Begasus[m]>
OscarL: tests pass without the cmd's cmp/diff
<OscarL>
Begasus[m]: yes, mentioned that on the PR.
<Begasus[m]>
ah :)
<Begasus[m]>
should check first I guess ;P
<nipos>
OscarL: Oh,sounds funny :D
<OscarL>
Begasus[m]: I was just trying to avoid the warning messages from ./configure, really.
<OscarL>
nipos: I hate planned obsolesence... also... couldn't afford new models even if I wanted, LOL :-D
<nipos>
Yeah,planned obsolescence sucks.And the fast moving phone market today also
<Begasus[m]>
OscarL: if you try to avoid those warnings you're in for a great sometimes :)
<OscarL>
(I was using a perfectly OK Android 4.1 till a few days ago)
<nipos>
I wish I could still use my good old Blackberrys,never liked Android.But I think they don't even support Let's Encrypt certificates,which affects basically every website
<OscarL>
nipos: hate when a perfectly fine (mandatory) home-banking app/phone combo stops working after an app update... and bank offers as solution: buy a newer phone :-(
<nipos>
I don't do online-banking
<nipos>
And generally I don't use proprietary apps at all
<nephele>
No banking website anymore?
<OscarL>
home-banking website, that was functional in the past... now requires a "token" (TOTP) that only their mandatory app will give you :-/
<nipos>
I need my phone only for calls/SMS and I do some surfing and XMPP with it as well,but could probably life without it.Sometimes thinking about going back to a simple stupid featurephone.
<Begasus[m]>
OK, no change with the cmd's included or not OscarL , just libcheck is needed/required
<Begasus[m]>
haven't checked if it could be moved to TEST_REQUIRES yet
<OscarL>
it can't.
<OscarL>
will drop cmp/diff then.
<OscarL>
there.
<Begasus[m]>
nope, you're right
<Begasus[m]>
+1 on dropping the cmd's :)
<nephele>
How should i check severeall boolean flags? I want to check in fFlags the FLAG_FLAT is set, and that FLAG_INSIDE is not set
<OscarL>
Begasus[m]: re: sigil... I *seem* to recall that you tried patching it in the past, right? (that collections vs collection.abc). Seems functional enough at least.
<Begasus[m]>
I'll do a check on that also to make sure, did a check not that long ago iirc again, but yeah, we tackled that some time ago :)
<Begasus[m]>
wasn't there an issue upstreamed about it?
<OscarL>
this uses 3dEyes fork, last updated 6 years ago.
<nekobot>
• OscarL (a5e7de04): libsigrokdecode: switch to Python 3.10. (#11707)…
<OscarL>
Begasus[m]: so I thought... better patch it now until 3dEyes decides to update its fork :-)
<Begasus[m]>
maybe also upstream with a PR?
<Begasus[m]>
building sigil now
<OscarL>
Mmm, he has issues disabled on his fork. I think that's why I haven reported it before :-)
<Begasus[m]>
ah :) makes it hard then :)
<OscarL>
(I've tested it opening a ~200 KB .epub file, and a 6 MB one.)
<Begasus[m]>
should silense the deprecated warnings :)
<OscarL>
Would also be nice to move it to Qt6 (if at all possible). But out of my area :-D
<Begasus[m]>
opening the epub file I have here works fine (last time without patching that one failed) +1
<Begasus[m]>
haven't checked with Qt6 (6.8.2 here by now) :)
<OscarL>
yeah, the error dialag was pretty annoying. Good thing the fix was pretty simple.
<Begasus[m]>
one downside, I have to check those Qt6 builds done here on another laptop that still contains 6.7.2 :P
<OscarL>
No hurries in trying to move Sigil to Qt6. was just thinking outloud :-)
<Begasus[m]>
need to run a test on the other laptop for GCompris latest release (moved to Qt6)
<OscarL>
Only reason for me to touch Sigil is because the other altearnative .epub "reader", is insane with its depedencies (calibre)
<OscarL>
not touching that again any time soon :-D
<Begasus[m]>
and update qml_box2d for that also
phcoder has quit [Quit: Connection closed for inactivity]
<Begasus[m]>
there is another one from KDE (name slips my mind atm)
<Begasus[m]>
arianna, but that uses qtwebengine (Qt5)
<Begasus[m]>
if Korli has no objections Sigil is fine with me
<OscarL>
thanks for checking.
<Begasus[m]>
np, thanks for fixing :)
Nasina has quit [Read error: Connection reset by peer]
<OscarL>
moving Sigil to 3.10 also allows to drop at least a couple of *_python39 packages.
<Begasus[m]>
OscarL: for cffi: /bin/sh: line 1: more: command not found
<OscarL>
(lxml, for example)
Nasina has joined #haiku
<OscarL>
mmm
<Begasus[m]>
(for the tests)
nephele_xmpp has joined #haiku
<OscarL>
ah... tests... forgot to mention I haven't run those (need to download/install py/pytest and can't do that till later this night.
<Begasus[m]>
lol, thought you smoke tested it? :P
<OscarL>
smoke tested yes.
<OscarL>
not run the actuall full test suite
<Begasus[m]>
/bin/sh: line 1: more: command not found
<Begasus[m]>
still not found?
<Begasus[m]>
heh
<OscarL>
let me see... pytest package might not be as big/heavy as I remembered... will try now
<OscarL>
(quite a few deps though, oh well)
<Begasus[m]>
not seeing any reference to cmd:more in the source
<OscarL>
Will just remove the call to "help(cffi) from the "smoke test", that should fix it.
<OscarL>
Still "Ran 0 tests in 0.002s" :-/
<OscarL>
No idea if this ever worked before (thus why I add reference results when I do run --test)
Nasina has quit [Read error: Connection reset by peer]
<Begasus[m]>
small script to call "more --version" works
<OscarL>
Begasus[m]: in this case, the call to help(cffi) was being "paginated" by the Python interpreter, think it got confused thiking it was running on an interactive shell.
<Begasus[m]>
maybe it's missing cmd:sh?
<OscarL>
no idea how Python calls more for the pagination in that case, and why it "fails", yet it does use for pagination still.
<OscarL>
No need, we can really just remove that smoke test from the recipe. I just left it there (slightly modified) from what older recipe version did.
<OscarL>
mmm, recipe was calling the test pretty wrongly.
<OscarL>
using "py.test c/ testing/" seems the way to go.
<nephele>
one thing i like about ssh keys over http auth, no username :D
<nephele>
just my password
<OscarL>
using passwordless ssh keys for local net ssh is also pretty convenient once the initial setup is done.
<OscarL>
Begasus[m]: "test session starts"... "collected 2043 items", this might take a while :-D
<Begasus[m]>
ok, now running tests :)
<nephele>
nipos: still there? I uploaded my patch to fix all dem controls. Except for the scrollbars. I figured I'd rather not introduce a merge conflict with that
<nephele>
but if you want you can now use control colors aswell for the scrollbars, we can then merge both patches
Nasina has joined #haiku
<OscarL>
Begasus[m]: will leave only that "py.test c/ testing/" call inside the for loop on TESTS().
<nipos>
Yes,I'm still here.Will try your new patch.
<OscarL>
Begasus[m]: will update recipe with reference results too. Has some fails, but given that no one complained before... I think this is OK for now.
<Begasus[m]>
+1
<OscarL>
still at 80% here :-D
<nipos>
I don't really want to change the scrollbar stuff yet again,at least as long as your patch isn't merged and there's still the risk that someone starts a big discussion about it.That prevents getting stuff done.Better merge scrollbars as-is and then add the panel -> control color change to your patch.
<Begasus[m]>
OscarL: you can compare then :)
Nasina has quit [Read error: Connection reset by peer]
<OscarL>
good thing I'm running this *now" (at 32 degrees Celcius). Expected 40/42 around noon :-(
<Begasus[m]>
nephele: see, we "are" running a code sprint here too :)
<OscarL>
(min temps at >26 even at night ain't fun either :-D)
<nephele>
Begasus[m]: i was planning to attend anyway. But now I am "online" instead
<nephele>
Oh well
<Begasus[m]>
still good nephele +1
<nipos>
Updating 7932 targets (for haiku.hpkg).I should really consider switching to the Intel laptop for that stuff,since I can compile with multiple cores there
<Begasus[m]>
added or "libroot" in name: in line 19 for testing/cffi0/test_parsing.py
<Begasus[m]>
" or "libroot" in name"
<Begasus[m]>
had a bit of a talk at FOSDEM in regards to libc/libroot ;)
<OscarL>
pushed my changes to the PR.
<OscarL>
if you think it makes sense to patch the test case, feel free to do so.
<Begasus[m]>
would be nice to figure out where it calls -ld :)
<Begasus[m]>
maybe this? "m = ffi.dlopen(lib_m)"
<Begasus[m]>
nope
diver has quit [Quit: Leaving.]
<nephele>
libm is already a empty library to satisy linkers :P
<OscarL>
as errors are pretty harmless, not sure I'd expend much time patching it (unless to scratch that itch).
<nephele>
nipos: hmm, maybe i can rebase ontop of yours later then.
<Begasus[m]>
you know me too well :P
<OscarL>
I should try (again) to solve the ones from the main Python packages... but ain't always easy :-D
<OscarL>
(darn str{f,p}time and datetime errors I can't pin-point)
<OscarL>
might wait till summer ends before running those test suites though!
<OscarL>
(I need a more efficient CPU :-D)
<nephele>
Okay, mostly done with controls now, except for the scrollbars. But those can be done "soonish" and I also have to fix these for Webkit
<nipos>
My computer just crashed while building your patch,even with using only one core to compile
<Begasus[m]>
lots of "U dlopen" with lgrep :)
<OscarL>
stupid test log messages give wrong info. it says: "testing/cffi0/test_ffi_backend.py::TestFFI::test_dlopen_handle" but "test_dlopen_handle" is on a different file :-D
<OscarL>
(test_ownlib.py)
<OscarL>
that "dl" test is skipped for Win32 and musl. I guess you could add an "or sys.platform[:5] == 'haiku'" there (line 391)
<Begasus[m]>
yeah trying there :)
<OscarL>
feel free to either ammend my PR, or apply your changes after merging mine as-is.
<Begasus[m]>
testing/cffi0/test_ffi_backend.py without failure +1 :D
<Begasus[m]>
running another test round
<Begasus[m]>
eeps at least on error reported
* OscarL
re-reads logs and thanks nephele for the tip regarding Steam.
<OscarL>
Too bad my only gamer friend lives in NZ :-D
<Begasus[m]>
ah, that file is missing the same part
<OscarL>
k time for me to go for now (darn data meter only updates when I disconnect/reconnect, LOL).
<OscarL>
Have a good one, everyone!
OscarL has quit [Quit: Gone with the wind]
nephele_xmpp has left #haiku [#haiku]
talos9 has joined #haiku
<Begasus[m]>
bugger, just a tad too late: 1917 passed, 122 skipped, 4 xfailed, 564 warnings in 233.85s (0:03:53)
nephele_xmpp has joined #haiku
talos has quit [Read error: Connection reset by peer]
talos has joined #haiku
talos9 has quit [Ping timeout: 480 seconds]
<nephele>
going out for a bit. Hopefully I can tackle other issues than control rendering when back :P
nephele has quit [Quit: Vision[]: i've been blurred!]
chilledfrogs has quit [Quit: connection reset by purr]
<nekobot>
[haiku/haiku] 6670202514af - ControlLook: Use B_PANEL_TEXT_COLOR for scrollbar arrows
<nephele>
waddlesplash: You said the file copy dialog needed layout kit treatment iirc. I could do that next, I think it also deserves to have the remaining time and finish time displayed at the same time. There is not much reason why this disalog has to be so tiny
<nephele>
It also has a hover popup that stays for like half a second :D
<waddlesplash>
we should start by replicating the existing layouts mostly 1:1 and not change too much
<waddlesplash>
if we want to adjust the dialog's design then we can do that after switching it to use layouts
x10z has joined #haiku
<waddlesplash>
right now it's tiny for me because I'm on HiDPI. without that, it looks fine
<waddlesplash>
so if we just scale with the system font size the metrics it already uses it will probably be OK
x10z has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
diver has joined #haiku
nephele_xmpp has left #haiku [Error from remote client]
<nephele>
waddlesplash: I'm not a fan of text that switches between two representations often. I don't know if this is done anywhere else in the UI. Anyhow, sure. I can make it use the layout kit, and in an additional commit change the style. Though this is more complicated than "just" rebuilding the layout to what we want anyway
<waddlesplash>
how so?
<nephele>
what is your question targeted towards?
<waddlesplash>
the last statement, "more complicated than..."
<waddlesplash>
the code already has logic for the switching
nephele_xmpp has joined #haiku
<nephele>
Yes, but making code for this once with a new layout is a one-time thing. Making a layout with the switching, and then going back and re-doing it is more work. If I "just" do the new layout I don't have to take care to make sure the switching code does not break (and does not break the new layout, because the label frequently changes size)
<nipos>
There's no simple command to add new strings from en.catkeys to all other locales,right?
<waddlesplash>
you don't need to do that
<waddlesplash>
the files will be automatically updated
<waddlesplash>
at least the ones in the Haiku tree anyway
<waddlesplash>
ones out-of-tree don't need to be updated either I don't think
<nipos>
I'm talking about my BeAIM app
<waddlesplash>
ah. I think you don't need to do that there either
<waddlesplash>
missing strings just fall back
<waddlesplash>
meanwhile, I am trying to install Alpine Linux in a VM because I want to test some things against musl, and I wanted something more than a container
<nipos>
Having them added would make it easier to translate using CatKeysEditor,but on the other hand it's maybe misleading to have the strings added but not translated
<waddlesplash>
got it installed (purely TUI), used "setup-desktop" to install a DE, rebooted ... and now my network connection is gone
<waddlesplash>
as in, ifconfig/ip now show only the loopback interface
BrunoSpr has quit [Quit: Vision[]: Ich wurde eingeweicht!]
<waddlesplash>
and the default install contains nothing to check the PCI device tree
<waddlesplash>
and dmesg doesn't have anything that looks helpful
<waddlesplash>
it feels like *every single time* I use Linux, something like this happens. and some people are still surprised that I quit using Linux bare-metal years ago due to all the problems I kept having to fix on it
<nephele>
waddlesplash: probably "conman" or "network manager" took over
<nephele>
normally alpine linux uses "ifup ifdown" from debian
<waddlesplash>
this is Alpine, I don't know if those exist
<nephele>
I know they exist, but only if you install a DE
<waddlesplash>
but even if they do, why would my network interface be totally gone?
diver has quit [Ping timeout: 480 seconds]
<waddlesplash>
the DE also doesn't start anyway
<waddlesplash>
sddm complains about missing symbols
<waddlesplash>
nephele: "eth0" doesn't even exist anymore I mean
<waddlesplash>
it's just not there
<nephele>
the network interface iirc only "appears" once ifup is called, try checking if network manager or conman is running as a service
<waddlesplash>
I don't mean "I don't have a DHCP lease", I mean "the network interface is nonexistent"
<nephele>
ifup eth0 doesn't work?
<waddlesplash>
no, it says "no such device"
<waddlesplash>
lol
<nipos>
Yeah,Linux is always frustrating.Yet they still have their fans who don't understand why I find Haiku or FreeBSD (for servers) a lot better
<nephele>
oh, perhaps then your problem is udev
<nephele>
you see either busybox provides files in /dev, or udev does (or devd)
<nephele>
and DE sometimes installs udev too
<waddlesplash>
nipos: I installed GhostBSD last week in another VM, it was somewhat annoying but much less so than this
<waddlesplash>
nephele: lol, how/when did you learn this? I never did
<nipos>
I mean,network management should really be a system thing and not depend on having one from dozens of userland network management tools installed.
<waddlesplash>
that's kind of crazy that it's up to userspace to lay out /dev properly
<waddlesplash>
what a mess
<nephele>
waddlesplash: Honestly, I don't know :(
<nephele>
But it's not knowledge that helps me, unless someone has linux problems
<waddlesplash>
well, Google is getting worse and worse, and it's not helping me find solutions here
<nipos>
Oh no,Weblate/Codeberg Translate doesn't support the catkeys format.Would have been too easy I guess :/
<nephele>
well there is *also* a filesystem provided by the kernel that might lay this out properly called devfs i think?
<nipos>
I mean,I could possible spin up a Polyglot instance for my stuff,but that's a bit much work for translating a single and yet rather unknown app.
<waddlesplash>
nipos: read the comments: it's already supported by the underlying library
<waddlesplash>
so it just needs to be added to weblate
<waddlesplash>
and a maintainer linked to an example commit adding support for a new format
<nipos>
Seems easy to do
<nipos>
Getting it somehow submitted is always the most difficult part for stuff hosted on M$' walled garden
<nephele>
maybe waddlesplash can submit it for you
<x512[m]>
waddlesplash: Linux use sysfs instead of devfs to provide device information by kernel.
janking has joined #haiku
janking has quit []
nephele_xmpp has left #haiku [Error from remote client]
<waddlesplash>
nephele: indeed there'e a message at he end of dmesg about eudev starting
<waddlesplash>
so, I guess it got installed indeed
<waddlesplash>
so... what is it not doing correctly then lol
<nephele>
oh noes, now downloads in Gnome Web can be erronously opened in konqueror in Haiku too xD
x10z has joined #haiku
<Begasus[m]>
lol
<nephele>
#justLinuxThings
<Begasus[m]>
just use the flatpak stuff :P
<nephele>
Yeah. right. I installed luanti with flatpack. It came up an alert that "Luanti is running in the background!" and then i dismissed the alert... and then it killed luanti every time after 5 seconds
<Begasus[m]>
Falkon and Iceweasel being buggy on github too
<nephele>
untill i removed that flag that it's not "allowed" to run in the background
<Begasus[m]>
heh
<Begasus[m]>
well, you don't get systemd with Haiku :)
<nephele>
the alert was my only chance to not make it kill luanti after 5 seconds for daring to run. There was no gui way to unset this flag
<Begasus[m]>
bug report done?
<nephele>
No. No clue who to send bug reports about linux misbehaving
x10z has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
nephele_xmpp has joined #haiku
<Begasus[m]>
heh, good point there :)
<Begasus[m]>
closing down here
<Begasus[m]>
cu peeps!
janking has joined #haiku
janking has quit []
nephele has quit [Quit: Vision[]: i've been blurred!]
<nephele_xmpp>
waddlesplash, up to 150MiB/s on linux
<nephele_xmpp>
but it does go down to 90MiB/s partially
MisthaLu has quit [Quit: Leaving]
<nephele_xmpp>
but that's still more than double the speed of haiku highest in the worst case
<nephele_xmpp>
(and haiku went down to like 13MiB/s)
jmairboeck has joined #haiku
chilledfrogs has quit [Quit: connection reset by purr]
_justin_kelly166686824 has quit []
gouchi has joined #haiku
illwieckz__ has quit [Remote host closed the connection]