Kde-application - KDE-Applications - telepathy -kde Faiulre and konsole about archpower

[mrbubzie93]

I see some of the kde stuff is now in testing…do those get installed automatically or do I have to switch to the testing branch to install…sorry I’m coming from Ubuntu just wasn’t sure how the testing branch worked (still reading the manual on there wiki). Thanks in advance :)

[kth5]

@mrbubzie93 You need to enable the [testing] and [testing-any] repos that are commented in the shipped /etc/pacman.conf. Then just pacmna -Syu like you would on any Arch-like.

The new versions in [testing] to not solve the konsole issues though, I've been experimenting with debug builds but it's tough on 32bit systems with just 980Mbyte usable memory so haven't gotten any results other than finding out that drkonqi also can't submit crash reports on powerpc. :-/

[mrbubzie]

that’s understandable besides xfce are there any other desktop environment available like lxqt, lxde, gnome…have had a chance to look yes! Sent from my iPhone On Jan 9, 2024, at 9:49 AM, Alexander Baldeck ***@***.***> wrote:  @mrbubzie93<https://github.com/mrbubzie93> You need to enable the [testing] and [testing-any] repos that are commented in the shipped /etc/pacman.conf. Then just pacmna -Syu like you would on any Arch-like. The new versions in [testing] to not solve the konsole issues though, I've been experimenting with debug builds but it's tough on 32bit systems with just 980Mbyte usable memory so haven't gotten any results other than finding out that drkonqi also can't submit crash reports on powerpc. :-/ — Reply to this email directly, view it on GitHub<#67 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/BC5SKVZECEGN3JVGFULCJZ3YNVKH5AVCNFSM6AAAAABA2GCIVWVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTQOBTGE4DSOJUGE>. You are receiving this because you authored the thread.Message ID: ***@***.***>

[mrbubzie]

that’s understandable besides xfce are there any other desktop environment available like lxqt, lxde, gnome?…haven’t had a chance to look yet! But I would go as far to say maybe just pull kde all together on the ppc32 side and maybe just focus on lxqt and xfce or even mate. Just my opinion…kde is a lot more complicated than any other desktop environment so maybe just keep kde for the two higher architectures? As I said if you need a remote machine to compile what not then hit me up

…

Sent from my iPhone On Jan 9, 2024, at 9:49 AM, Alexander Baldeck ***@***.***> wrote:  @mrbubzie93<https://github.com/mrbubzie93> You need to enable the [testing] and [testing-any] repos that are commented in the shipped /etc/pacman.conf. Then just pacmna -Syu like you would on any Arch-like. The new versions in [testing] to not solve the konsole issues though, I've been experimenting with debug builds but it's tough on 32bit systems with just 980Mbyte usable memory so haven't gotten any results other than finding out that drkonqi also can't submit crash reports on powerpc. :-/ — Reply to this email directly, view it on GitHub<#67 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/BC5SKVZECEGN3JVGFULCJZ3YNVKH5AVCNFSM6AAAAABA2GCIVWVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTQOBTGE4DSOJUGE>. You are receiving this because you authored the thread.Message ID: ***@***.***>

[mrbubzie93]

Although I just installed the latest test, everything is actually working it’s really just Konsole, which is not a big deal because you can switch to the terminal when you're on the desktop with shortcuts. It's not really a big deal. It's pretty stable! I've been playing around with the testing branch, and I'm really happy with this and impressed. Power management is now working – I mean, it's still slow obviously, but you sir are doing the work of a god 😂. So thank you for making this possible. Couldn't get KDE under Debian, older versions of Kubuntu, or other PPC distro that offered KDE. This is absolutely insane how well it works on ancient hardware.

[mrbubzie]

If your looking add another browser to the repo Eolie is epiphany on steroids it actually works and renders correctly. Were you able to get epiphany working on g4 it just hangs for me Sent from my iPhone On Jan 9, 2024, at 11:24 PM, mrbubzie93 ***@***.***> wrote:  Although I just installed the latest test everyone is actually it’s just konsole which is not a big deal because you can switch to terminal when your on the desktop with shortcuts it’s not really a big deal. It pretty stable! — Reply to this email directly, view it on GitHub<#67 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/BC5SKV43SX2CJMKHAJ73OTLYNYJYVAVCNFSM6AAAAABA2GCIVWVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTQOBUGE3TAMZQGI>. You are receiving this because you authored the thread.Message ID: ***@***.***>

[kth5]

Just as a side note:

I'm currently working on Plasma 6 instead of fixing the old 5.x branch of things and the problem with Konsole seems to persist.

[mrbubzie]

Hey, thanks for following up. The issue with Konsole is due to its reliance on graphics acceleration, which is currently broken, causing the observed problems. Compiling from source won't address this issue. The key now is to ensure Mesa is working properly and then focus on secondary issues. I've been looking into it in the last few days. It's also affecting other things like Epiphany and other WebKit browsers, defaulting due to broken graphics drivers. If you check the logs, you can see where it's segfaulting, and below that is an error related to the GPU.

…

Sent from my iPhone On Feb 8, 2024, at 7:16 AM, Alexander Baldeck ***@***.***> wrote:  Just as a side note: I'm currently working on Plasma 6 instead of fixing the old 5.x branch of things and the problem with Konsole seems to persist. — Reply to this email directly, view it on GitHub<#67 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/BC5SKV3MGLBBXCWYHOA7I2DYSS6YFAVCNFSM6AAAAABA2GCIVWVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSMZTHE4TINJYG4>. You are receiving this because you authored the thread.Message ID: ***@***.***>

[kth5]

It's actually a two fold sword. I've been working on this on and off over the past month or so ($DAYJOB & $FAMILY).

qt6 has GPU accelleration disabled by default now as it was and always has been with qt5 as well (Void and a few others do this as well). Even so, konsole needed disabling the so called "BidiRenderer" to get it to work. I submitted a few crashes using drkonqi though in various circumstances. Let's see what comes off of that. (commit: 28df381)

When it comes to Plasma 6, which accelleration disabled in qt6 (commit: 34f0428) kwin_x11 can start on powerpc and render properly. Launch typical applications and all. However, startplasma-x11 won't work due to plasmashell crashing due to a buffer overflow pretty close to libc and I just haven't had the time to traverse the debug symbols down to what it could be.

My test chain for Plasma 6 is as follows:

• launch X.org using the following .xinitrc

export QMLSCENE_DEVICE=softwarecontext
export QT_DEBUG_PLUGINS=1
KWIN_COMPOSE=N kwin_x11 &
#exec startplasma-x11 # the ultimate goal
exec xterm

• launch prerequisites manually

/usr/lib/kglobalacceld &
/usr/lib/kactivitymanagerd &

• launch plasmashell with debugging

plasmashell --no-respawn --qmljsdebugger

Hint: This won't be "fun" on actual hardware with less than 2GB RAM even when not using full debug symbols.

The lowest levels of the backtrace will get into libc territory where buffer overflow protection kicks in and kills the process. Someone with more time may want to have a look at this in more detail.

#0  0xf4a096d0 in ??? () at /usr/lib/libc.so.6
#1  0xf49b3434 in raise () at /usr/lib/libc.so.6
#2  0xf499aca8 in abort () at /usr/lib/libc.so.6
#3  0xf49f8c30 in ??? () at /usr/lib/libc.so.6
#4  0xf4a9f708 in __fortify_fail () at /usr/lib/libc.so.6
#5  0xf4a9ecf0 in __chk_fail () at /usr/lib/libc.so.6
#6  0xf49cdaf8 in ??? () at /usr/lib/libc.so.6
#7  0xf49ce034 in ??? () at /usr/lib/libc.so.6
#8  0xf49d5e28 in ??? () at /usr/lib/libc.so.6
#9  0xf49ee750 in ??? () at /usr/lib/libc.so.6
#10 0xf4aa0670 in __sprintf_chk () at /usr/lib/libc.so.6
#11 0xf461de6c in sprintf (__fmt=0xf473d570 "%6llx_%2x_", __s=0xfffeee00 "400000055_ 2") at /usr/include/bits/stdio2.h:30
#12 KIO::ConnectionBackend::sendCommand (data=..., cmd=<optimized out>, this=0x254b460)
    at /usr/src/debug/kio/kio-6.0.0/src/core/connectionbackend.cpp:177
#13 KIO::Connection::sendnow (this=<optimized out>, cmd=<optimized out>, data=...) at /usr/src/debug/kio/kio-6.0.0/src/core/connection.cpp:186
#14 0xf461dfbc in KIO::ConnectionPrivate::dequeue (this=0x209d910) at /usr/src/debug/kio/kio-6.0.0/src/core/connection.cpp:26
#15 KIO::ConnectionPrivate::dequeue (this=0x209d910) at /usr/src/debug/kio/kio-6.0.0/src/core/connection.cpp:19
#16 0xf46f5ea8 in KIO::ConnectionServer::setNextPendingConnection (conn=0x1c28da0, this=<optimized out>)
    at /usr/include/c++/13.2.1/bits/unique_ptr.h:199
#17 KIO::Worker::accept (this=0x1f2c8e0) at /usr/src/debug/kio/kio-6.0.0/src/core/worker.cpp:57
#18 0xf501c4d4 in QtPrivate::QSlotObjectBase::call
    (a=0xfffeeeec, r=0x1f2c8e0, this=0x2018db0, this=<optimized out>, r=<optimized out>, a=<optimized out>)
    at /usr/src/debug/qt6-base/qtbase-everywhere-src-6.6.2/src/corelib/kernel/qobjectdefs_impl.h:433
#19 doActivate<false> (sender=0xed405930, signal_index=3, argv=0xfffeeeec)
    at /usr/src/debug/qt6-base/qtbase-everywhere-src-6.6.2/src/corelib/kernel/qobject.cpp:4039
#20 0xf501c4d4 in QtPrivate::QSlotObjectBase::call
    (a=0xfffeef6c, r=0xed405930, this=0x1c7bbd0, this=<optimized out>, r=<optimized out>, a=<optimized out>)
    at /usr/src/debug/qt6-base/qtbase-everywhere-src-6.6.2/src/corelib/kernel/qobjectdefs_impl.h:433
#21 doActivate<false> (sender=0x1efb870, signal_index=5, argv=0xfffeef6c)
    at /usr/src/debug/qt6-base/qtbase-everywhere-src-6.6.2/src/corelib/kernel/qobject.cpp:4039
#22 0xf501c4d4 in QtPrivate::QSlotObjectBase::call
    (a=0xfffeefec, r=0x1efb870, this=0x2018d50, this=<optimized out>, r=<optimized out>, a=<optimized out>)
    at /usr/src/debug/qt6-base/qtbase-everywhere-src-6.6.2/src/corelib/kernel/qobjectdefs_impl.h:433
#23 doActivate<false> (sender=0x1e2e060, signal_index=3, argv=0xfffeefec)
    at /usr/src/debug/qt6-base/qtbase-everywhere-src-6.6.2/src/corelib/kernel/qobject.cpp:4039
#24 0xf40e9724 in QLocalServerPrivate::_q_onNewConnection (this=0x1bfbc60)
    at /usr/src/debug/qt6-base/qtbase-everywhere-src-6.6.2/src/network/socket/qlocalserver_unix.cpp:275

I am still open to it being a bug somewhere in Arch POWER of course and patches are welcome.