Why does GitKraken crash after my last Fedora 36 update? - fedora

I enjoyed using GitKraken (v8.9) until I updated my current Fedora 36 distribution.
After the update, I had the problem that GitKraken crashed every time I wanted to open a repository tab or when I wanted to open the "File/Settings..." view.
I've updated GitKraken to the latest version and also looked at what the Fedora update includes, which may have caused the problem.
Additionally, I collected the coredump of the crash (via coredumpctl gdb and the crash log via journalctl -xf)
Command Line: $'/usr/share/gitkraken/gitkraken --type=renderer --enable-crashpad --crashpad-handler-pid=3178 --enable-crash-reporter=dde15cd3-eccc-4db0-abbf-4626ba853f5f,no_channel --user-data-dir=/home/hugo/.config/GitKraken --standard-schemes --secure-schemes --bypasscsp-schemes=sentry-ipc --cors-schemes=sentry-ipc --fetch-schemes=sentry-ipc --service-worker-schemes --streaming-schemes --app-path=/usr/share/gitkraken/resources/app.asar --no-sandbox --no-zygote --node-integration-in-worker --disable-gpu-compositing --lang=en-US --num-raster-threads=4 --enable-main-frame-before-activation --renderer-client-id=9 --launch-time-ticks=54958457852 --shared-files=v8_context_snapshot_data:100 --field-trial-handle=0,890488252484140780,4810308880598401477,131072 --disable-features=PlzServiceWorker,SpareRendererForSitePerProcess --enable-crashpad'
Executable: /usr/share/gitkraken/gitkraken
Control Group: /user.slice/user-1000.slice/user#1000.service/app.slice/app-gnome-gitkraken-2553.scope
Unit: user#1000.service
User Unit: app-gnome-gitkraken-2553.scope
Slice: user-1000.slice
Owner UID: 1000 (hugo)
Boot ID: 8bdd7f9dd7194682b2286579c96c8ec7
Machine ID: 70b436a43d284e9f9e7f1073bc0cf15b
Hostname: tag009442760151
Storage: /var/lib/systemd/coredump/core.gitkraken.1000.8bdd7f9dd7194682b2286579c96c8ec7.56008.1667909034000000.zst (present)
Disk Size: 77.8M
Message: Process 56008 (gitkraken) of user 1000 dumped core.
Module linux-vdso.so.1 with build-id 28de22885e5a5f761e8b05fe6d610d65bb875b04
Module nsfw.node without build-id.
Module libxkbfile.so.1 with build-id e9e99444598d67ff8514b6adfb390156611311d6
Stack trace of thread 56072:
#0 0x00007f26c336cffc __strlen_evex (libc.so.6 + 0x16cffc)
#1 0x00007f26bf3d351e _ZN14FontDescriptorC2EPKcS1_S1_S1_10FontWeight9FontWidthbb (fontmanager.node + 0xb51e)
#2 0x00007f26bf3d8f19 _Z20createFontDescriptorP10_FcPattern (fontmanager.node + 0x10f19)
#3 0x00007f26bf3d8fb9 _Z12getResultSetP10_FcFontSet (fontmanager.node + 0x10fb9)
#4 0x00007f26bf3d9182 _ZN15FontManagerImpl17getAvailableFontsEPP9ResultSet (fontmanager.node + 0x11182)
#5 0x00007f26bf3d7c7e _ZN18PromiseAsyncWorkerI9ResultSetE7ExecuteEv (fontmanager.node + 0xfc7e)
#6 0x00005602f9af87c1 n/a (gitkraken + 0x20257c1)
#7 0x00007f26c328cded start_thread (libc.so.6 + 0x8cded)
#8 0x00007f26c3312370 __clone3 (libc.so.6 + 0x112370)
With this information I contacted GitKraken support. The information I got from them is that it's a known issue that GitKraken can't handle some custom fonts, and the most common culprit is JetBrains fonts.
I also got a list of certain fonts that are known to cause problems viz
RedHatFont
JetBrains Mono font on Fedora 34 and Ubuntu 20.04 (see comments on this issue).
Vazir
Rubik
I didn't have any of these fonts installed, so it wasn't applicable to me.
Does someone face the same problem and has a solution to that?

I played around a bit with reverting packages I updated with in the buggy update, taking into account the information that GitKraken is having issues with some custom fonts.
And finally, I found a bunch of fonts related to the wine package that fixed the problem on rollback.
I didn't actually do the rollback, but completely uninstalled all wine related fonts because I don't need them anyway.
Removing following fonts solved my problem:
$ sudo dnf remove wine-\*-fonts
Dependencies resolved.
=====================================================================================================================================================
Package Architecture Version Repository Size
=====================================================================================================================================================
Removing:
wine-arial-fonts noarch 7.19-1.fc36 #updates 157 k
wine-courier-fonts noarch 7.19-1.fc36 #updates 170 k
wine-fixedsys-fonts noarch 7.19-1.fc36 #updates 37 k
wine-marlett-fonts noarch 7.19-1.fc36 #updates 32 k
wine-ms-sans-serif-fonts noarch 7.19-1.fc36 #updates 4.6 M
wine-small-fonts noarch 7.19-1.fc36 #updates 65 k
wine-symbol-fonts noarch 7.19-1.fc36 #updates 51 k
wine-system-fonts noarch 7.19-1.fc36 #updates 121 k
wine-tahoma-fonts noarch 7.19-1.fc36 #updates 300 k
wine-times-new-roman-fonts noarch 7.19-1.fc36 #updates 170 k
wine-webdings-fonts noarch 7.19-1.fc36 #updates 30 k
wine-wingdings-fonts noarch 7.19-1.fc36 #updates 35 k
Removing dependent packages:
wine-fonts noarch 7.19-1.fc36 #updates 0
Removing unused dependencies:
liberation-narrow-fonts noarch 2:1.07.6-8.fc36 #fedora 504 k
Transaction Summary
=====================================================================================================================================================
Remove 14 Packages
BTW, the 'old' (rollback) version of these wine fonts where GitKraken still worked was 7.12-2, but I haven't explicitly tried this rollback alone.

After searching around I found, that the problem might be loading fonts.
There is a support request ( might be done already ): https://feedback.gitkraken.com/suggestions/218000/add-ligature-font-support
With that hint I installed fira-code-fonts from the repository and this solved my problem to open preferences.
My System:
Gitkraken 8.10.3
Fedora 36

Related

how to use the new ntfs3 driver on fedora 34

linux kernel 5.15 comes with new ntfs3 driver, I'm on fedora 34 with kernel 5.15.6, but when I mount a device using ntfs3, it reports error:
# uname -a
Linux fedora 5.15.6-100.fc34.x86_64 #1 SMP Wed Dec 1 13:41:51 UTC 2021 x86_64 x86_64 x86_64 GNU/Linux
# mount -t ntfs3 /dev/sda1 /run/media/sify/Elements\ SE
mount: /run/media/sify/Elements SE: wrong fs type, bad option, bad superblock on /dev/sda1, missing codepage or helper program, or other error.
do I need to do anything special to use the new ntfs3 driver?
I solved this by repairing the hard disk on windows using chkdisk, it took a whole day and I lost some data. But after that, I can mount the disk using ntfs3. I suspect I've saved some malformed data on that disk before.

Missing library on 64 bit riscv Fedora

I am working on an assembly-based Forth (on top of Linux - https://github.com/mcmenaminadrian/riscyforth) for Risc-V and I can build it and run it on my real 64 bit hardware (a Nezha SBC running a Debian-based install).
I decided to test it also on QEMU with Fedora but it won't run:
[root#fedora-riscv riscyforth]# ./riscyforth
-bash: ./riscyforth: No such file or directory
This seems to be because I am missing a library:
readelf -a ./riscyforth
....
Program Headers:
Type Offset VirtAddr PhysAddr
FileSiz MemSiz Flags Align
PHDR 0x0000000000000040 0x0000000000010040 0x0000000000010040
0x0000000000000150 0x0000000000000150 R 0x8
INTERP 0x0000000000000190 0x0000000000010190 0x0000000000010190
0x000000000000000d 0x000000000000000d R 0x1
[Requesting program interpreter: /lib/ld.so.1]
And my understanding this is a 32 bit library - but my OS is 64 bit:
[root#fedora-riscv riscyforth]# uname -a
Linux fedora-riscv 5.10.6-200.0.riscv64.fc33.riscv64 #1 SMP Tue Jan 12 13:46:56 UTC 2021
riscv64 riscv64 riscv64 GNU/Linux
And, rather more importantly, so is the ELF:
[root#fedora-riscv riscyforth]# readelf -a ./riscyforth | more
ELF Header:
Magic: 7f 45 4c 46 02 01 01 00 00 00 00 00 00 00 00 00
Class: ELF64
Data: 2's complement, little endian
Version: 1 (current)
OS/ABI: UNIX - System V
ABI Version: 0
Type: EXEC (Executable file)
Machine: RISC-V
Version: 0x1
Entry point address: 0x10690
Start of program headers: 64 (bytes into file)
Start of section headers: 98992 (bytes into file)
Flags: 0x5, RVC, double-float ABI
Size of this header: 64 (bytes)
Size of program headers: 56 (bytes)
Number of program headers: 6
Size of section headers: 64 (bytes)
Number of section headers: 23
Section header string table index: 22
I have updated the Fedora install and rebooted (as there was a mix of -33 and -32 packages before) but this has not fixed the issue.
Why do I need to install a 32 library and which one do I need to install? I cannot find an obviously titled RPM.
Or do I need to put something in the Makefile for Fedora that I don't need on the Debian system?
This seems to be because I am missing a library:
You are not missing a library, you are missing the ELF interpreter (/lib/ld.so.1), also known as the dynamic linker or the loader.
And my understanding this is a 32 bit library
I am not sure what "this" in your statement refers to; your binary is definitely 64-bit, and couldn't use any 32-bit libraries in any case.
Why do I need to install a 32 library
You don't.
It appears that normally /lib/ld.so.1 should be present on a RISC-V system and be a symlink to /lib/ld-linux-riscv64-lp64d.so.1 (at least on GLIBC-based systems).
So your QEMU target environment is missing required parts.
I fixed the immediate problem with ln -s /lib/ld-linux-riscv64-lp64d.so.1 /lib/ld.so.1 but I still hope someone can answer why I needed this.
Because you somehow didn't install GLIBC into your QEMU environment properly.

How to recreate venv with mysql#5.5 with openssl 1.0.0 dependency on mac

I used to have a venv with mysql#5.5 installed. This somehow broke (I guess because I upgraded my python version) so I'm now trying to recreate it.
However I am getting the following error:
ImportError: dlopen(/Users/bartjonk/Envs/w-poc/lib/python3.9/site-packages/MySQLdb/_mysql.cpython-39-darwin.so, 2): Library not loaded: /usr/local/opt/openssl/lib/libssl.1.0.0.dylib Referenced from: /usr/local/opt/mysql#5.5/lib/libmysqlclient.18.dylib Reason: image not found.
I don't know how to fix it.
I can't do a brew switch ore create symlinks as recommended here, since openssl version 1.0.0 is no longer in my cellar:
`> ls -al /usr/local/Cellar/op*
total 0
drwxr-xr-x 3 bartjonk staff 96 2 nov 14:42 .
drwxrwxr-x 106 bartjonk admin 3392 6 nov 17:59 ..
drwxr-xr-x 14 bartjonk staff 448 2 nov 14:42 1.1.1h`
and brew install openssl#1.0.0s or any other variants (brew install openssl#1.0 etc.) only result in :
==> Searching for similarly named formulae...
Error: No similarly named formulae found.
Error: No available formula or cask with the name "openssl#1.0.2s".
==> Searching for a previously deleted formula (in the last month)...
Error: No previously deleted formula found.
==> Searching taps on GitHub...
Error: No formulae found in taps.
The link in the comment given by #Reinier Torenbeek (answer to: OpenSSL 1.0.2m on macOS) works like a charm.
I can now switch between the old and the new implementation using:
brew switch openssl 1.0.2m
You can check which versions are available using:
ls -al /usr/local/Cellar/op*

Octave 4.4.1 fails to start on 32 bit Windows 7

Octave 4.0.0 works on Windows 7 32 bit. I have just downloaded 4.4.1 and it has never worked.
Following information:
Problem signature:
Problem Event Name: APPCRASH
Application Name: octave-gui.exe
Application Version: 0.0.0.0
Application Timestamp: 00000000
Fault Module Name: Qt5Core.dll
Fault Module Version: 5.11.0.0
Fault Module Timestamp: 00000000
Exception Code: 40000015
Exception Offset: 002e8696
OS Version: 6.1.7601.2.1.0.256.48
Locale ID: 3081
Additional Information 1: d1e0
Additional Information 2: d1e00fdc0b78c108f75564157f84c2f0
Additional Information 3: f785
Additional Information 4: f7855e1e76ad3f0614cfa9b98110597e
Found a vaguely similar problem, failing to start GUI on Windows 10, applied suggested solution of setting "Disable display scaling on high DPI settings" under Compatibility tab of executable Properties.
If you download GNU octave from octave.org you'll see a list:
Windows-64 (recommended)
octave-4.4.1-w64-installer.exe (~ 238 MB) [signature]
octave-4.4.1-w64.7z (~ 267 MB) [signature]
octave-4.4.1-w64.zip (~ 481 MB) [signature]
Windows-32 (old computers)
octave-4.4.1-w32-installer.exe (~ 238 MB) [signature]
octave-4.4.1-w32.7z (~ 267 MB) [signature]
octave-4.4.1-w32.zip (~ 481 MB) [signature]
Of couse a 64bit built will NOT run on a Windoze7 32bit so you have to get octave-4.4.1-w32-installer.exe and install this

Symbolic package and Octave is working only for supersimple tasks. Am I doing anything wrong?

I have installed (on Windows machine (Win7 64bit)):
python 3.5.0 and sympy 0.7.6.1
octave 4.0.0.0 and symbolic 2.2.2
This is what I get when I try to do anything with more than one operator:
>> pkg load symbolic
>> syms x
OctSymPy v2.2.2: this is free software without warranty, see source.
Initializing communication with SymPy using a popen2() pipe.
Detected Windows: using "winwrapy.bat" to workaround Octave bug #43036
Some output from the Python subprocess (pid 28092) might appear next.
OctSymPy: Communication established. SymPy v0.7.6.1.
Python 3.5.0 (v3.5.0:374f501f4567, Sep 13 2015, 02:16:59) [MSC v.1900 32 bit (Int
el)]
>> x^2
ans = (sym)
2
x
>> x^2 + 2*x + 1
PYTHON: Error in var export
Resetting the octsympy communication mechanism
Closing the Python pipe...
error: python_cmd: unexpected return
error: called from
python_cmd at line 143 column 6
mtimes at line 42 column 5
>>
What is wrong?