Librem5 issueshttps://source.puri.sm/groups/Librem5/-/issues2023-06-20T13:26:26Zhttps://source.puri.sm/Librem5/debs/gtk4/-/issues/8GTK4 file chooser not adaptive2023-06-20T13:26:26ZGuido GuntherGTK4 file chooser not adaptiveHere's what you get e.g. with phosh-mobile-settings on current byzantium:
![image](/uploads/6c21a2bad73217951325f7893f611926/image.png)
There's things cut off to the right (and single touch doesn't move into folders).
Funny enough nh...Here's what you get e.g. with phosh-mobile-settings on current byzantium:
![image](/uploads/6c21a2bad73217951325f7893f611926/image.png)
There's things cut off to the right (and single touch doesn't move into folders).
Funny enough nheko (a Qt app) gets an adaptive file portal:
![image](/uploads/ada1b545b16add3e4c1be0b7bf1d52c9/image.png)
@alexander.mikhaylenko filedialog is ticket off as adaptive via portal in #4 hence i'm filing this separately.
/cc @agxhttps://source.puri.sm/Librem5/feedbackd/-/issues/71Request to add profile-setting scheduling2023-04-27T12:36:19ZAdam PlumbRequest to add profile-setting schedulingIt would be great if I could set an automatic do-not-disturb schedule in my phone so I don't need to manually set it to silent at night and un-silent in the morning. I considered creating a separate scheduling daemon that uses the exist...It would be great if I could set an automatic do-not-disturb schedule in my phone so I don't need to manually set it to silent at night and un-silent in the morning. I considered creating a separate scheduling daemon that uses the existing dbus interface to set the profile, but it seems that this feature might need to be in feedbackd to work effectively. If the phone is asleep while passing through a profile change time, then when a call/text comes in, it would need to immediately know which profile to use.
My thinking is feedbackd can set up a new profile name like "auto", then have a separate interface for setting the profile-switching scheduling. I'm not sure how the interface would look, but we could have two settings maybe like this
```
gsettings set org.sigxcpu.feedbackd auto1_time 0800
gsettings set org.sigxcpu.feedbackd auto1_profile full
gsettings set org.sigxcpu.feedbackd auto2_time 2200
gsettings set org.sigxcpu.feedbackd auto2_profile silent
```https://source.puri.sm/Librem5/debs/squeekboard/-/issues/6Should provide /usr/share/applications/sm.puri.OSK0.desktop via /etc/alternat...2023-07-18T19:54:28ZGuido GuntherShould provide /usr/share/applications/sm.puri.OSK0.desktop via /etc/alternatives/Phosh-OSKI'd be good if squeekboard would handle sm.puri.OSK0 via `update-alternatives` like [librem5-base](https://source.puri.sm/Librem5/librem5-base/-/merge_requests/328) and [phosh-osk-stub](https://gitlab.gnome.org/guidog/phosh-osk-stub/-/me...I'd be good if squeekboard would handle sm.puri.OSK0 via `update-alternatives` like [librem5-base](https://source.puri.sm/Librem5/librem5-base/-/merge_requests/328) and [phosh-osk-stub](https://gitlab.gnome.org/guidog/phosh-osk-stub/-/merge_requests/32) do.
Otherwise there's no OSK for phosh's session which doesn't end well (with a black screen in byzantium (not nice) with a fail whale in crimson (correct way), see See https://source.puri.sm/Librem5/OS-issues/-/issues/319#note_236833 .
The easiest fix would be to base this package on Debian's squeekboard packaging which handles this since 1.20.0-2 (Nov 22).https://source.puri.sm/Librem5/linux/-/issues/477After software update, no WiFi adapter found2023-05-02T20:31:56ZM GamboniAfter software update, no WiFi adapter foundMy Wi-Fi adapter has been flaky for a long time (*usually* working, but sometimes disappearing. When that would happen cycling the kill-switch or at worst a reboot would bring it back) but since April 31st it’s not coming up at all. I di...My Wi-Fi adapter has been flaky for a long time (*usually* working, but sometimes disappearing. When that would happen cycling the kill-switch or at worst a reboot would bring it back) but since April 31st it’s not coming up at all. I did a software update a few days before that, so I believe it's stopped working at that point.
There’s no Wi-Fi indicator at the top, opening Wi-Fi settings says “No Wi-Fi adapter found”. I tried flicking the kill switch up and down, rebooting, shutting down and restarting, waiting a couple hours (it used to appear a couple minutes after boot before).
Apparently some people had this issue because their `/boot` was full but `df -h /boot` says 35% used, 281M available which I suppose is plenty.
When I disable the kill switch (i.e. supposedly enable the adapter), the following lines appear in `dmesg -w`:
```
[ 751.655912] mmc1: new high speed SDIO card at address fffd
[ 751.657097] redpine_91x: rsi_probe: ***** 9116 Module *****
[ 751.657879] redpine_91x: redpine_hal_device_init: oper_mode = 5, coex_mode = 2
[ 751.659052] redpine_91x: Received Board Version Number: 10
[ 751.660155] redpine_91x: rsi_load_firmware: Loading file RS9116_NLINK_WLAN_BT_IMAGE.rps
[ 753.054966] redpine_91x: *** Flash is Empty ***
[ 753.054990] redpine_91x: FW Length = 329216 bytes
[ 753.055000] redpine_91x: ================================================
[ 753.055004] redpine_91x: ================ RSI Version Info ==============
[ 753.055009] redpine_91x: ================================================
[ 753.055014] redpine_91x: FW Version : 2.1.0.0
[ 753.055021] redpine_91x: RSI FW Version : 1610.2.1.ab.ffbf
[ 753.055030] redpine_91x: Driver Version : RS9116.NB0.NL.GNU.LNX.OSD.2.0.0.0024
[ 753.055036] redpine_91x: Operating mode : 5 [Wi-Fi STA + BT EDR]
[ 753.055042] redpine_91x: Firmware file : RS9116_NLINK_WLAN_BT_IMAGE.rps
[ 753.055047] redpine_91x: ================================================
[ 753.063343] redpine_91x: Burn Command Pass.. Upgrading the firmware
[ 756.635879] redpine_91x: bl_cmd: Command EOF_REACHED (45) writing failed..
[ 756.635905] redpine_91x: Firmware upgrade failed - attempt RAM load
[ 756.635911] redpine_91x: Ram file magic 0xa55a size 329152
[ 756.637265] redpine_91x: ================================================
[ 756.637274] redpine_91x: ================ RSI Version Info ==============
[ 756.637279] redpine_91x: ================================================
[ 756.637285] redpine_91x: FW Version : 2.1.0.0
[ 756.637292] redpine_91x: RSI FW Version : 1610.2.1.ab.ffbf
[ 756.637300] redpine_91x: Driver Version : RS9116.NB0.NL.GNU.LNX.OSD.2.0.0.0024
[ 756.637306] redpine_91x: Operating mode : 5 [Wi-Fi STA + BT EDR]
[ 756.637313] redpine_91x: Firmware file : RS9116_NLINK_WLAN_BT_IMAGE.rps
[ 756.637318] redpine_91x: ================================================
[ 756.684738] redpine_91x: ***** Firmware Loading successful *****
[ 756.684960] redpine_91x: RAM load successful
[ 756.684967] redpine_91x: ***** Firmware Loading successful *****
```
which I suppose imply the card was found?
I see that firmware file in `/usr/lib/firmware`:
`-rw-r–r-- 1 root root 329216 4 mai 2022 /usr/lib/firmware/RS9116_NLINK_WLAN_BT_IMAGE.rps`
Note that it reports the file is one year old.
However `nmcli device` doesn't find the card:
```
DEVICE TYPE STATE CONNECTION
cdc-wdm0 gsm connected Salt Internet - Postpaid
lxcbr0 bridge connected (externally) lxcbr0
usb0 ethernet unavailable --
lo loopback unmanaged --
```
(There should be a wlan0 somewhere there, right?)
`uname -r`:
`6.2.0-1-librem5`
The WWAN adapter works fine.
What can I try next?https://source.puri.sm/Librem5/linux/-/issues/4796.3: don't build linux-image-librem5 binary package for linux-next-librem52023-05-09T07:52:10ZMartin Kepplinger6.3: don't build linux-image-librem5 binary package for linux-next-librem5debian packaing was reworked for 6.3 quite a bit, see `git log --oneline v6.2..v6.3 scripts/package/*` and dpkg-buildpackage now builds linux-image-librem5 in any case. We must not build that for linux-next-librem5 source packages though.debian packaing was reworked for 6.3 quite a bit, see `git log --oneline v6.2..v6.3 scripts/package/*` and dpkg-buildpackage now builds linux-image-librem5 in any case. We must not build that for linux-next-librem5 source packages though.https://source.puri.sm/Librem5/zephyr/-/issues/1add Debian packaging2023-06-30T10:23:55ZMartin Kepplingeradd Debian packagingsuggested source package name: zephyr-librem5
suggested binary package name: zephyr-image-librem5
the binary package should simply install the *.elf binary to /lib/firmware/. (and get the license info correct)
what do you think?
@evan...suggested source package name: zephyr-librem5
suggested binary package name: zephyr-image-librem5
the binary package should simply install the *.elf binary to /lib/firmware/. (and get the license info correct)
what do you think?
@evangelos.tzarashttps://source.puri.sm/Librem5/feedbackd/-/issues/72Default vibration theme for incoming call are too intensive2023-05-11T12:46:45ZVasiliy BelyjDefault vibration theme for incoming call are too intensiveRight now on incoming call the phone vibrates non-stop. This appears some kind of annoying and in most of cases out there vibration alarm on incoming call is short vibrating with pauses in between.
The patch changes a behaviour to a mor...Right now on incoming call the phone vibrates non-stop. This appears some kind of annoying and in most of cases out there vibration alarm on incoming call is short vibrating with pauses in between.
The patch changes a behaviour to a more convenient one.[feedbackd.patch](/uploads/dd34be80ebdcbe1edc8bd0106e0392f3/feedbackd.patch)https://source.puri.sm/Librem5/debs/ModemManager/-/issues/54Unable to send SMS on 4G with VoLTE enabled2023-05-29T14:00:19ZJan VlugUnable to send SMS on 4G with VoLTE enabledI cannot send SMS messages any more with Chatty. This used to work fine in the past.
I see this in the logging when sending an SMS message:
```
May 10 20:24:47 avalon mmsdtng[840]: ../plugins/modemmanager.c:cb_dbus_signal_sms_added() Go...I cannot send SMS messages any more with Chatty. This used to work fine in the past.
I see this in the logging when sending an SMS message:
```
May 10 20:24:47 avalon mmsdtng[840]: ../plugins/modemmanager.c:cb_dbus_signal_sms_added() Got new SMS
May 10 20:24:47 avalon mmsdtng[840]: ../plugins/modemmanager.c:mmsd_check_pdu_type() This is not an SMS being received, do not care
```
Because I had to send an important message, I put the SIM-card back in my Android phone, where sending worked fine.
What could be wrong? How can I get more details on what goes wrong?
Receiving SMS messages still works fine.
Another interesting observation is that there appears a little red cross after approximately 25 seconds next to the SMS messages. So, somehow Chatty knows that something went wrong. But I do not know how to get any additional information, like an error message. In the logging is nothing else than the lines above.
Now I’m thinking of it, the issue of not being able to send SMS messages might have started after a modem firmware update that I did a while ago. But it might also be totally unrelated, it is just a hypothesis.
See also [this forum post](https://forums.puri.sm/t/sending-sms-messages-stopped-working-how-to-debug/20233/3).
Chatty version: 0.7.2-1pureos1https://source.puri.sm/Librem5/millipixels/-/issues/85Auto White Balance going mental2023-05-25T13:33:57ZEvangelos Ribeiro TzarasAuto White Balance going mentalIn lower light conditions I have seen very rapidly switching backwards and forward from F11 and A profiles.
This will produce interesting effects on a video, as can be seen below.
Lighting conditions were: artificial light in hackspace...In lower light conditions I have seen very rapidly switching backwards and forward from F11 and A profiles.
This will produce interesting effects on a video, as can be seen below.
Lighting conditions were: artificial light in hackspace at night (reproducible at Debian Reunion HH after ~22:00)
![awb_gone_crazy](/uploads/fe5a56bf53889ce086cbeaea844c35da/awb_gone_crazy.webm)
It would maybe great to rate-limit that :Phttps://source.puri.sm/Librem5/OS-issues/-/issues/326Cannot scale-to-fit mumble2023-06-02T16:15:39ZEvangelos Ribeiro TzarasCannot scale-to-fit mumbleThe preferences window in Mumble overflows a lot.
p-m-s does not show a separate entry in the compositor panel.
It was previously suspected that the problem being that it's run through xwayland where scaling does not work,
however I've...The preferences window in Mumble overflows a lot.
p-m-s does not show a separate entry in the compositor panel.
It was previously suspected that the problem being that it's run through xwayland where scaling does not work,
however I've verified that it runs indeed under wayland (`WAYLAND_DEBUG=1` works).
See this video (with ~13MB too big to attach to gitlab directly):
https://fortysixandtwo.eu/upload/mumble1.webm
I will check wayland logs for what type of surface is being opened for the preferences window, but I wanted to share my initial findings already (so I won't forget to do it later).https://source.puri.sm/Librem5/linux/-/issues/483release to crimson and backport to byzantium2023-07-06T09:52:26ZMartin Kepplingerrelease to crimson and backport to byzantiumShould we start releasing to crimson? I'm not yet aware of patches that do not apply to a crimson system, but if so, I'd handle additional patches to the byzantium kernel against the crimson kernel - exactly how we did it with byzantium ...Should we start releasing to crimson? I'm not yet aware of patches that do not apply to a crimson system, but if so, I'd handle additional patches to the byzantium kernel against the crimson kernel - exactly how we did it with byzantium and amber-phone.https://source.puri.sm/Librem5/librem5-flash-image/-/issues/14Use luks variant by default2023-06-25T18:42:17ZSebastian KrzyszkowiakUse luks variant by defaultWe're still defaulting to `plain`. `luks` variant isn't experimental anymore and should probably be used by default now.We're still defaulting to `plain`. `luks` variant isn't experimental anymore and should probably be used by default now.https://source.puri.sm/Librem5/librem5-devkit-tools/-/issues/54setup-pureos-pkg-repo: ProjectBranch object has no attribute protect2023-06-16T05:20:39ZEvangelos Ribeiro Tzarassetup-pureos-pkg-repo: ProjectBranch object has no attribute protectWith python3-gitlab 1:3.12.0-1 the setup-pureos-pkg-repo is unhappy:
```
$ setup-pureos-pkg-repo -i 166
setup-pureos-pkg-repo:102: DeprecationWarning: `as_list=False` is deprecated and will be removed in a future version. Use `iterator...With python3-gitlab 1:3.12.0-1 the setup-pureos-pkg-repo is unhappy:
```
$ setup-pureos-pkg-repo -i 166
setup-pureos-pkg-repo:102: DeprecationWarning: `as_list=False` is deprecated and will be removed in a future version. Use `iterator=True` instead. (python-gitlab: /usr/lib/python3/dist-packages/gitlab/client.py:917)
projects = gl.projects.list(as_list=False)
Traceback (most recent call last):
File "/home/fortysixandtwo/git/librem5-dev-tools/setup-pureos-pkg-repo", line 199, in <module>
sys.exit(main(sys.argv))
^^^^^^^^^^^^^^
File "/home/fortysixandtwo/git/librem5-dev-tools/setup-pureos-pkg-repo", line 130, in main
def_branch.protect()
^^^^^^^^^^^^^^^^^^
File "/usr/lib/python3/dist-packages/gitlab/base.py", line 134, in __getattr__
raise AttributeError(message)
AttributeError: 'ProjectBranch' object has no attribute 'protect'. Did you mean: 'protected'?
```
```
>>> def_branch = project.branches.get('pureos/latest')
>>> def_branch.
def_branch.asdict( def_branch.developers_can_push def_branch.pprint()
def_branch.attributes def_branch.encoded_id def_branch.project_id
def_branch.can_push def_branch.get_id() def_branch.protected
def_branch.commit def_branch.manager def_branch.to_json(
def_branch.default def_branch.merged def_branch.web_url
def_branch.delete( def_branch.name
def_branch.developers_can_merge def_branch.pformat()
```
So there is now the `protected` attribute, but the `protect()` method seems to have moved elsewhere.
The upstream change is [here](https://github.com/python-gitlab/python-gitlab/commit/9656a16f9f34a1aeb8ea0015564bad68ffb39c26).
[API reference](https://python-gitlab.readthedocs.io/en/stable/api/gitlab.v4.html#gitlab.v4.objects.ProjectProtectedBranchManager) and
[upstream docs](https://github.com/python-gitlab/python-gitlab/blob/main/docs/gl_objects/protected_branches.rst)https://source.puri.sm/Librem5/librem5-dev-tools/-/issues/12setup-pureos-pkg-repo: ProjectBranch object has no attribute protect2023-06-20T09:54:59ZEvangelos Ribeiro Tzarassetup-pureos-pkg-repo: ProjectBranch object has no attribute protectWith python3-gitlab 1:3.12.0-1 the setup-pureos-pkg-repo is unhappy:
```
$ setup-pureos-pkg-repo -i 166
setup-pureos-pkg-repo:102: DeprecationWarning: `as_list=False` is deprecated and will be removed in a future version. Use `iterator...With python3-gitlab 1:3.12.0-1 the setup-pureos-pkg-repo is unhappy:
```
$ setup-pureos-pkg-repo -i 166
setup-pureos-pkg-repo:102: DeprecationWarning: `as_list=False` is deprecated and will be removed in a future version. Use `iterator=True` instead. (python-gitlab: /usr/lib/python3/dist-packages/gitlab/client.py:917)
projects = gl.projects.list(as_list=False)
Traceback (most recent call last):
File "/home/fortysixandtwo/git/librem5-dev-tools/setup-pureos-pkg-repo", line 199, in <module>
sys.exit(main(sys.argv))
^^^^^^^^^^^^^^
File "/home/fortysixandtwo/git/librem5-dev-tools/setup-pureos-pkg-repo", line 130, in main
def_branch.protect()
^^^^^^^^^^^^^^^^^^
File "/usr/lib/python3/dist-packages/gitlab/base.py", line 134, in __getattr__
raise AttributeError(message)
AttributeError: 'ProjectBranch' object has no attribute 'protect'. Did you mean: 'protected'?
```
```
>>> def_branch = project.branches.get('pureos/latest')
>>> def_branch.
def_branch.asdict( def_branch.developers_can_push def_branch.pprint()
def_branch.attributes def_branch.encoded_id def_branch.project_id
def_branch.can_push def_branch.get_id() def_branch.protected
def_branch.commit def_branch.manager def_branch.to_json(
def_branch.default def_branch.merged def_branch.web_url
def_branch.delete( def_branch.name
def_branch.developers_can_merge def_branch.pformat()
```
So there is now the `protected` attribute, but the `protect()` method seems to have moved elsewhere.
The upstream change is [here](https://github.com/python-gitlab/python-gitlab/commit/9656a16f9f34a1aeb8ea0015564bad68ffb39c26).
[API reference](https://python-gitlab.readthedocs.io/en/stable/api/gitlab.v4.html#gitlab.v4.objects.ProjectProtectedBranchManager) and
[upstream docs](https://github.com/python-gitlab/python-gitlab/blob/main/docs/gl_objects/protected_branches.rst)https://source.puri.sm/Librem5/librem5-base/-/issues/72Switch to pureos/latest, make pureos/byzantium a backport branch2023-07-18T12:39:20ZSebastian KrzyszkowiakSwitch to pureos/latest, make pureos/byzantium a backport branchStill no need to do it *right now*, but filling an issue with a label so it won't be forgotten.
CI will have to be switched from `landing` to `byzantium-updates-proposed` in pureos/byzantium branch.Still no need to do it *right now*, but filling an issue with a label so it won't be forgotten.
CI will have to be switched from `landing` to `byzantium-updates-proposed` in pureos/byzantium branch.https://source.puri.sm/Librem5/geary/-/issues/26Revisit adaptiveness patches2023-06-21T09:34:08ZGuido GuntherRevisit adaptiveness patchesRecent versions should be mostly good for crimsonRecent versions should be mostly good for crimsonhttps://source.puri.sm/Librem5/librem5-base/-/issues/73Use "Protected" field2023-07-12T06:32:29ZSebastian KrzyszkowiakUse "Protected" fieldSince dpkg 1.20.1, deb packages can set `Protected: yes` to make dpkg refuse to uninstall the package without applying a force flag. Some binary packages produced by librem5-base could use it to prevent metapackages from being uninstalle...Since dpkg 1.20.1, deb packages can set `Protected: yes` to make dpkg refuse to uninstall the package without applying a force flag. Some binary packages produced by librem5-base could use it to prevent metapackages from being uninstalled as conflict resolution on upgrades.
dpkg in byzantium is already fresh enough to support this field.
https://wiki.debian.org/Teams/Dpkg/Spec/ProtectedFieldhttps://source.puri.sm/Librem5/linux/-/issues/486Need a newer kernel for crimson2023-07-17T09:38:46ZGuido GuntherNeed a newer kernel for crimsonWe currently have there
```
librem5-base-defaults : Depends: linux-image-librem5 (>= 6.3.7pureos2) but 6.2.13pureos3 is to be installed or
linux-image-amd64 but it is not installable
```
@martin.kepp...We currently have there
```
librem5-base-defaults : Depends: linux-image-librem5 (>= 6.3.7pureos2) but 6.2.13pureos3 is to be installed or
linux-image-amd64 but it is not installable
```
@martin.kepplinger I can handle that (I'd just bump the version and push that to `pureos/landing`) but wanted to check with you first.https://source.puri.sm/Librem5/linux/-/issues/487fix linux-image-librem5 binary package2023-06-29T07:10:40ZMartin Kepplingerfix linux-image-librem5 binary packagethere is a bug with how linux-image-librem5 is created. It's created/installed just like the actual linux-image-6.3.0-1-librem5, but it should be an exception: It should contain nothing and exists only do have the "Depends" in d/control....there is a bug with how linux-image-librem5 is created. It's created/installed just like the actual linux-image-6.3.0-1-librem5, but it should be an exception: It should contain nothing and exists only do have the "Depends" in d/control. deb-packaging scripts completely changed from v6.2..v6.3. Where should we best tweak them (probably scripts/package/builddeb ) to have an empty linux-image-librem5 again?https://source.puri.sm/Librem5/linux/-/issues/488edt_ft5x06: Reading LIC/FW version is broken on 6.42023-10-02T08:36:32ZSebastian Krzyszkowiakedt_ft5x06: Reading LIC/FW version is broken on 6.4On 6.3:
```
purism@evergreen:~$ sudo cat /sys/kernel/debug/edt_ft5x06/lic_version /sys/kernel/debug/edt_ft5x06/fw_version
7
2
```
On 6.4:
```
purism@evergreen:~$ sudo cat /sys/kernel/debug/edt_ft5x06/lic_version /sys/kernel/debug/edt_...On 6.3:
```
purism@evergreen:~$ sudo cat /sys/kernel/debug/edt_ft5x06/lic_version /sys/kernel/debug/edt_ft5x06/fw_version
7
2
```
On 6.4:
```
purism@evergreen:~$ sudo cat /sys/kernel/debug/edt_ft5x06/lic_version /sys/kernel/debug/edt_ft5x06/fw_version
18446603336221196295
18446603336221196290
```