Bug 500203

I can confirm this issue on Arch Linux.

I cannot reproduce, can you share the output of 'env' in a command line

(In reply to David Edmundson from comment #2)
> I cannot reproduce, can you share the output of 'env' in a command line

BROWSER=firefox
DBUS_SESSION_BUS_ADDRESS=unix:path=/run/user/1000/bus
DISPLAY=:0
EDITOR=nvim
ELECTRON_OZONE_PLATFORM_HINT=wayland
ELECTRON_TRASH=kioclient
GTK2_RC_FILES=/home/pibeng/.config/gtk-2.0/gtkrc
GTK_RC_FILES=/etc/gtk/gtkrc:/home/pibeng/.gtkrc:/home/pibeng/.config/gtkrc
GTK_USE_PORTAL=1
KDE_APPLICATIONS_AS_SCOPE=1
KDE_FULL_SESSION=true
KDE_SESSION_UID=1000
KDE_SESSION_VERSION=6
KGLOBALACCELD_PLATFORM=org.kde.kwin
LANG=en_US.UTF-8
LANGUAGE=en_US
LC_ADDRESS=en_US.UTF-8
LC_IDENTIFICATION=en_US.UTF-8
LC_MEASUREMENT=en_US.UTF-8
LC_MONETARY=en_US.UTF-8
LC_NAME=en_US.UTF-8
LC_NUMERIC=en_US.UTF-8
LC_PAPER=en_US.UTF-8
LC_TELEPHONE=en_US.UTF-8
LC_TIME=en_US.UTF-8
PATH=/home/pibeng/.local/bin:/usr/local/sbin:/usr/local/bin:/usr/local/sbin:/usr/local/bin:/usr/bin:/var/lib/flatpak/exports/bin:/usr/lib/jvm/default/bin:/usr/bin/site_perl:/usr/bin/vendor_perl:/usr/bin/core_perl
QT_AUTO_SCREEN_SCALE_FACTOR=0
QT_FONT_DPI=96
QT_WAYLAND_RECONNECT=1
SESSION_MANAGER=local/pbngtower1:@/tmp/.ICE-unix/1266,unix/pbngtower1:/tmp/.ICE-unix/1266
USER=pibeng
WAYLAND_DISPLAY=wayland-0
XDG_CURRENT_DESKTOP=KDE
XDG_MENU_PREFIX=plasma-
XDG_RUNTIME_DIR=/run/user/1000
XDG_SEAT=seat0
XDG_SESSION_TYPE=wayland
XDG_VTNR=1
XKB_DEFAULT_LAYOUT=us

Changing the status back to REPORTED because I already posted my `env` output and it's been 10 days since the last comment.

Cannot reproduce on X11

> text/plain;charset=ANSI_X3.4-1968

This might be the culprit, but my clip doesn't have it.

A possibly relevant merge request was started @ https://invent.kde.org/plasma/plasma-workspace/-/merge_requests/5266

Git commit 709f0738216fee7523f09f1169965c3a30b32095 by Fushan Wen.
Committed on 28/02/2025 at 00:12.
Pushed by fusionfuture into branch 'master'.

appiumtests: test copying UTF-8 string

M  +6    -4    appiumtests/applets/clipboardtest.py

https://invent.kde.org/plasma/plasma-workspace/-/commit/709f0738216fee7523f09f1169965c3a30b32095

A possibly relevant merge request was started @ https://invent.kde.org/plasma/plasma-workspace/-/merge_requests/5268

Git commit 183764fdc34b1438d632255d5b655e12ab38570e by Fushan Wen.
Committed on 28/02/2025 at 11:23.
Pushed by fusionfuture into branch 'master'.

klipper: ignore non-UTF-8 encoded plain text

Qt clipboard doesn't support other encodings.
FIXED-IN: 6.3.3

M  +7    -4    appiumtests/applets/clipboardtest.py
M  +7    -0    klipper/updateclipboardjob.cpp
M  +1    -0    klipper/updateclipboardjob.h

https://invent.kde.org/plasma/plasma-workspace/-/commit/183764fdc34b1438d632255d5b655e12ab38570e

Git commit d79fbbb31decb97e985e0a89718fc05605f9e4de by Fushan Wen.
Committed on 28/02/2025 at 12:22.
Pushed by fusionfuture into branch 'Plasma/6.3'.

klipper: ignore non-UTF-8 encoded plain text

Qt clipboard doesn't support other encodings.
FIXED-IN: 6.3.3


(cherry picked from commit 183764fdc34b1438d632255d5b655e12ab38570e)

Co-authored-by: Fushan Wen <qydwhotmail@gmail.com>

M  +7    -4    appiumtests/applets/clipboardtest.py
M  +7    -0    klipper/updateclipboardjob.cpp
M  +1    -0    klipper/updateclipboardjob.h

https://invent.kde.org/plasma/plasma-workspace/-/commit/d79fbbb31decb97e985e0a89718fc05605f9e4de

I updated to Plasma 6.3.3 on Arch Linux and sadly this bug can still be reproduced in the same way.

For instance, if I copy the text 狐狸 in Firefox, there's a new row in the main table with the correct mimetype:

INSERT INTO main VALUES('7f1efc3b619b12d38164b3610171e17bf13680a2',1741727038.727999926,1741727038.727999926,'text/plain;charset=utf-8,UTF8_STRING,COMPOUND_TEXT,TEXT,text/plain,STRING','狐狸',NULL);

However, its corresponding rows in the aux table look like this:

INSERT INTO aux VALUES('7f1efc3b619b12d38164b3610171e17bf13680a2','text/plain;charset=utf-8','7f1efc3b619b12d38164b3610171e17bf13680a2');
INSERT INTO aux VALUES('7f1efc3b619b12d38164b3610171e17bf13680a2','text/plain','ce2b5da05d513aed6cf70611ecc367cd333f69ef');

Indeed, if I run `cat .local/share/klipper/data/7f1efc3b619b12d38164b3610171e17bf13680a2/7f1efc3b619b12d38164b3610171e17bf13680a2 ` I get 狐狸 and `cat .local/share/klipper/data/7f1efc3b619b12d38164b3610171e17bf13680a2/ce2b5da05d513aed6cf70611ecc367cd333f69ef` prints \u72d0\u72f8, which I'm guessing is the source of this bug.

However, if I delete 狐狸 from my clipboard history and repeat this experiment inside KWrite, here are the rows that are added:

INSERT INTO main VALUES('7f1efc3b619b12d38164b3610171e17bf13680a2',1741727174.533999919,1741727174.533999919,'text/plain,text/plain;charset=utf-8','狐狸',NULL);
INSERT INTO aux VALUES('7f1efc3b619b12d38164b3610171e17bf13680a2','text/plain','7f1efc3b619b12d38164b3610171e17bf13680a2');
INSERT INTO aux VALUES('7f1efc3b619b12d38164b3610171e17bf13680a2','text/plain;charset=utf-8','7f1efc3b619b12d38164b3610171e17bf13680a2');

In this case, both aux rows point to the same file which contains 狐狸, not \u72d0\u72f8.

Can confirm that the patch mentioned in comment 11 did not fix the issue. Copying "а ạ ą ä à á ą" above and following the steps to reproduce still pastes as "\u0430 \u1ea1 \u0105 � � � \u0105" in Kate and the text doesn't paste at all on Firefox

I took a glance at https://invent.kde.org/plasma/plasma-workspace/-/blob/master/klipper/updateclipboardjob.cpp which was edited in the patch and I'm stumped on why copying Unicode text from Firefox would lead to different results than in KWrite or other apps, so I'm guessing maybe Firefox is the one that's converting Unicode text into \uXXXX form for the text/plain mimetype.

Two possible solutions could be making Klipper only save the text/plain;charset=utf-8 version of the text if both that and text/plain are given to Klipper by an app, or making Klipper prioritize text/plain;charset=utf-8 over text/plain when selecting an item from clipboard history (but I couldn't find the code that handles this). However, I don't know much about the Klipper codebase or what consequences either of these fixes would have.

(In reply to Anthony Wang from comment #14)
> I'm guessing maybe Firefox is the one that's converting Unicode text into \uXXXX form for the text/plain mimetype.

The text appears with the expected characters on the Plasma clipboard history popup and in the corresponding database file if you were to inspect it with a SQLite browser or CLI utility, so I believe the problem is on Klipper's end.

*** Bug 501130 has been marked as a duplicate of this bug. ***

A possibly relevant merge request was started @ https://invent.kde.org/plasma/plasma-workspace/-/merge_requests/5325

(In reply to Timothy B from comment #0)
> Sometimes, if you paste the history item to a textbox in Firefox, it would
> output nothing.

I noticed this too on Fedora 41 (Plasma 6.3.3, Frameworks 6.12.0, Qt 6.8.2 on Wayland). Nothing happens when I paste the second item in Klipper into either Firefox's Location/URL field or an HTML textarea; but I see the garbled text if I paste it into Firefox's "Save As..." dialog. I've never seen Firefox decline to paste text clipboard contents.

Strangely, the Klipper text pastes fine into a Thunderbird e-mail composition window, but not Thunderbird's message search input field or "Search Messages" dialog.

Git commit 9c014bad595b743de4ca7b236c4d467356505990 by Fushan Wen.
Committed on 22/03/2025 at 10:14.
Pushed by fusionfuture into branch 'master'.

klipper: let Qt handle plain text to deal with non-UTF-8 encodings

Data from text/plain might use a weird encoding on some systems, so
don't manually deal with plain text from raw bytes.
FIXED-IN: 6.3.4

M  +4    -4    klipper/autotests/v3migrationtest.py
M  +41   -36   klipper/updateclipboardjob.cpp
M  +2    -1    klipper/updateclipboardjob.h

https://invent.kde.org/plasma/plasma-workspace/-/commit/9c014bad595b743de4ca7b236c4d467356505990

Git commit cd7d0215a7a0bc3503640f0f8c4e44bde08f7b20 by Fushan Wen.
Committed on 22/03/2025 at 10:49.
Pushed by fusionfuture into branch 'Plasma/6.3'.

klipper: let Qt handle plain text to deal with non-UTF-8 encodings

Data from text/plain might use a weird encoding on some systems, so
don't manually deal with plain text from raw bytes.
FIXED-IN: 6.3.4


(cherry picked from commit 9c014bad595b743de4ca7b236c4d467356505990)

Co-authored-by: Fushan Wen <qydwhotmail@gmail.com>

M  +4    -4    klipper/autotests/v3migrationtest.py
M  +41   -36   klipper/updateclipboardjob.cpp
M  +2    -1    klipper/updateclipboardjob.h

https://invent.kde.org/plasma/plasma-workspace/-/commit/cd7d0215a7a0bc3503640f0f8c4e44bde08f7b20

This does not correct the situation in case of synchronization of selection with the clipboard and display of Cyrillic.
If the "Keep the selection and clipboard the same" option is enabled, and the value "Text selection: Always save in history" is selected,
then in this case, selecting text in Cyrillic in a web browser, when you subsequently paste text from the clipboard, it looks like \u0441.
Now I have highlighted the text in Cyrillic in Firefox: "Вырабатывающая и передающая", as a result, I then pasted the following text from the clipboard:
"\u0412\u044b\u0440\u0430\u0431\u0430\u0442\u044b\u0432\u0430\u044e\u0449\u0430\u044f \u0438 \u043f\u0435\u0440\u0435\u0434\u0430\u044e\u0449\u0430\u044f"

(In reply to Victor Ryzhykh from comment #21)
> This does not correct the situation in case of synchronization of selection
> with the clipboard and display of Cyrillic.
> If the "Keep the selection and clipboard the same" option is enabled, and
> the value "Text selection: Always save in history" is selected,
> then in this case, selecting text in Cyrillic in a web browser, when you
> subsequently paste text from the clipboard, it looks like \u0441.
> Now I have highlighted the text in Cyrillic in Firefox: "Вырабатывающая и
> передающая", as a result, I then pasted the following text from the
> clipboard:
> "\u0412\u044b\u0440\u0430\u0431\u0430\u0442\u044b\u0432\u0430\u044e\u0449\u04
> 30\u044f \u0438 \u043f\u0435\u0440\u0435\u0434\u0430\u044e\u0449\u0430\u044f"

It should be the same symptom, but the code is in another place.

A possibly relevant merge request was started @ https://invent.kde.org/plasma/plasma-workspace/-/merge_requests/5340

(In reply to Bug Janitor Service from comment #23)
> A possibly relevant merge request was started @
> https://invent.kde.org/plasma/plasma-workspace/-/merge_requests/5340

This patch fixed the bug. Thank you.

Git commit a4b5f1d4e9878ab425a1273d6bfb48252ecedf0b by Fushan Wen.
Committed on 26/03/2025 at 04:14.
Pushed by fusionfuture into branch 'master'.

klipper: let Qt handle plain text to deal with non-UTF-8 encodings when syncing selections
FIXED-IN: 6.3.4

M  +11   -6    klipper/systemclipboard.cpp

https://invent.kde.org/plasma/plasma-workspace/-/commit/a4b5f1d4e9878ab425a1273d6bfb48252ecedf0b

Git commit 8d81c6816e64a9b9df7a093cc50f093dc258ff7a by Fushan Wen.
Committed on 26/03/2025 at 12:57.
Pushed by fusionfuture into branch 'Plasma/6.3'.

klipper: let Qt handle plain text to deal with non-UTF-8 encodings when syncing selections
FIXED-IN: 6.3.4


(cherry picked from commit a4b5f1d4e9878ab425a1273d6bfb48252ecedf0b)

Co-authored-by: Fushan Wen <qydwhotmail@gmail.com>

M  +11   -6    klipper/systemclipboard.cpp

https://invent.kde.org/plasma/plasma-workspace/-/commit/8d81c6816e64a9b9df7a093cc50f093dc258ff7a

I compiled Klipper from the Plasma/6.3 branch and can confirm this bug has finally been fixed, unlike the first PR attempt a few weeks ago. Thanks Fushan Wen!

*** Bug 503429 has been marked as a duplicate of this bug. ***