In the meantime I am shure that my trouble is not a problem of seafile. I installed successful an alder version so now i can work.
But after the last update I did not start virtualbox, but some minutes ago I tried but it did not work. The information about the problem is almost the same as trying the newest version of seafile:
"virtualbox
VirtualBox: supR3HardenedMainGetTrustedMain: dlopen("/usr/lib/virtualbox/VirtualBox.so",) failed: /usr/lib/libharfbuzz.so.0: undefined symbol: FT_Get_Var_Blend_Coordinates"
Search Criteria
Package Details: seafile-client 9.0.13-2
Package Actions
Git Clone URL: | https://aur.archlinux.org/seafile-client.git (read-only, click to copy) |
---|---|
Package Base: | seafile-client |
Description: | GUI client for synchronizing your local files with seafile server |
Upstream URL: | https://github.com/haiwen/seafile-client |
Licenses: | Apache |
Submitter: | Localizator |
Maintainer: | Joffrey |
Last Packager: | Joffrey |
Votes: | 168 |
Popularity: | 0.000000 |
First Submitted: | 2012-12-10 17:34 (UTC) |
Last Updated: | 2025-04-05 18:40 (UTC) |
Dependencies (7)
- qt6-5compat
- qt6-base (qt6-base-gitAUR, qt6-base-headlessAUR)
- qt6-webengine
- seafileAUR
- cmake (cmake-gitAUR, cmake3AUR) (make)
- qt6-tools (make)
- gtk-update-icon-cache (gtk-update-icon-cache-gitAUR) (optional)
Required by (0)
Sources (2)
Latest Comments
« First ‹ Previous 1 .. 22 23 24 25 26 27 28 29 30 31 32 .. 44 Next › Last »
a-bostaurus commented on 2017-01-19 09:54 (UTC)
eolianoe commented on 2017-01-18 21:22 (UTC)
@a-bostaurus: you cannot rename libharfbuzz as seafile depends on it!
Try to find the special configurations you made and if you cannot find it, build all seafile packages in a clean chroot [a].
As a side note, please post build logs in english in order to make the debug easier. ('export LANG=C' before the build should translate it)
[a] https://wiki.archlinux.org/index.php/DeveloperWiki:Building_in_a_Clean_Chroot
a-bostaurus commented on 2017-01-18 20:00 (UTC)
Sorry, that did not help. So I renamed libharfbuzz... and that mistake did not appear again. But the problem with seafile-applet still exist:
"/usr/lib/libQt5Gui.so.5.7.1: undefined reference to `hb_unicode_funcs_set_compose_func'
/usr/lib/libQt5Gui.so.5.7.1: undefined reference to `hb_font_funcs_destroy'
/usr/lib/libQt5WebEngineCore.so.5.7.1: undefined reference to `hb_ot_layout_has_positioning'
/usr/lib/libQt5Gui.so.5.7.1: undefined reference to `hb_unicode_funcs_set_script_func'
collect2: Fehler: ld gab 1 als Ende-Status zurück
make[2]: *** [CMakeFiles/seafile-applet.dir/build.make:5385: seafile-applet] Fehler 1
make[1]: *** [CMakeFiles/Makefile2:997: CMakeFiles/seafile-applet.dir/all] Fehler 2
make: *** [Makefile:128: all] Fehler 2
==> FEHLER: Ein Fehler geschah in build().
Breche ab...
==> FEHLER:Makepkg konnte seafile-client nicht erstellen."
Of course I rename the libharfbuzz...-file again.
eolianoe commented on 2017-01-18 14:28 (UTC)
@a-bostaurus: try to remove all seafile packages, then reinstall harfbuzz and all seafile packages.
If it's working on your five others computers, there is a special configuration on the sixth which causes trouble.
a-bostaurus commented on 2017-01-18 09:57 (UTC) (edited on 2017-01-18 13:23 (UTC) by a-bostaurus)
There is harfbuzz on my computer, but no infinality ...
"extra/harfbuzz 1.4.1-1 [installed]
OpenType text shaping engine
extra/harfbuzz-icu 1.4.1-1 [installed]
OpenType text shaping engine (ICU integration)
multilib/lib32-harfbuzz 1.4.1-1 [installed]
OpenType text shaping engine (32-bit)"
Should I deinstall harfbuzz?
Some hours later: I had a look at the dependicies of harfbuzz. I fear deinstalling harfbuzz will mean to make my system unusabel. I.e. chromium needs it and so on. Because of i deinstalled seafile-client to get a clean new installation I just now cannot install seafile, although I need it strongly for my early work. Very annoying.
eolianoe commented on 2017-01-18 09:33 (UTC)
@snack: I think that they just regenerate the sources from the git tree without recreating a proper tag. It's a bit annoying and I think that I will use the commit hash to get the sources rather than the archive.
@a-bostaurus: This may me linked to some harfbuzz and Infinality fonts bugs. DO you use them ?
Distorted commented on 2017-01-18 09:32 (UTC)
@a-bostaurus
are you using infinality?
snack commented on 2017-01-18 09:26 (UTC)
I had similar problems with checksums this morning. When upgrading with yaourt I got this message:
==> Validating source files with sha256sums...
seafile-6.0.2.tar.gz ... FAILED
The cached version of seafile-6.0.2.tar.gz on my disk retrieved from previous updates did not pass the checksum of the current PKGBUILD, but evidently passed that of old PKGBUILD since I had no problems in previous updates.
So I deleted the file and started the update again, to force a fresh download, and this time everything worked.
So I'd say that seafile-6.0.2.tar.gz has changed upstream without changing the filename. Sounds a bit strange (to not say suspicious), but maybe there's an explanation for this?
a-bostaurus commented on 2017-01-18 09:23 (UTC)
I'm working with six arch computers. Everyone works with seafile. Whenever I try to update seafile-client there is some trouble. And never the same ... This time it is:
"/usr/lib/libharfbuzz.so.0: undefined reference to `FT_Get_Var_Blend_Coordinates'
collect2: Fehler: ld gab 1 als Ende-Status zurück
make[2]: *** [CMakeFiles/seafile-applet.dir/build.make:5385: seafile-applet] Fehler 1
make[1]: *** [CMakeFiles/Makefile2:997: CMakeFiles/seafile-applet.dir/all] Fehler 2
make: *** [Makefile:128: all] Fehler 2
==> FEHLER: Ein Fehler geschah in build().
Breche ab...
==> FEHLER:Makepkg konnte nicht erstellen." Deinstalling seafile-client with all its dependies and installing again did not help ...
MrKMG commented on 2017-01-17 15:52 (UTC)
SHA256SUM appears to be different:
8ac2eb3611e69b4fd85a9572714153566705c73e443cf2ae24da556c7cf12078 seafile-6.0.2.tar.gz
Pinned Comments
Joffrey commented on 2021-05-30 20:06 (UTC) (edited on 2021-05-30 20:11 (UTC) by Joffrey)
Please, when you have compilation or execution errors, recompile each component without using an AUR helper before reporting an issue.