| Git Clone URL: | https://aur.archlinux.org/librewolf-bin.git (read-only, click to copy) |
|---|---|
| Package Base: | librewolf-bin |
| Description: | Community-maintained fork of Firefox, focused on privacy, security and freedom. |
| Upstream URL: | https://librewolf-community.gitlab.io/ |
| Keywords: | browser web |
| Licenses: | GPL, MPL, LGPL |
| Conflicts: | librewolf |
| Provides: | librewolf |
| Submitter: | lsf |
| Maintainer: | lsf |
| Last Packager: | lsf |
| Votes: | 465 |
| Popularity: | 34.57 |
| First Submitted: | 2019-06-16 13:12 (UTC) |
| Last Updated: | 2025-03-27 22:43 (UTC) |
libxul.so is corrupted in librewolf-bin 130.0.1-1
librewolf
XPCOMGlueLoad error for file /usr/lib/librewolf/libxul.so:
/usr/lib/librewolf/libxul.so: file too short
Couldn't load XPCOM.
size is 0 bytes
ls -l /usr/lib/librewolf/libxul.so
-rwxr-xr-x 1 root root 0 Sep 23 13:14 /usr/lib/librewolf/libxul.so
Today was my third attempt at installing the current version (130) of LibreWolf to my up-to-date Manjaro KDE system. Two issues:
Has anyone here experienced the same issue?
EDIT 2024-09-15: Don’t bother, I installed the Flatpak version.
For quite some time now, the -bin version of the package has not been based on the Arch specific (source) PKGBUILD anymore, and thus does not include profiling anymore, for example, and probably some performance related build options.
So: It might be worth checking out the upstream repo(s) https://codeberg.org/librewolf/source and potentially open a PR (or an issue at https://codeberg.org/librewolf/issues/issues) regarding potential improvements on that front. Thanks!
@blutuna: Your results are even worse than mine: 30 % performance drop for LW compared to FF. Although they share the same source code. That is odd.
@karoyli... I get the following on my 5950, no tweaks to the installs... FF = 21.1 LW = 15.7 Floorp = 13.6 SRWare Iron = 20.6 Mullvad = 8.64
May be librewolf disable some runtime optimization for privacy reasons
Nice joke. But still, from my point of view librewolf should at least reach the same performance ratings as firefox.
@mabod, if that matters: I get a value of Infinity (no kidding) with a self-compiled librewolf (only my own patch is applied extra, it's why I compile manually), and 21.4 on FF 128.0 on my 5950.
I guess my librewolf surpassed lightspeed? :)
Something is wrong with the performance of this binary on my Ryzen 9 5900X. When I test performance with https://browserbench.org/Speedometer3.0/ I get a score of only 11,9. Firefox with the same version 128.0.3 reaches a score of 13,7, which is 15 % better. Since both are using the same source code I would assume librewolf-bin is using different compiler options. When I compile librewolf by myself with
CFLAGS="-march=x86-64-v3 -O2 -pipe -fno-plt -fexceptions \
-Wp,-D_FORTIFY_SOURCE=2 -Wformat -Werror=format-security \
-fstack-clash-protection -fcf-protection"
I get a score of 15,9, which is excellent. Only exceeded by Chromium with 17,9. So something seems to be odd with the compiler options used to build librewolf-bin.
Pinned Comments
lsf commented on 2021-11-10 12:14 (UTC) (edited on 2023-04-17 07:18 (UTC) by lsf)
https://wiki.archlinux.org/title/Arch_User_Repository#Acquire_a_PGP_public_key_if_needed
gpg --keyserver hkp://keyserver.ubuntu.com --search-keys 031F7104E932F7BD7416E7F6D2845E1305D6E801/edit: starting with 112.0-1, the binaries are signed with the maintainers shared key, so
gpg --keyserver hkp://keyserver.ubuntu.com --search-keys 662E3CDD6FE329002D0CA5BB40339DD82B12EF16should do the trick instead. I've also signed the key with the previously used key, so you have at least some guarantee that it's not a malicious attack :)