Package Details: webkit2gtk 2.50.6-7

Git Clone URL: https://aur.archlinux.org/webkit2gtk.git (read-only, click to copy)
Package Base: webkit2gtk
Description: Web content engine for GTK
Upstream URL: https://webkitgtk.org
Licenses: MIT, ISC, OFL-1.1, ICU, Apache-2.0, NCSA, BSD-3-Clause, BSD-2-Clause, BSL-1.0, MPL-1.1, MPL-2.0, GPL-2.0-only, LGPL-2.1-or-later, LGPL-2.1-only, BSD-2-Clause-Views, Apache-2.0 WITH LLVM-exception, AFL-2.0 OR GPL-2.0-or-later, BSD-Source-Code, bzip2-1.0.6, GPL-3.0-only WITH Autoconf-exception-3.0, GPL-3.0-or-later WITH Bison-exception-2.2, NCSA OR MIT, SunPro, Unicode-TOU
Provides: libjavascriptcoregtk-4.0.so, libwebkit2gtk-4.0.so
Submitter: andreas_baumann
Maintainer: andreas_baumann
Last Packager: andreas_baumann
Votes: 19
Popularity: 11.36
First Submitted: 2026-04-10 12:35 (UTC)
Last Updated: 2026-04-28 16:18 (UTC)

Sources (2)

Latest Comments

« First ‹ Previous 1 2 3 4 5 6 7 Next › Last »

moormaster commented on 2026-04-28 16:05 (UTC)

@andreas_baumann, did you forget to do a "git push"? :)

FabioLolix commented on 2026-04-28 15:46 (UTC)

I have builded the package yesterday and is here https://sourceforge.net/projects/fabiololix-os-archive/files/Packages/ but that's not a repo

Unfortunately it's listed as a dependency for BambuStudio

Lancaban on bambustudio-appimage page comments it is reported that it work with webkit2gtk-4.1

Lancaban commented on 2026-04-28 15:18 (UTC)

Still unable to install this without running out of memory and system crashing, I've got 32 Gb and made sure to close things beforehand.

Unfortunately it's listed as a dependency for BambuStudio. (not sure why, that's on them not you) I'm open to suggestions.

andreas_baumann commented on 2026-04-28 13:13 (UTC) (edited on 2026-04-28 13:13 (UTC) by andreas_baumann)

@moormaster: thanks a bunch for the patch, added. :-)

moormaster commented on 2026-04-27 16:55 (UTC) (edited on 2026-04-27 16:57 (UTC) by moormaster)

I have improved Maravedis suggestions to match more cases of how "number of parallel jobs" can be extracted from MAKEFLAGS - combined with a fallback to nproc / 2 if MAKEFLAGS does not contain a -j argument:

Proof/tests that regexp works in the five cases -jN, -j N, -j=N, --jobs N, and --jobs=N:

$ { echo '-j2'; echo '-j 3'; echo '-j=4'; echo '--jobs 5'; echo '--jobs=6'; } | sed 's/.*\(-j[ =]\?\|--jobs[ =]\)\([0-9]\+\).*/\2/'
2
3
4
5
6

Suggested lines to insert into the PKGBUILD:

# extract N from MAKEFLAGS arguments of forms
#   -jN or -j N or -j=N 
#   or --jobs N or --jobs=N
export CMAKE_BUILD_PARALLEL_LEVEL=$(echo "$MAKEFLAGS" | sed 's/.*\(-j[ =]\?\|--jobs[ =]\)\([0-9]\+\).*/\2/' )

# If CMAKE_BUILD_PARALLEL_LEVEL is still empty assign half of nproc as default value
export CMAKE_BUILD_PARALLEL_LEVEL=${CMAKE_BUILD_PARALLEL_LEVEL:=$(( $(nproc) / 2 ))}

jghodd commented on 2026-04-27 13:43 (UTC)

@andreas_baumann - thank you. that fixed it.

Maravedis commented on 2026-04-27 06:28 (UTC)

I'm all for extracting the parallel number from makeflags, but it is missing a couple of cases.

  export CMAKE_BUILD_PARALLEL_LEVEL=$(echo "$MAKEFLAGS" | sed 's/.*-j\([0-9]\+\).*/\1/')

This breaks on --jobs=8, for example.

andreas_baumann commented on 2026-04-27 05:14 (UTC) (edited on 2026-04-27 05:38 (UTC) by andreas_baumann)

@jghodd: yes, apparently nobody tests the make build path anymore, so back to ninja (I guess it's a timestamp issue around GIR generation).

Also, extracting the N from -j<N> in the MAKFLAGS and setting them as CMAKE_BUILD_PARALLEL_LEVEL seems to be a good solution for respecting the parallelity of the build.

jghodd commented on 2026-04-27 01:34 (UTC)

Getting the following build error:


[ 99%] Built target WebKit
make[2]: *** No rule to make target 'JavaScriptCore-4.0.gir', needed by 'WebKit2WebExtension-4.0.gir'.  Stop.
make[1]: *** [CMakeFiles/Makefile2:2490: Source/WebKit/CMakeFiles/typelib-WebKit2WebExtension.dir/all] Error 2
make: *** [Makefile:156: all] Error 2
==> ERROR: A failure occurred in build().
    Aborting...

un_pogaz commented on 2026-04-26 21:49 (UTC) (edited on 2026-04-27 04:51 (UTC) by un_pogaz)

So here a proposed patch to have a better compromise.

This patch reenable the Ninja generator to take advantage of faster compilation speeds, BUT dynamically limit the ressources used by the compilation to half of the number of processors available on the user's computer.

If some user continues to encounter compilation issues, it will be easy to adjust by modifying the calculation operations.

EDIT: this should also fix error reported by @guglovich and @jghodd

   # https://gitlab.archlinux.org/archlinux/packaging/packages/webkitgtk-6.0/-/issues/4
   export CC=clang CXX=clang++
   LDFLAGS+=" -fuse-ld=lld"

+  # Limit the number of processors used by Ninja,
+  # else the compilation can use too much ressources
+  # and crashe the machine
+  TOTAL_PRC=$(nproc --all)     # total number of processors
+  USED_PRC=$(($TOTAL_PRC / 2)) # use half of them
+  if [ $USED_PRC -eq 0 -o $USED_PRC -lt 0 ]; then
+    USED_PRC=1  # ensure at least one
+  fi
+  export CMAKE_BUILD_PARALLEL_LEVEL=$USED_PRC
+
   # Skia uses malloc_usable_size
   CFLAGS="${CFLAGS/_FORTIFY_SOURCE=3/_FORTIFY_SOURCE=2}"
   CXXFLAGS="${CXXFLAGS/_FORTIFY_SOURCE=3/_FORTIFY_SOURCE=2}"

   # JITted code crashes when CET is used
   CFLAGS+=' -fcf-protection=none'
   CXXFLAGS+=' -fcf-protection=none'

-  cmake -S webkitgtk-$pkgver -B build "${cmake_options[@]}"
+  cmake -S webkitgtk-$pkgver -B build -G Ninja "${cmake_options[@]}"
   cmake --build build
 }

 package_webkit2gtk() {

« First ‹ Previous 1 2 3 4 5 6 7 Next › Last »