Ping. On Tue Aug 25, 2026 at 7:32 PM CEST, Igor Zornik wrote: > Hello, ports. > > A new version of esbuild with a few stylistic changes to the Makefile. > JS and CSS bundling tested on today's amd64 snap. QP-encoded diff below > the changelog. Someone please commit if OK. > > Out of curiosity I ran make test and got the following error: > > no required module provides package github.com/evanw/esbuild; to add it: > go get github.com/evanw/esbuild > FAIL github.com/evanw/esbuild [setup failed] > > Does this make it a candidate for NO_TEST? > > Changelog: > - Fix tree shaking bug due to TypeScript import alias. > - Fix CSS minification bug involving '&'. > - Avoid overwriting input files without '--allow-overwrite'. > - Fix incorrect code generated when using top-level await. > - Fix a minification bug with lowered logical assignment operators. > - Fix a potential deadlock when the JavaScript API is used incorrectly. > - Handle target collisions. > - Force '.mp3' files to use the 'audio/mpeg' MIME type. > - Add a new TypeScript syntax warning. > - Add support for formatting errors for Visual Studio. > - Fix a bug with CSS gamut mapping. > > Detailed changelog is available at > https://github.com/evanw/esbuild/releases/tag/v0.28.2 > Index: Makefile =================================================================== RCS file: /cvs/ports/devel/esbuild/Makefile,v retrieving revision 1.21 diff -u -p -u -p -r1.21 Makefile --- Makefile 30 Jun 2026 17:20:17 -0000 1.21 +++ Makefile 1 Sep 2026 15:18:33 -0000 @@ -1,19 +1,20 @@ COMMENT = fast bundler for the web MODGO_MODNAME = github.com/evanw/esbuild -MODGO_VERSION = v0.28.1 +MODGO_VERSION = v0.28.2 DISTNAME = esbuild-${MODGO_VERSION} -CATEGORIES = devel -MAINTAINER = Igor Zornik <mocheryl@mocheryl.org> +CATEGORIES = devel HOMEPAGE = https://esbuild.github.io/ +MAINTAINER = Igor Zornik <mocheryl@mocheryl.org> + # MIT PERMIT_PACKAGE = Yes -WANTLIB += c pthread +WANTLIB += c pthread MODULES = lang/go Index: distinfo =================================================================== RCS file: /cvs/ports/devel/esbuild/distinfo,v retrieving revision 1.20 diff -u -p -u -p -r1.20 distinfo --- distinfo 30 Jun 2026 17:20:17 -0000 1.20 +++ distinfo 1 Sep 2026 15:18:33 -0000 @@ -1,6 +1,6 @@ -SHA256 (esbuild-v0.28.1.zip) = zoWIu+ztbLJ6wybOMGFDfkd17rDo9aU64Pjcvid8Clw= +SHA256 (esbuild-v0.28.2.zip) = 2w/CkY730XabF0btEj1A20+SY24lNt2k5A9dCArjDeI= SHA256 (go_modules/golang.org/x/sys/@v/v0.0.0-20220715151400-c0bba94af5f8.mod) = 8DMzMJb+GY8xUd7tk/LeunTlC7/nc5E0BFvDt85KUCQ= SHA256 (go_modules/golang.org/x/sys/@v/v0.0.0-20220715151400-c0bba94af5f8.zip) = OxgJNyFuk1WfFrYHbQm69UpXBzePEbhntuuRTFawm5E= -SIZE (esbuild-v0.28.1.zip) = 2155515 +SIZE (esbuild-v0.28.2.zip) = 2242951 SIZE (go_modules/golang.org/x/sys/@v/v0.0.0-20220715151400-c0bba94af5f8.mod) = 33 SIZE (go_modules/golang.org/x/sys/@v/v0.0.0-20220715151400-c0bba94af5f8.zip) = 1798284
Tuesday, September 01, 2026
mips64 bulk build report
bulk build on octeon.ports.openbsd.org started on Mon Aug 24 15:32:23 UTC 2026 finished at Tue Sep 1 11:50:00 UTC 2026 lasted 08D20h17m done with kern.version=OpenBSD 8.0-beta (GENERIC.MP) #74: Sun Aug 23 13:29:32 MDT 2026 built packages:9041 Aug 24:1939 Aug 25:1374 Aug 26:1718 Aug 27:1378 Aug 28:659 Aug 29:617 Aug 30:614 Aug 31:739 Sep 1:2 build failures: 87 http://build-failures.rhaalovely.net/mips64/2026-08-24/archivers/blosc2.log http://build-failures.rhaalovely.net/mips64/2026-08-24/astro/wmglobe.log http://build-failures.rhaalovely.net/mips64/2026-08-24/audio/gogglesmm.log http://build-failures.rhaalovely.net/mips64/2026-08-24/cad/coil64.log http://build-failures.rhaalovely.net/mips64/2026-08-24/cad/netgen-mesher.log http://build-failures.rhaalovely.net/mips64/2026-08-24/cad/ngspice/ngspice.log http://build-failures.rhaalovely.net/mips64/2026-08-24/chinese/libpinyin.log http://build-failures.rhaalovely.net/mips64/2026-08-24/comms/liquid-dsp.log http://build-failures.rhaalovely.net/mips64/2026-08-24/comms/qrq.log http://build-failures.rhaalovely.net/mips64/2026-08-24/comms/sdrpp.log http://build-failures.rhaalovely.net/mips64/2026-08-24/converters/k2pdfopt.log http://build-failures.rhaalovely.net/mips64/2026-08-24/databases/duckdb.log http://build-failures.rhaalovely.net/mips64/2026-08-24/databases/postgresql-pllua.log http://build-failures.rhaalovely.net/mips64/2026-08-24/databases/sqlite3-odbc.log http://build-failures.rhaalovely.net/mips64/2026-08-24/devel/abseil-cpp.log http://build-failures.rhaalovely.net/mips64/2026-08-24/devel/arm-none-eabi/gcc,aarch64.log http://build-failures.rhaalovely.net/mips64/2026-08-24/devel/boost.log http://build-failures.rhaalovely.net/mips64/2026-08-24/devel/cbmc.log http://build-failures.rhaalovely.net/mips64/2026-08-24/devel/clang-tools-extra.log http://build-failures.rhaalovely.net/mips64/2026-08-24/devel/cvsplot.log http://build-failures.rhaalovely.net/mips64/2026-08-24/devel/mips-elf/gcc.log http://build-failures.rhaalovely.net/mips64/2026-08-24/devel/py-greenlet.log http://build-failures.rhaalovely.net/mips64/2026-08-24/devel/py-llvmlite.log http://build-failures.rhaalovely.net/mips64/2026-08-24/devel/py-thrift.log http://build-failures.rhaalovely.net/mips64/2026-08-24/devel/py-unicorn.log http://build-failures.rhaalovely.net/mips64/2026-08-24/devel/riscv32-esp-elf/newlib.log http://build-failures.rhaalovely.net/mips64/2026-08-24/devel/tbb.log http://build-failures.rhaalovely.net/mips64/2026-08-24/devel/xtensa-esp32-elf/newlib.log http://build-failures.rhaalovely.net/mips64/2026-08-24/devel/xtensa-esp32s2-elf/newlib.log http://build-failures.rhaalovely.net/mips64/2026-08-24/devel/xtensa-esp32s3-elf/newlib.log http://build-failures.rhaalovely.net/mips64/2026-08-24/devel/xtensa-lx106-elf/newlib.log http://build-failures.rhaalovely.net/mips64/2026-08-24/emulators/advancemame.log http://build-failures.rhaalovely.net/mips64/2026-08-24/emulators/desmume.log http://build-failures.rhaalovely.net/mips64/2026-08-24/emulators/dosbox-x.log http://build-failures.rhaalovely.net/mips64/2026-08-24/emulators/dynamips.log http://build-failures.rhaalovely.net/mips64/2026-08-24/emulators/libretro-pcsx-rearmed.log http://build-failures.rhaalovely.net/mips64/2026-08-24/emulators/openmsx.log http://build-failures.rhaalovely.net/mips64/2026-08-24/games/astromenace.log http://build-failures.rhaalovely.net/mips64/2026-08-24/games/blockrage.log http://build-failures.rhaalovely.net/mips64/2026-08-24/games/capitan-sevilla.log http://build-failures.rhaalovely.net/mips64/2026-08-24/games/endless-sky.log http://build-failures.rhaalovely.net/mips64/2026-08-24/games/freeciv,-client.log http://build-failures.rhaalovely.net/mips64/2026-08-24/games/gnukem.log http://build-failures.rhaalovely.net/mips64/2026-08-24/games/godot/pack3,-editor.log http://build-failures.rhaalovely.net/mips64/2026-08-24/games/hyperrogue.log http://build-failures.rhaalovely.net/mips64/2026-08-24/games/nblood.log http://build-failures.rhaalovely.net/mips64/2026-08-24/games/openomf.log http://build-failures.rhaalovely.net/mips64/2026-08-24/games/scummvm.log http://build-failures.rhaalovely.net/mips64/2026-08-24/games/wargus.log http://build-failures.rhaalovely.net/mips64/2026-08-24/geo/gpstk.log http://build-failures.rhaalovely.net/mips64/2026-08-24/graphics/py-matplotlib.log http://build-failures.rhaalovely.net/mips64/2026-08-24/graphics/stable-diffusion.cpp.log http://build-failures.rhaalovely.net/mips64/2026-08-24/graphics/vtk.log http://build-failures.rhaalovely.net/mips64/2026-08-24/inputmethods/ibus.log http://build-failures.rhaalovely.net/mips64/2026-08-24/lang/STk.log http://build-failures.rhaalovely.net/mips64/2026-08-24/lang/apl.log http://build-failures.rhaalovely.net/mips64/2026-08-24/lang/binaryen.log http://build-failures.rhaalovely.net/mips64/2026-08-24/lang/gambit.log http://build-failures.rhaalovely.net/mips64/2026-08-24/lang/gforth.log http://build-failures.rhaalovely.net/mips64/2026-08-24/lang/librep.log http://build-failures.rhaalovely.net/mips64/2026-08-24/lang/verilator.log http://build-failures.rhaalovely.net/mips64/2026-08-24/multimedia/libyuv:libyuv/8aeb3a9ca36341a640528e59b34b5d641080dca8.tar.gz.log http://build-failures.rhaalovely.net/mips64/2026-08-24/math/flann.log http://build-failures.rhaalovely.net/mips64/2026-08-24/math/flintlib.log http://build-failures.rhaalovely.net/mips64/2026-08-24/math/lean.log http://build-failures.rhaalovely.net/mips64/2026-08-24/math/lrs.log http://build-failures.rhaalovely.net/mips64/2026-08-24/math/med.log http://build-failures.rhaalovely.net/mips64/2026-08-24/math/ntl.log http://build-failures.rhaalovely.net/mips64/2026-08-24/net/gtk-gnutella.log http://build-failures.rhaalovely.net/mips64/2026-08-24/net/libcares.log http://build-failures.rhaalovely.net/mips64/2026-08-24/net/toxic,no_x11.log http://build-failures.rhaalovely.net/mips64/2026-08-24/plan9/drawterm.log http://build-failures.rhaalovely.net/mips64/2026-08-24/print/luametatex.log http://build-failures.rhaalovely.net/mips64/2026-08-24/security/botan2.log http://build-failures.rhaalovely.net/mips64/2026-08-24/security/distorm3.log http://build-failures.rhaalovely.net/mips64/2026-08-24/sysutils/fastfetch.log http://build-failures.rhaalovely.net/mips64/2026-08-24/sysutils/flashrom.log http://build-failures.rhaalovely.net/mips64/2026-08-24/sysutils/fprintd.log http://build-failures.rhaalovely.net/mips64/2026-08-24/sysutils/mangl.log http://build-failures.rhaalovely.net/mips64/2026-08-24/telephony/asterisk/16.log http://build-failures.rhaalovely.net/mips64/2026-08-24/textproc/aspell/core.log http://build-failures.rhaalovely.net/mips64/2026-08-24/textproc/py-rapidfuzz.log http://build-failures.rhaalovely.net/mips64/2026-08-24/x11/cde.log http://build-failures.rhaalovely.net/mips64/2026-08-24/x11/gnome/keyring.log http://build-failures.rhaalovely.net/mips64/2026-08-24/x11/notion.log http://build-failures.rhaalovely.net/mips64/2026-08-24/x11/qt5/qtscript.log http://build-failures.rhaalovely.net/mips64/2026-08-24/x11/qt6/qtdeclarative.log
Re: Hyprland + OpenBSD
On Tue, 1 Sep 2026 00:10:31 +0200 Nazarenko Mykyta <mn@superpc.suicide.solutions> wrote: > On Sun, 30 Aug 2026 17:30:05 -0400 (EDT) > Daniel Dickman <didickman@gmail.com> wrote: > > > Here's a proof of concept for the Hyprland wayland compositor which I've > > been running for a few weeks on amd64. > > > > Note that the ports may need a good amount of polishing in terms of > > checking deps, WANTLIBs, etc, etc. So I'm posting it here in case someone > > wants to complete the work as I'm out of time and energy. > > > > Note that lua 5.5 is now in ports thanks to sthen@'s ok. but we still need > > glaze imported. I've sent out a port for glaze separately and if another > > developer wants to ok and import that would be helpful. > > > > Hope sharing this work-in-progress is helpful to folks with an interest in > > getting official Hyprland packages on OpenBSD. > > Hola Daniel! > > Too many warnings during the build process, but I spotted that upstream > successfully fixed this : > > https://github.com/hyprwm/Hyprland/commit/af0d014cb26f536d8cb7cab2b9d5784f69767c8a > > I'm very interested in Hyprland on OpenBSD, so... > I'll try to finish it all up properly. > > Big thanks! > > With best regards, > Nazarenko Mykyta > Hola ports@, Daniel! I made a few minor changes to WANTLIB in hyprland and hyprlang. I also added a patch that eliminates cascade of warnings during compilation by placing overrides in the appropriate places. Once version 0.56.3 is released, I'll remove my patch, but for now, it's best to import it along with this patch, just to keep things clean. I tested this on a laptop with an i915 and on another laptop with an amdgpu. It works great ( at least I haven't noticed any hidden issues ). Tarball attached. Thanks again to Daniel for all his work. With best regards, Nazarenko Mykyta
Re: www/librewolf: bump to 155.0-1, this time correctly
Le Tue, Sep 01, 2026 at 02:59:45PM +0100, Leah Rowe a écrit : > > question: > > how can I know reliably when to update SO_VERSION? i assume this is when the > one or more of the library dependencies gets updated, but I can't be sure. > > Can someone tell me? https://www.openbsd.org/faq/ports/specialtopics.html#SharedLibs but that's a bit special with monsters, i bump by habit at each major release but noone should care, since nothing links against the mozilla libs. fwiw i've commited the previous iteration of the diff because i had to move on to other things. follow cvs commits...
Re: www/librewolf: bump to 155.0-1, this time correctly
Hi. I see that you merged the earlier version of my patch. I will diff the version you merged with this one, and then adapt that to send a revision patch instead. There were changes to PLIST and SO_VERSION in this newer version of the patch. Am 01.09.26 um 14:54 schrieb Leah Rowe: > Hi everyone, > > Here is an updated version of my LibreWolf 155.0-1 update, compared to > the one I sent earlier. It updates the PLIST properly, and that's the > only difference. > > This should be safe to merge, but I'm building it one more time just > to be sure. > > To re-cap, in addition to the 155.0-1 update, this patch also makes > the following changes: > > * PGO data re-used from firefox builds, for performance gain depending > on workload > > * Re-enables --enable-rust-simd in mozconfig for performance gain > depending on workload > > * Moves DIST_SUBDIR to mozilla instead of librewolf, so that PGO data > can be re-used from mozilla-firefox porting > > * Cleans up some stale patches accordingly; one patch was removed due > to it having already been applied upstream > > One other change: i forgot to update SO_VERSION in the previous > version of the patch. This has now been done. > > I'm now 99.9% sure that this patch is correct. I'll be 100% sure when > it finishes compiled (ETA one to two hours on my machine, after I send > this email). > -- Company director, Minifree Ltd Registered in England, No. 9361826 | VAT No. GB202190462 Registered Office: 19 Hilton Road, Canvey Island, Essex SS8 9QA, UK
Re: www/librewolf: bump to 155.0-1, this time correctly
question: how can I know reliably when to update SO_VERSION? i assume this is when the one or more of the library dependencies gets updated, but I can't be sure. Can someone tell me? I saw that the firefox 155.0 update also updated this variable, so I updated it accordingly in my librewolf 155.0-1 update, but some confirmation would be nice. I assume also it is when there are breaking ABI changes in libraries. Am 01.09.26 um 14:54 schrieb Leah Rowe: > Hi everyone, > > Here is an updated version of my LibreWolf 155.0-1 update, compared to > the one I sent earlier. It updates the PLIST properly, and that's the > only difference. > > This should be safe to merge, but I'm building it one more time just > to be sure. > > To re-cap, in addition to the 155.0-1 update, this patch also makes > the following changes: > > * PGO data re-used from firefox builds, for performance gain depending > on workload > > * Re-enables --enable-rust-simd in mozconfig for performance gain > depending on workload > > * Moves DIST_SUBDIR to mozilla instead of librewolf, so that PGO data > can be re-used from mozilla-firefox porting > > * Cleans up some stale patches accordingly; one patch was removed due > to it having already been applied upstream > > One other change: i forgot to update SO_VERSION in the previous > version of the patch. This has now been done. > > I'm now 99.9% sure that this patch is correct. I'll be 100% sure when > it finishes compiled (ETA one to two hours on my machine, after I send > this email). > -- Company director, Minifree Ltd Registered in England, No. 9361826 | VAT No. GB202190462 Registered Office: 19 Hilton Road, Canvey Island, Essex SS8 9QA, UK
www/librewolf: bump to 155.0-1, this time correctly
Hi everyone, Here is an updated version of my LibreWolf 155.0-1 update, compared to the one I sent earlier. It updates the PLIST properly, and that's the only difference. This should be safe to merge, but I'm building it one more time just to be sure. To re-cap, in addition to the 155.0-1 update, this patch also makes the following changes: * PGO data re-used from firefox builds, for performance gain depending on workload * Re-enables --enable-rust-simd in mozconfig for performance gain depending on workload * Moves DIST_SUBDIR to mozilla instead of librewolf, so that PGO data can be re-used from mozilla-firefox porting * Cleans up some stale patches accordingly; one patch was removed due to it having already been applied upstream One other change: i forgot to update SO_VERSION in the previous version of the patch. This has now been done. I'm now 99.9% sure that this patch is correct. I'll be 100% sure when it finishes compiled (ETA one to two hours on my machine, after I send this email). -- Company director, Minifree Ltd Registered in England, No. 9361826 | VAT No. GB202190462 Registered Office: 19 Hilton Road, Canvey Island, Essex SS8 9QA, UK
From 557254d6400029666c252592755d0b95bc7d4e9d Mon Sep 17 00:00:00 2001
From: Leah Rowe <leah@libreboot.org>
Date: Sun, 30 Aug 2026 20:51:18 +0100
Subject: [PATCH 1/1] www/librewolf: bump to release 155.0-1
patches/patch-xpcom_ds_nsAtom_h removed, because this change seems
to have been made upstream. changes in this patch, besides 155.0:
use firefox PGO data, matched to the corresponding firefox
version (155.0) matching librewolf 155.0-1.
although librewolf doesn't provide PGO data, the two codebases
are so similar that using PGO for speed optimisation will work
on areas of the code that remain untouched by librewolf, which
is most of the code. therefore, using mozilla profdata works fine.
openbsd already re-uses mozilla PGO data despite modifying firefox.
the changes made by librewolf aren't aggressive enough to really
invalidate the profdata, so it will still provide a viable boost
to most workloads.
i've re-added the rust simd build option, which is fine so long
as you keep both librewolf and rustc/cargo up to date in the ports
system. i initially removed this option during development of the
port, and i was often using older versions of librewolf or its
libraries, which includes rust libraries, which sometimes caused
build errors that removing this option mitigated. for regular ports
maintenance, it is best that this option remains enabled as it
enables a nice performance boost for users in certain workloads.
MERGE NOTE:
DIST_SUBDIR is changed to mozilla, so now librewolf shares the same
subdir as firefox. This avoids duplicating identical copies of
mozilla PGO data on the openbsd dist server.
This way, we avoid any duplication. This means that there will
forever by a stale librewolf/ directory on the dist server. I tested
it with this change, and it works fine. LibreWolf tarballs now
download into mozilla/ on the dist server.
I waited until the next release update before doing this, to
avoid the 154.0.1-3 release being duplicated on the openbsd
dist server.
Signed-off-by: Leah Rowe <leah@libreboot.org>
---
www/librewolf/Makefile | 13 ++++++---
www/librewolf/distinfo | 6 ++--
.../patches/patch-config_makefiles_rust_mk | 2 +-
www/librewolf/patches/patch-mozconfig | 12 +++-----
www/librewolf/patches/patch-xpcom_ds_nsAtom_h | 29 -------------------
www/librewolf/pkg/PLIST | 4 +--
6 files changed, 19 insertions(+), 47 deletions(-)
delete mode 100644 www/librewolf/patches/patch-xpcom_ds_nsAtom_h
diff --git a/www/librewolf/Makefile b/www/librewolf/Makefile
index ba70810ef68..b25823a96bb 100644
--- a/www/librewolf/Makefile
+++ b/www/librewolf/Makefile
@@ -13,16 +13,22 @@ ONLY_FOR_ARCHS = amd64 aarch64 riscv64
# If upstream adds revision number -x, it will become plx in OpenBSD, e.g.
# librewolf-149.0.2-2 (upstream) becomes librewolf-149.0.2pl2 (patch level 2)
-MOZILLA_DIST_VERSION = 154.0.1-3
+MOZILLA_DIST_VERSION = 155.0-1
MOZILLA_VERSION = ${MOZILLA_DIST_VERSION:C/-([0-9]+)$/pl\1/}
-REVISION = 0
MOZILLA_PROJECT = librewolf
MOZILLA_CODENAME = browser
MAINTAINER = Leah Rowe <leah@libreboot.org>
+# LibreWolf doesn't provide PGO data, but we use the data
+# provided by FireFox releases, re-using the OpenBSD distfile
+MOZILLA_PROFDATA_TASKID = cNOKkWdaQfOqEzllBpy9Rw
+FFVERSION = ${MOZILLA_DIST_VERSION:C/-([0-9]+)$//}
+FFDIST = firefox-${FFVERSION}.source
+DISTFILES.profdata = ${FFDIST}-profdata.tar.xz{profdata.tar.xz}
+
HOMEPAGE = https://librewolf.net/
-SO_VERSION = 3.0
+SO_VERSION = 4.0
# Bump minor ver if shlibs deleted in components dir to avoid pkg_add -r issues
MOZILLA_LIBS = mozavcodec mozavutil mozgtk mozinference mozwayland xul \
clearkey gkcodecs lgpllibs mozsqlite3
@@ -32,7 +38,6 @@ CATEGORIES = www
MOZSITE = https://librewolf.dev/api/packages/librewolf/generic
SITES = ${MOZSITE}/librewolf-source/${MOZILLA_DIST_VERSION}/
EXTRACT_SUFX = .tar.gz
-DIST_SUBDIR = librewolf
# mozilla public license
PERMIT_PACKAGE= Yes
diff --git a/www/librewolf/distinfo b/www/librewolf/distinfo
index a9b3cc0a0e6..0e403401391 100644
--- a/www/librewolf/distinfo
+++ b/www/librewolf/distinfo
@@ -1,2 +1,4 @@
-SHA256 (librewolf/librewolf-154.0.1-3.source.tar.gz) = /ubKrDD5r/bf2P6GjOhDYCsDwwcmCchv9/xsF8OXpsw=
-SIZE (librewolf/librewolf-154.0.1-3.source.tar.gz) = 1175015645
+SHA256 (mozilla/firefox-155.0.source-profdata.tar.xz) = ML66F7kaeSloWTP2fL28H2w3ep0UMrnRd60HQeXg+NY=
+SHA256 (mozilla/librewolf-155.0-1.source.tar.gz) = XZUdgHHva8xOq4u6FJLiaa9yjINYZDfsKn2xHUa+NvY=
+SIZE (mozilla/firefox-155.0.source-profdata.tar.xz) = 18883580
+SIZE (mozilla/librewolf-155.0-1.source.tar.gz) = 1175547857
diff --git a/www/librewolf/patches/patch-config_makefiles_rust_mk b/www/librewolf/patches/patch-config_makefiles_rust_mk
index 6956d2ba803..177aa8e43ee 100644
--- a/www/librewolf/patches/patch-config_makefiles_rust_mk
+++ b/www/librewolf/patches/patch-config_makefiles_rust_mk
@@ -5,7 +5,7 @@ https://bugzilla.mozilla.org/show_bug.cgi?id=1640982
Index: config/makefiles/rust.mk
--- config/makefiles/rust.mk.orig
+++ config/makefiles/rust.mk
-@@ -118,7 +118,8 @@ ifndef rustflags_sancov
+@@ -91,7 +91,8 @@ ifeq (,$(RUST_SANCOV_FLAGS))
# Never enable when coverage is enabled to work around https://github.com/rust-lang/rust/issues/90045.
ifndef MOZ_CODE_COVERAGE
ifeq (,$(findstring gkrust_gtest,$(RUST_LIBRARY_FILE)))
diff --git a/www/librewolf/patches/patch-mozconfig b/www/librewolf/patches/patch-mozconfig
index 011bd84fadf..327fa7440ef 100644
--- a/www/librewolf/patches/patch-mozconfig
+++ b/www/librewolf/patches/patch-mozconfig
@@ -10,7 +10,7 @@ restored on re-builds, matching what was produced by this patch.
Index: mozconfig
--- mozconfig.orig
+++ mozconfig
-@@ -1,20 +1,29 @@
+@@ -1,9 +1,16 @@
+# %%VARIABLE%% is replaced with a string set by the
+# OpenBSD port Makefile at build time. In addition to
+# these changes, the LibreWolf OpenBSD port Makefile
@@ -29,12 +29,8 @@ Index: mozconfig
ac_add_options --disable-updater
ac_add_options --disable-cargo-incremental
ac_add_options --enable-hardening
- ac_add_options --enable-stl-hardening
- ac_add_options --enable-optimize
- ac_add_options --enable-release
--ac_add_options --enable-rust-simd
-+# Rust SIMD option causes build errors on OpenBSD:
-+# ac_add_options --enable-rust-simd
+@@ -13,8 +20,9 @@ ac_add_options --enable-release
+ ac_add_options --enable-rust-simd
# allow replacing malloc manually, will not affect regular systems
# not compatible with ac_add_options --disable-jemalloc, which only works when preloading custom malloc
-ac_add_options --enable-replace-malloc
@@ -45,7 +41,7 @@ Index: mozconfig
ac_add_options --with-app-name=librewolf
ac_add_options --with-branding=browser/branding/librewolf
-@@ -33,7 +42,19 @@
+@@ -33,7 +41,19 @@ mk_add_options MOZ_TELEMETRY_REPORTING=0
export CFLAGS="-ftrivial-auto-var-init=zero -fwrapv -Wno-backend-plugin $CFLAGS"
export CXXFLAGS="-ftrivial-auto-var-init=zero -fwrapv -Wno-backend-plugin $CXXFLAGS"
diff --git a/www/librewolf/patches/patch-xpcom_ds_nsAtom_h b/www/librewolf/patches/patch-xpcom_ds_nsAtom_h
deleted file mode 100644
index 851ab0078b3..00000000000
--- a/www/librewolf/patches/patch-xpcom_ds_nsAtom_h
+++ /dev/null
@@ -1,29 +0,0 @@
-fix build with libcxx22
-https://phabricator.services.mozilla.com/D316723
-
-Index: xpcom/ds/nsAtom.h
---- xpcom/ds/nsAtom.h.orig
-+++ xpcom/ds/nsAtom.h
-@@ -36,11 +36,19 @@ class nsDynamicAtom;
- class nsAtom {
- public:
- // Returns true if ToLowercaseASCII would return the string unchanged.
-+ //
-+ // This is deliberately a plain loop rather than std::all_of: gGkAtoms calls
-+ // this for every static atom within a single constant expression, and
-+ // std::all_of costs enough extra constexpr steps to exceed the default
-+ // -fconstexpr-steps budget with some standard library implementations.
- static constexpr bool ComputeIsAsciiLowercase(const char16_t* aString,
- const uint32_t aLength) {
-- return std::all_of(aString, aString + aLength, [](char16_t c) {
-- return !mozilla::IsAsciiUppercaseAlpha(c);
-- });
-+ for (uint32_t i = 0; i < aLength; ++i) {
-+ if (mozilla::IsAsciiUppercaseAlpha(aString[i])) {
-+ return false;
-+ }
-+ }
-+ return true;
- }
-
- template <size_t N>
diff --git a/www/librewolf/pkg/PLIST b/www/librewolf/pkg/PLIST
index 8a89d2c1e3b..c73c905d5b8 100644
--- a/www/librewolf/pkg/PLIST
+++ b/www/librewolf/pkg/PLIST
@@ -49,7 +49,7 @@ lib/${MOZILLA_PROJECT}/dependentlibs.list
lib/${MOZILLA_PROJECT}/distribution/
lib/${MOZILLA_PROJECT}/distribution/distribution.ini
lib/${MOZILLA_PROJECT}/distribution/policies.json
-@bin lib/${MOZILLA_PROJECT}/glxtest
+@bin lib/${MOZILLA_PROJECT}/gfxtest
lib/${MOZILLA_PROJECT}/gmp-clearkey/
lib/${MOZILLA_PROJECT}/gmp-clearkey/0.1/
@lib lib/${MOZILLA_PROJECT}/gmp-clearkey/0.1/libclearkey.so.${LIBclearkey_VERSION}
@@ -69,8 +69,6 @@ lib/${MOZILLA_PROJECT}/${MOZILLA_PROJECT}.cfg
lib/${MOZILLA_PROJECT}/omni.ja
lib/${MOZILLA_PROJECT}/platform.ini
lib/${MOZILLA_PROJECT}/removed-files
-@bin lib/${MOZILLA_PROJECT}/vaapitest
-@bin lib/${MOZILLA_PROJECT}/vulkantest
@man man/man1/${MOZILLA_PROJECT}.1
share/applications/${MOZILLA_PROJECT}.desktop
share/doc/pkg-readmes/${PKGSTEM}
--
2.55.0
Re: www/librewolf: update to 155.0-1
Here's a new version of the patch, with make-update-patches applied My build failed on the last patch, because I ran out of disk space, so I took the opportunity to do it properly this time, thus: make patch make update-patches make make update-plist make package i used to do it in this order: make package, then make update-patches, then make update-plist. I'm told that make-package can fail if files specified no longer exist in the tarball, so I'm doing it this way now. I'll run update-plist later and send a new diff if that changes anything. Am 01.09.26 um 11:48 schrieb Leah Rowe: > Hi everyone, > > Thank you Landry for telling me how to derive PGO task IDs from > mozilla's CI. > > The attached patch updates LibreWolf to 155.0-1 > > I've enabled PGO, by re-using mozilla profdata as the firefox port does. > > I've enabled --enable-rust-simd for further speed improvement under > certain workloads. > > DIST_SUBDIR is now mozilla, so librewolf shares the same dist server > subdirectory as mozilla. This means the "librewolf" directory will no > longer have new tarballs added to it on the pkg server; perhaps > openbsd could merge the contents of it into mozilla and make a 301 > redirect on httpd? > > The patch contains more info in the commit description. I'm building > it now as we speak. A patch was removed as it seemed to be applied > upstream (xpcom patch). > > Happy hacking! > > PS: I still don't know how to use CVS. I patched the GitHub mirror. > -- Company director, Minifree Ltd Registered in England, No. 9361826 | VAT No. GB202190462 Registered Office: 19 Hilton Road, Canvey Island, Essex SS8 9QA, UK
From 1ad2eb59d36e8babfb2fa954eb1f463e917f1da7 Mon Sep 17 00:00:00 2001
From: Leah Rowe <leah@libreboot.org>
Date: Sun, 30 Aug 2026 20:51:18 +0100
Subject: [PATCH 1/1] www/librewolf: bump to release 155.0-1
patches/patch-xpcom_ds_nsAtom_h removed, because this change seems
to have been made upstream. changes in this patch, besides 155.0:
use firefox PGO data, matched to the corresponding firefox
version (155.0) matching librewolf 155.0-1.
although librewolf doesn't provide PGO data, the two codebases
are so similar that using PGO for speed optimisation will work
on areas of the code that remain untouched by librewolf, which
is most of the code. therefore, using mozilla profdata works fine.
openbsd already re-uses mozilla PGO data despite modifying firefox.
the changes made by librewolf aren't aggressive enough to really
invalidate the profdata, so it will still provide a viable boost
to most workloads.
i've re-added the rust simd build option, which is fine so long
as you keep both librewolf and rustc/cargo up to date in the ports
system. i initially removed this option during development of the
port, and i was often using older versions of librewolf or its
libraries, which includes rust libraries, which sometimes caused
build errors that removing this option mitigated. for regular ports
maintenance, it is best that this option remains enabled as it
enables a nice performance boost for users in certain workloads.
MERGE NOTE:
DIST_SUBDIR is changed to mozilla, so now librewolf shares the same
subdir as firefox. This avoids duplicating identical copies of
mozilla PGO data on the openbsd dist server.
This way, we avoid any duplication. This means that there will
forever by a stale librewolf/ directory on the dist server. I tested
it with this change, and it works fine. LibreWolf tarballs now
download into mozilla/ on the dist server.
I waited until the next release update before doing this, to
avoid the 154.0.1-3 release being duplicated on the openbsd
dist server.
Signed-off-by: Leah Rowe <leah@libreboot.org>
---
www/librewolf/Makefile | 11 +++++--
www/librewolf/distinfo | 6 ++--
.../patches/patch-config_makefiles_rust_mk | 2 +-
www/librewolf/patches/patch-mozconfig | 12 +++-----
www/librewolf/patches/patch-xpcom_ds_nsAtom_h | 29 -------------------
5 files changed, 17 insertions(+), 43 deletions(-)
delete mode 100644 www/librewolf/patches/patch-xpcom_ds_nsAtom_h
diff --git a/www/librewolf/Makefile b/www/librewolf/Makefile
index ba70810ef68..da124b8664e 100644
--- a/www/librewolf/Makefile
+++ b/www/librewolf/Makefile
@@ -13,14 +13,20 @@ ONLY_FOR_ARCHS = amd64 aarch64 riscv64
# If upstream adds revision number -x, it will become plx in OpenBSD, e.g.
# librewolf-149.0.2-2 (upstream) becomes librewolf-149.0.2pl2 (patch level 2)
-MOZILLA_DIST_VERSION = 154.0.1-3
+MOZILLA_DIST_VERSION = 155.0-1
MOZILLA_VERSION = ${MOZILLA_DIST_VERSION:C/-([0-9]+)$/pl\1/}
-REVISION = 0
MOZILLA_PROJECT = librewolf
MOZILLA_CODENAME = browser
MAINTAINER = Leah Rowe <leah@libreboot.org>
+# LibreWolf doesn't provide PGO data, but we use the data
+# provided by FireFox releases, re-using the OpenBSD distfile
+MOZILLA_PROFDATA_TASKID = cNOKkWdaQfOqEzllBpy9Rw
+FFVERSION = ${MOZILLA_DIST_VERSION:C/-([0-9]+)$//}
+FFDIST = firefox-${FFVERSION}.source
+DISTFILES.profdata = ${FFDIST}-profdata.tar.xz{profdata.tar.xz}
+
HOMEPAGE = https://librewolf.net/
SO_VERSION = 3.0
# Bump minor ver if shlibs deleted in components dir to avoid pkg_add -r issues
@@ -32,7 +38,6 @@ CATEGORIES = www
MOZSITE = https://librewolf.dev/api/packages/librewolf/generic
SITES = ${MOZSITE}/librewolf-source/${MOZILLA_DIST_VERSION}/
EXTRACT_SUFX = .tar.gz
-DIST_SUBDIR = librewolf
# mozilla public license
PERMIT_PACKAGE= Yes
diff --git a/www/librewolf/distinfo b/www/librewolf/distinfo
index a9b3cc0a0e6..0e403401391 100644
--- a/www/librewolf/distinfo
+++ b/www/librewolf/distinfo
@@ -1,2 +1,4 @@
-SHA256 (librewolf/librewolf-154.0.1-3.source.tar.gz) = /ubKrDD5r/bf2P6GjOhDYCsDwwcmCchv9/xsF8OXpsw=
-SIZE (librewolf/librewolf-154.0.1-3.source.tar.gz) = 1175015645
+SHA256 (mozilla/firefox-155.0.source-profdata.tar.xz) = ML66F7kaeSloWTP2fL28H2w3ep0UMrnRd60HQeXg+NY=
+SHA256 (mozilla/librewolf-155.0-1.source.tar.gz) = XZUdgHHva8xOq4u6FJLiaa9yjINYZDfsKn2xHUa+NvY=
+SIZE (mozilla/firefox-155.0.source-profdata.tar.xz) = 18883580
+SIZE (mozilla/librewolf-155.0-1.source.tar.gz) = 1175547857
diff --git a/www/librewolf/patches/patch-config_makefiles_rust_mk b/www/librewolf/patches/patch-config_makefiles_rust_mk
index 6956d2ba803..177aa8e43ee 100644
--- a/www/librewolf/patches/patch-config_makefiles_rust_mk
+++ b/www/librewolf/patches/patch-config_makefiles_rust_mk
@@ -5,7 +5,7 @@ https://bugzilla.mozilla.org/show_bug.cgi?id=1640982
Index: config/makefiles/rust.mk
--- config/makefiles/rust.mk.orig
+++ config/makefiles/rust.mk
-@@ -118,7 +118,8 @@ ifndef rustflags_sancov
+@@ -91,7 +91,8 @@ ifeq (,$(RUST_SANCOV_FLAGS))
# Never enable when coverage is enabled to work around https://github.com/rust-lang/rust/issues/90045.
ifndef MOZ_CODE_COVERAGE
ifeq (,$(findstring gkrust_gtest,$(RUST_LIBRARY_FILE)))
diff --git a/www/librewolf/patches/patch-mozconfig b/www/librewolf/patches/patch-mozconfig
index 011bd84fadf..327fa7440ef 100644
--- a/www/librewolf/patches/patch-mozconfig
+++ b/www/librewolf/patches/patch-mozconfig
@@ -10,7 +10,7 @@ restored on re-builds, matching what was produced by this patch.
Index: mozconfig
--- mozconfig.orig
+++ mozconfig
-@@ -1,20 +1,29 @@
+@@ -1,9 +1,16 @@
+# %%VARIABLE%% is replaced with a string set by the
+# OpenBSD port Makefile at build time. In addition to
+# these changes, the LibreWolf OpenBSD port Makefile
@@ -29,12 +29,8 @@ Index: mozconfig
ac_add_options --disable-updater
ac_add_options --disable-cargo-incremental
ac_add_options --enable-hardening
- ac_add_options --enable-stl-hardening
- ac_add_options --enable-optimize
- ac_add_options --enable-release
--ac_add_options --enable-rust-simd
-+# Rust SIMD option causes build errors on OpenBSD:
-+# ac_add_options --enable-rust-simd
+@@ -13,8 +20,9 @@ ac_add_options --enable-release
+ ac_add_options --enable-rust-simd
# allow replacing malloc manually, will not affect regular systems
# not compatible with ac_add_options --disable-jemalloc, which only works when preloading custom malloc
-ac_add_options --enable-replace-malloc
@@ -45,7 +41,7 @@ Index: mozconfig
ac_add_options --with-app-name=librewolf
ac_add_options --with-branding=browser/branding/librewolf
-@@ -33,7 +42,19 @@
+@@ -33,7 +41,19 @@ mk_add_options MOZ_TELEMETRY_REPORTING=0
export CFLAGS="-ftrivial-auto-var-init=zero -fwrapv -Wno-backend-plugin $CFLAGS"
export CXXFLAGS="-ftrivial-auto-var-init=zero -fwrapv -Wno-backend-plugin $CXXFLAGS"
diff --git a/www/librewolf/patches/patch-xpcom_ds_nsAtom_h b/www/librewolf/patches/patch-xpcom_ds_nsAtom_h
deleted file mode 100644
index 851ab0078b3..00000000000
--- a/www/librewolf/patches/patch-xpcom_ds_nsAtom_h
+++ /dev/null
@@ -1,29 +0,0 @@
-fix build with libcxx22
-https://phabricator.services.mozilla.com/D316723
-
-Index: xpcom/ds/nsAtom.h
---- xpcom/ds/nsAtom.h.orig
-+++ xpcom/ds/nsAtom.h
-@@ -36,11 +36,19 @@ class nsDynamicAtom;
- class nsAtom {
- public:
- // Returns true if ToLowercaseASCII would return the string unchanged.
-+ //
-+ // This is deliberately a plain loop rather than std::all_of: gGkAtoms calls
-+ // this for every static atom within a single constant expression, and
-+ // std::all_of costs enough extra constexpr steps to exceed the default
-+ // -fconstexpr-steps budget with some standard library implementations.
- static constexpr bool ComputeIsAsciiLowercase(const char16_t* aString,
- const uint32_t aLength) {
-- return std::all_of(aString, aString + aLength, [](char16_t c) {
-- return !mozilla::IsAsciiUppercaseAlpha(c);
-- });
-+ for (uint32_t i = 0; i < aLength; ++i) {
-+ if (mozilla::IsAsciiUppercaseAlpha(aString[i])) {
-+ return false;
-+ }
-+ }
-+ return true;
- }
-
- template <size_t N>
--
2.55.0
Re: www/librewolf: update to 155.0-1
build failed, but not because there's anything wrong with the patch. i ran out of disk space! i'm freeing up some space and restarting the build. i'm confident that it'll build and run perfectly though, once i've done that. Am 01.09.26 um 13:33 schrieb Leah Rowe: > > then in cases where the patch does apply but would be diffede when > running make-update, what is the regular build process doing? > > i won't really worry about it. i'll just make sure to run everything > the way i said in the last email from now on: > > make patch > > make update-patches > > make > > make update-plist > > and to test: make package (i prefer to install from a package instead > of doing make-install) > > nevermind my question. i'll just do it this way from now on. by the > way, thhe build has succeeded. it's just linking now. i'll test it > soon, but i think this update is ready to go. > > > > Am 01.09.26 um 13:28 schrieb Landry Breuil: >> Le Tue, Sep 01, 2026 at 01:19:27PM +0100, Leah Rowe a écrit : >>> Make update-patches confuses me. If it would produce a diff, but a >>> ran it >>> after running make, does that mean when building a package, and >>> patching, it >>> will do the equivalent of make update-patches just for applying >>> them, but >>> not actually update them? (sort of like git cherry pick without error) >> i don't think i'm able to process that sentence. >> >> make update-patches (see infrastructure/mk/bsd.port.mk) looks for >> .orig.port files in $WRKSRC, and regens the patches. if you've made a >> new patch (cp x x.orig.port, edit x) it will pick it up. nothing else >> updates patches. >> >> you can run make update-patches anytime if you know your patches just to >> update the churn from code moving around/line changes, but sometimes >> (when PKG_SUBST is ran for example) you'd get some chunks modified, in >> the general case it's better to run it after make patch. >> >> make does call make patch as a dependency. make update-patches is never >> run as a dependency of another target, you need to run it manually. > -- Company director, Minifree Ltd Registered in England, No. 9361826 | VAT No. GB202190462 Registered Office: 19 Hilton Road, Canvey Island, Essex SS8 9QA, UK