CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: o...@cvs.openbsd.org2024/05/28 04:04:23 Modified files: net/powerdns : Makefile distinfo Log message: Update to PowerDNS Authoritative Server 4.9.1
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: o...@cvs.openbsd.org2024/05/14 04:00:15 Modified files: net/powerdns_recursor: Makefile distinfo Log message: Update to PowerDNS Recursor 5.0.5
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: o...@cvs.openbsd.org2024/05/13 04:09:08 Modified files: net/dnsdist: Makefile distinfo Log message: Update to dnsdist-1.9.4, including security fix for for very specific configs. https://dnsdist.org/security-advisories/powerdns-advisory-for-dnsdist-2024-03.html
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: o...@cvs.openbsd.org2024/05/07 23:30:14 Modified files: net/quiche : Makefile crates.inc distinfo Log message: Update quiche to 0.21.0
udpate net/quiche
Hi, OK? -Otto Index: net/quiche/Makefile === RCS file: /home/cvs/ports/net/quiche/Makefile,v diff -u -p -r1.3 Makefile --- net/quiche/Makefile 18 Mar 2024 06:46:18 - 1.3 +++ net/quiche/Makefile 6 May 2024 12:30:42 - @@ -3,7 +3,7 @@ COMMENT = library implementing QUIC and # ring-v0.17 does not support this arch NOT_FOR_ARCHS =sparc64 -V =0.20.1 +V =0.21.0 PKG_NAME = quiche-${V} CATEGORIES = net HOMEPAGE = https://github.com/cloudflare/quiche Index: net/quiche/crates.inc === RCS file: /home/cvs/ports/net/quiche/crates.inc,v diff -u -p -r1.2 crates.inc --- net/quiche/crates.inc 20 Dec 2023 11:25:58 - 1.2 +++ net/quiche/crates.inc 6 May 2024 13:01:19 - @@ -1,121 +1,112 @@ -MODCARGO_CRATES += aho-corasick1.1.2 # Unlicense OR MIT -MODCARGO_CRATES += android-tzdata 0.1.1 # MIT OR Apache-2.0 -MODCARGO_CRATES += android_system_properties 0.1.5 # MIT/Apache-2.0 +MODCARGO_CRATES += aho-corasick1.1.3 # Unlicense OR MIT MODCARGO_CRATES += arrayvec0.7.4 # MIT OR Apache-2.0 -MODCARGO_CRATES += atty0.2.14 # MIT -MODCARGO_CRATES += autocfg 1.1.0 # Apache-2.0 OR MIT -MODCARGO_CRATES += base64 0.21.5 # MIT OR Apache-2.0 +MODCARGO_CRATES += autocfg 1.3.0 # Apache-2.0 OR MIT MODCARGO_CRATES += bindgen 0.68.1 # BSD-3-Clause -MODCARGO_CRATES += bitflags2.4.1 # MIT OR Apache-2.0 -MODCARGO_CRATES += boring 4.2.0 # Apache-2.0 -MODCARGO_CRATES += boring-sys 4.2.0 # MIT -MODCARGO_CRATES += bumpalo 3.14.0 # MIT OR Apache-2.0 -MODCARGO_CRATES += cc 1.0.83 # MIT OR Apache-2.0 +MODCARGO_CRATES += bitflags2.5.0 # MIT OR Apache-2.0 +MODCARGO_CRATES += boring 4.6.0 # Apache-2.0 +MODCARGO_CRATES += boring-sys 4.6.0 # MIT +MODCARGO_CRATES += cc 1.0.97 # MIT OR Apache-2.0 MODCARGO_CRATES += cexpr 0.6.0 # Apache-2.0/MIT MODCARGO_CRATES += cfg-if 1.0.0 # MIT/Apache-2.0 -MODCARGO_CRATES += chrono 0.4.31 # MIT OR Apache-2.0 -MODCARGO_CRATES += clang-sys 1.6.1 # Apache-2.0 +MODCARGO_CRATES += clang-sys 1.7.0 # Apache-2.0 MODCARGO_CRATES += cmake 0.1.50 # MIT OR Apache-2.0 -MODCARGO_CRATES += core-foundation-sys 0.8.6 # MIT OR Apache-2.0 -MODCARGO_CRATES += darling 0.20.3 # MIT -MODCARGO_CRATES += darling_core0.20.3 # MIT -MODCARGO_CRATES += darling_macro 0.20.3 # MIT -MODCARGO_CRATES += data-encoding 2.5.0 # MIT -MODCARGO_CRATES += deranged0.3.10 # MIT OR Apache-2.0 +MODCARGO_CRATES += darling 0.20.8 # MIT +MODCARGO_CRATES += darling_core0.20.8 # MIT +MODCARGO_CRATES += darling_macro 0.20.8 # MIT +MODCARGO_CRATES += data-encoding 2.6.0 # MIT MODCARGO_CRATES += docopt 1.1.1 # Unlicense/MIT -MODCARGO_CRATES += either 1.9.0 # MIT OR Apache-2.0 -MODCARGO_CRATES += env_logger 0.6.2 # MIT/Apache-2.0 +MODCARGO_CRATES += either 1.11.0 # MIT OR Apache-2.0 +MODCARGO_CRATES += env_logger 0.10.2 # MIT OR Apache-2.0 MODCARGO_CRATES += equivalent 1.0.1 # Apache-2.0 OR MIT MODCARGO_CRATES += fnv 1.0.7 # Apache-2.0 / MIT MODCARGO_CRATES += foreign-types 0.5.0 # MIT/Apache-2.0 MODCARGO_CRATES += foreign-types-macros0.2.3 # MIT/Apache-2.0 MODCARGO_CRATES += foreign-types-shared0.3.1 # MIT/Apache-2.0 +MODCARGO_CRATES += form_urlencoded 1.2.1 # MIT OR Apache-2.0 MODCARGO_CRATES += fs_extra1.3.0 # MIT MODCARGO_CRATES += fslock 0.2.1 # MIT -MODCARGO_CRATES += getrandom 0.2.11 # MIT OR Apache-2.0 +MODCARGO_CRATES += getrandom 0.2.15 # MIT OR Apache-2.0 MODCARGO_CRATES += glob0.3.1 # MIT OR Apache-2.0 -MODCARGO_CRATES += hashbrown 0.12.3 # MIT OR Apache-2.0 -MODCARGO_CRATES += hashbrown 0.14.3 # MIT OR Apache-2.0 -MODCARGO_CRATES += hermit-abi 0.1.19 # MIT/Apache-2.0 -MODCARGO_CRATES += hex 0.4.3 # MIT OR Apache-2.0 -MODCARGO_CRATES += humantime 1.3.0 # MIT/Apache-2.0 -MODCARGO_CRATES += iana-time-zone 0.1.58 # MIT OR Apache-2.0 -MODCARGO_CRATES += iana-time-zone-haiku0.1.2 # MIT OR Apache-2.0 +MODCARGO_CRATES += hashbrown 0.14.5 # MIT OR Apache-2.0 +MODCARGO_CRATES += hermit-abi 0.3.9 # MIT OR Apache-2.0 +MODCARGO_CRATES += humantime 2.1.0 # MIT/Apache-2.0 MODCARGO_CRATES += ident_case 1.0.1 # MIT/Apache-2.0 MODCARGO_CRATES += idna0.1.5 # MIT/Apache-2.0 -MODCARGO_CRATES += indexmap1.9.3 # Apache-2.0 OR MIT -MODCARGO_CRATES += indexmap2.1.0 # Apache-2.0 OR MIT
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: o...@cvs.openbsd.org2024/04/24 23:56:42 Modified files: net/powerdns_recursor: Makefile distinfo Log message: Update to PowerDNS Recursor 5.0.4. https://docs.powerdns.com/recursor/security-advisories/powerdns-advisory-2024-02.html
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: o...@cvs.openbsd.org2024/04/05 08:25:50 Modified files: net/dnsdist: Makefile distinfo Log message: Update to dnsdist 1.9.3
Re: meson vs boost
On Mon, Mar 25, 2024 at 01:03:44PM +0100, Omar Polo wrote: > On 2024/03/22 20:12:41 +0100, Otto Moerbeek wrote: > > How about this? If you think it's ok, I'll try to see what upstream thinks > > the boost.py diff fails to apply for me, maybe it was generated against > the previous version (1.3.x)? Right, sorry, I did not pay attantion. Below looks ok. -Otto > > > Index: patches/patch-mesonbuild_dependencies_boost_py > === > RCS file: > /home/cvs/ports/devel/meson/patches/patch-mesonbuild_dependencies_boost_py,v > diff -u -p -r1.1 patch-mesonbuild_dependencies_boost_py > --- patches/patch-mesonbuild_dependencies_boost_py25 Mar 2024 07:12:30 > - 1.1 > +++ patches/patch-mesonbuild_dependencies_boost_py25 Mar 2024 11:49:49 > - > @@ -1,13 +1,13 @@ > Index: mesonbuild/dependencies/boost.py > --- mesonbuild/dependencies/boost.py.orig > +++ mesonbuild/dependencies/boost.py > -@@ -609,7 +609,8 @@ class BoostDependency(SystemDependency): > - mlog.debug(f'Static {len(libs)}') > +@@ -594,7 +594,8 @@ class BoostDependency(SystemDependency): > + # mlog.debug('- vscrt: {}'.format(vscrt)) > + libs = [x for x in libs if x.static == self.static or not > self.explicit_static] > libs = [x for x in libs if x.mt == self.multithreading] > - mlog.debug(f'MT {len(libs)}') > -libs = [x for x in libs if x.version_matches(lib_vers)] > +if not mesonlib.is_openbsd(): > +libs = [x for x in libs if x.version_matches(lib_vers)] > - mlog.debug(f'Vers {len(libs)} {lib_vers}') > libs = [x for x in libs if x.arch_matches(self.arch)] > - mlog.debug(f'Arch {len(libs)}') > + libs = [x for x in libs if x.vscrt_matches(vscrt)] > + libs = [x for x in libs if x.nvsuffix != 'dll'] # Only link to > import libraries > >
Re: meson vs boost
On Fri, Mar 22, 2024 at 08:55:19PM +0100, Antoine Jacoutot wrote: > Ok Thanks, Committed and upstreamed: https://github.com/mesonbuild/meson/pull/13006 -Otto > > — > Antoine > > > On 22 Mar 2024, at 20:12, Otto Moerbeek wrote: > > > > On Fri, Mar 22, 2024 at 05:55:22PM +0100, Antoine Jacoutot wrote: > > > >>> On Fri, Mar 22, 2024 at 01:54:13PM +0100, Otto Moerbeek wrote: > >>> Hi, > >>> > >>> at PowerDNS we are starting to use meson builds and I hit upon a snag > >>> wrt boost. > >>> > >>> When searching for libs, only .a's are found, all boost .so's are > >>> skipped. I tracked it down to line 608 in > >>> > >>> ./mesonbuild/dependencies/boost.py:filter_libraries: > >>> > >>>libs = [x for x in libs if x.version_matches(lib_vers)] > >>> > >>> Here lib_vers is set to the boost version 1_84, but presumably the .so > >>> version is used to compare against. On other systems that is the > >>> convention, but on OpenBSD that does not match. > >>> > >>> Disabling this line makes things work. > >>> > >>> So there seem to be two possible solutions: make the so version match > >>> the boost version, or just skip the test. > >>> > >>> I think the latter is preferred. If that's ok, I'll prep a diff. > >> > >> Agreed. > >> Otherwise we defeat the purpose of SHARED_LIBS. > >> It would be nice to have this discussed / pushed upstream if you could. > >> https://github.com/mesonbuild/meson > >> > >> Thanks. > >> > >> -- > >> Antoine > > > > How about this? If you think it's ok, I'll try to see what upstream thinks > > > >-Otto > > > > Index: Makefile > > === > > RCS file: /home/cvs/ports/devel/meson/Makefile,v > > diff -u -p -r1.132 Makefile > > --- Makefile14 Feb 2024 11:11:04 -1.132 > > +++ Makefile22 Mar 2024 19:10:08 - > > @@ -5,6 +5,7 @@ GH_ACCOUNT=mesonbuild > > GH_PROJECT=meson > > GH_TAGNAME=1.3.2 > > EPOCH=0 > > +REVISION=0 > > > > CATEGORIES=devel > > > > Index: patches/patch-mesonbuild_dependencies_boost_py > > === > > RCS file: patches/patch-mesonbuild_dependencies_boost_py > > diff -N patches/patch-mesonbuild_dependencies_boost_py > > --- /dev/null1 Jan 1970 00:00:00 - > > +++ patches/patch-mesonbuild_dependencies_boost_py22 Mar 2024 19:10:08 > > - > > @@ -0,0 +1,13 @@ > > +Index: mesonbuild/dependencies/boost.py > > +--- mesonbuild/dependencies/boost.py.orig > > mesonbuild/dependencies/boost.py > > +@@ -609,7 +609,8 @@ class BoostDependency(SystemDependency): > > + mlog.debug(f'Static {len(libs)}') > > + libs = [x for x in libs if x.mt == self.multithreading] > > + mlog.debug(f'MT {len(libs)}') > > +-libs = [x for x in libs if x.version_matches(lib_vers)] > > ++if not mesonlib.is_openbsd(): > > ++libs = [x for x in libs if x.version_matches(lib_vers)] > > + mlog.debug(f'Vers {len(libs)} {lib_vers}') > > + libs = [x for x in libs if x.arch_matches(self.arch)] > > + mlog.debug(f'Arch {len(libs)}') >
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: o...@cvs.openbsd.org2024/03/25 01:12:30 Modified files: devel/meson: Makefile Added files: devel/meson/patches: patch-mesonbuild_dependencies_boost_py Log message: On OpenBSD we do not have the convention that the .so version numbers are equal to the boost version number. Skip the check filtering on that condition. ok @ajacoutot
Re: meson vs boost
On Fri, Mar 22, 2024 at 05:55:22PM +0100, Antoine Jacoutot wrote: > On Fri, Mar 22, 2024 at 01:54:13PM +0100, Otto Moerbeek wrote: > > Hi, > > > > at PowerDNS we are starting to use meson builds and I hit upon a snag > > wrt boost. > > > > When searching for libs, only .a's are found, all boost .so's are > > skipped. I tracked it down to line 608 in > > > > ./mesonbuild/dependencies/boost.py:filter_libraries: > > > > libs = [x for x in libs if x.version_matches(lib_vers)] > > > > Here lib_vers is set to the boost version 1_84, but presumably the .so > > version is used to compare against. On other systems that is the > > convention, but on OpenBSD that does not match. > > > > Disabling this line makes things work. > > > > So there seem to be two possible solutions: make the so version match > > the boost version, or just skip the test. > > > > I think the latter is preferred. If that's ok, I'll prep a diff. > > Agreed. > Otherwise we defeat the purpose of SHARED_LIBS. > It would be nice to have this discussed / pushed upstream if you could. > https://github.com/mesonbuild/meson > > Thanks. > > -- > Antoine How about this? If you think it's ok, I'll try to see what upstream thinks -Otto Index: Makefile === RCS file: /home/cvs/ports/devel/meson/Makefile,v diff -u -p -r1.132 Makefile --- Makefile14 Feb 2024 11:11:04 - 1.132 +++ Makefile22 Mar 2024 19:10:08 - @@ -5,6 +5,7 @@ GH_ACCOUNT= mesonbuild GH_PROJECT=meson GH_TAGNAME=1.3.2 EPOCH= 0 +REVISION= 0 CATEGORIES=devel Index: patches/patch-mesonbuild_dependencies_boost_py === RCS file: patches/patch-mesonbuild_dependencies_boost_py diff -N patches/patch-mesonbuild_dependencies_boost_py --- /dev/null 1 Jan 1970 00:00:00 - +++ patches/patch-mesonbuild_dependencies_boost_py 22 Mar 2024 19:10:08 - @@ -0,0 +1,13 @@ +Index: mesonbuild/dependencies/boost.py +--- mesonbuild/dependencies/boost.py.orig mesonbuild/dependencies/boost.py +@@ -609,7 +609,8 @@ class BoostDependency(SystemDependency): + mlog.debug(f'Static {len(libs)}') + libs = [x for x in libs if x.mt == self.multithreading] + mlog.debug(f'MT {len(libs)}') +-libs = [x for x in libs if x.version_matches(lib_vers)] ++if not mesonlib.is_openbsd(): ++libs = [x for x in libs if x.version_matches(lib_vers)] + mlog.debug(f'Vers {len(libs)} {lib_vers}') + libs = [x for x in libs if x.arch_matches(self.arch)] + mlog.debug(f'Arch {len(libs)}')
meson vs boost
Hi, at PowerDNS we are starting to use meson builds and I hit upon a snag wrt boost. When searching for libs, only .a's are found, all boost .so's are skipped. I tracked it down to line 608 in ./mesonbuild/dependencies/boost.py:filter_libraries: libs = [x for x in libs if x.version_matches(lib_vers)] Here lib_vers is set to the boost version 1_84, but presumably the .so version is used to compare against. On other systems that is the convention, but on OpenBSD that does not match. Disabling this line makes things work. So there seem to be two possible solutions: make the so version match the boost version, or just skip the test. I think the latter is preferred. If that's ok, I'll prep a diff. -Otto
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: o...@cvs.openbsd.org2024/03/18 00:50:45 Modified files: net/powerdns : Makefile distinfo Log message: Update to PowerDNS Authoritative Server 4.9.0
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: o...@cvs.openbsd.org2024/03/18 00:47:53 Modified files: net/dnsdist: Makefile distinfo Log message: Update to dnsdist 1.9.1
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: o...@cvs.openbsd.org2024/03/18 00:46:18 Modified files: net/quiche : Makefile distinfo Log message: Update to quiche 0.20.1 This fixes a potential DOS. See https://github.com/cloudflare/quiche/releases/tag/0.20.1
Re: quiche security update
On Thu, Mar 14, 2024 at 08:15:53AM +0100, Christian Weisgerber wrote: > Otto Moerbeek: > > > maybe I was not being clear... it would be ncie to have this in the release. > > Too late. > > > > This is a security update to quiche. > > > See https://github.com/cloudflare/quiche/releases/tag/0.20.1 > > That only prevents a type of denial-of-service attack, anyway. Ok, I'll commit this when the ports tree is open again. -Otto
Re: quiche security update
Hi, maybe I was not being clear... it would be ncie to have this in the release. -Otto On Wed, Mar 13, 2024 at 08:05:47AM +0100, Otto Moerbeek wrote: > Hi, > > This is a security update to quiche. > > See https://github.com/cloudflare/quiche/releases/tag/0.20.1 > > sthen, naddy, OK? > > -Otto > > Index: Makefile > === > RCS file: /home/cvs/ports/net/quiche/Makefile,v > diff -u -p -r1.2 Makefile > --- Makefile 20 Dec 2023 11:25:58 - 1.2 > +++ Makefile 13 Mar 2024 07:03:03 - > @@ -3,7 +3,7 @@ COMMENT = library implementing QUIC and > # ring-v0.17 does not support this arch > NOT_FOR_ARCHS = sparc64 > > -V = 0.20.0 > +V = 0.20.1 > PKG_NAME = quiche-${V} > CATEGORIES = net > HOMEPAGE = https://github.com/cloudflare/quiche > Index: distinfo > === > RCS file: /home/cvs/ports/net/quiche/distinfo,v > diff -u -p -r1.2 distinfo > --- distinfo 20 Dec 2023 11:25:58 - 1.2 > +++ distinfo 13 Mar 2024 07:03:03 - > @@ -119,7 +119,7 @@ SHA256 (cargo/windows_i686_msvc-0.48.5.t > SHA256 (cargo/windows_x86_64_gnu-0.48.5.tar.gz) = > U9QKvSWD0j5HGP3fHr7ITb/4OBwHyuZ/93aLvxnGcY4= > SHA256 (cargo/windows_x86_64_gnullvm-0.48.5.tar.gz) = > C3tSdnhooj1bq3aOOQ3F9cVYJbbTC4bIRP8tx0FARMw= > SHA256 (cargo/windows_x86_64_msvc-0.48.5.tar.gz) = > 7ZT85hVxpABoUrc4mgY6uYPALrG7N7R/gnLOktBtlTg= > -SHA256 (quiche-0.20.0.tar.gz) = cSW8gt3POPv7xpiCzLJyO/tNW/60Jxi4KR0m7AYELjg= > +SHA256 (quiche-0.20.1.tar.gz) = nEYNjs9sgMBr+bQvkSAe8z+RLiYVqHH/LQ5QGXuQHHE= > SIZE (cargo/aho-corasick-1.1.2.tar.gz) = 183136 > SIZE (cargo/android-tzdata-0.1.1.tar.gz) = 7674 > SIZE (cargo/android_system_properties-0.1.5.tar.gz) = 5243 > @@ -241,4 +241,4 @@ SIZE (cargo/windows_i686_msvc-0.48.5.tar > SIZE (cargo/windows_x86_64_gnu-0.48.5.tar.gz) = 801619 > SIZE (cargo/windows_x86_64_gnullvm-0.48.5.tar.gz) = 418486 > SIZE (cargo/windows_x86_64_msvc-0.48.5.tar.gz) = 798412 > -SIZE (quiche-0.20.0.tar.gz) = 681940 > +SIZE (quiche-0.20.1.tar.gz) = 683362 >
quiche security update
Hi, This is a security update to quiche. See https://github.com/cloudflare/quiche/releases/tag/0.20.1 sthen, naddy, OK? -Otto Index: Makefile === RCS file: /home/cvs/ports/net/quiche/Makefile,v diff -u -p -r1.2 Makefile --- Makefile20 Dec 2023 11:25:58 - 1.2 +++ Makefile13 Mar 2024 07:03:03 - @@ -3,7 +3,7 @@ COMMENT = library implementing QUIC and # ring-v0.17 does not support this arch NOT_FOR_ARCHS =sparc64 -V =0.20.0 +V =0.20.1 PKG_NAME = quiche-${V} CATEGORIES = net HOMEPAGE = https://github.com/cloudflare/quiche Index: distinfo === RCS file: /home/cvs/ports/net/quiche/distinfo,v diff -u -p -r1.2 distinfo --- distinfo20 Dec 2023 11:25:58 - 1.2 +++ distinfo13 Mar 2024 07:03:03 - @@ -119,7 +119,7 @@ SHA256 (cargo/windows_i686_msvc-0.48.5.t SHA256 (cargo/windows_x86_64_gnu-0.48.5.tar.gz) = U9QKvSWD0j5HGP3fHr7ITb/4OBwHyuZ/93aLvxnGcY4= SHA256 (cargo/windows_x86_64_gnullvm-0.48.5.tar.gz) = C3tSdnhooj1bq3aOOQ3F9cVYJbbTC4bIRP8tx0FARMw= SHA256 (cargo/windows_x86_64_msvc-0.48.5.tar.gz) = 7ZT85hVxpABoUrc4mgY6uYPALrG7N7R/gnLOktBtlTg= -SHA256 (quiche-0.20.0.tar.gz) = cSW8gt3POPv7xpiCzLJyO/tNW/60Jxi4KR0m7AYELjg= +SHA256 (quiche-0.20.1.tar.gz) = nEYNjs9sgMBr+bQvkSAe8z+RLiYVqHH/LQ5QGXuQHHE= SIZE (cargo/aho-corasick-1.1.2.tar.gz) = 183136 SIZE (cargo/android-tzdata-0.1.1.tar.gz) = 7674 SIZE (cargo/android_system_properties-0.1.5.tar.gz) = 5243 @@ -241,4 +241,4 @@ SIZE (cargo/windows_i686_msvc-0.48.5.tar SIZE (cargo/windows_x86_64_gnu-0.48.5.tar.gz) = 801619 SIZE (cargo/windows_x86_64_gnullvm-0.48.5.tar.gz) = 418486 SIZE (cargo/windows_x86_64_msvc-0.48.5.tar.gz) = 798412 -SIZE (quiche-0.20.0.tar.gz) = 681940 +SIZE (quiche-0.20.1.tar.gz) = 683362
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: o...@cvs.openbsd.org2024/03/07 04:34:25 Modified files: net/powerdns_recursor: Makefile distinfo Log message: Update to PowerDNS Recursor 5.0.3
Re: [new] net/dnsq
On Mon, Feb 26, 2024 at 11:30:20AM +0100, Renaud Allard wrote: > > > On 2/20/24 11:04, Stuart Henderson wrote: > > On 2024/02/20 10:37, Renaud Allard wrote: > > > > > > > > > On 2/20/24 10:19, Stuart Henderson wrote: > > > > On 2024/02/20 09:30, Renaud Allard wrote: > > > > > Yes, the name doesn't tell anything by itself. > > > > > Is this one better? The binary is still called q, but the package is > > > > > dnsq. > > > > > > > > If naming the package dnsq, I'd suggest also adding a symlink, dnsq -> > > > > q. > > > > > > I have added the symlink > > > > No need for ONLY_FOR_ARCHS, it builds ok on i386. > > > > OK sthen@ to import. > > > > ping OK otto@, -Otto
Re: [new] net/q
On Tue, Feb 20, 2024 at 09:09:49AM +0100, Renaud Allard wrote: > > > On 2/19/24 16:50, Renaud Allard wrote: > > Hello, > > > > Here is a preliminary port for "q" which is a tiny and feature-rich > > command line DNS client with support for UDP, TCP, DoT, DoH, DoQ, and > > ODoH. > > > > Here is the port cleaned up. There is no man page (although you can get > usage with --help), so I included the readme.md in /usr/local/share/doc/q > > Any suggestions or OKs? I occasiaonally use it, so a port is welcome. I do have one gripe: the name, it's way too short and thus non-descriptive. dnsq or anything more descriptive would be better. -Otto
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: o...@cvs.openbsd.org2024/02/16 07:22:44 Removed files: www/h2o: Makefile distinfo www/h2o/files : h2o.conf www/h2o/patches: patch-CMakeLists_txt patch-deps_neverbleed_neverbleed_c patch-deps_neverbleed_neverbleed_h patch-lib_common_socket_c patch-src_main_c www/h2o/pkg: DESCR PLIST h2o.rc Log message: Zap h2o; its only consumer was dnsdist and that's no longer true; ok tb@
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: o...@cvs.openbsd.org2024/02/16 07:21:42 Modified files: www: Makefile Log message: Unhook h2o; ok tb@
Re: CVS: cvs.openbsd.org: ports
On Fri, Feb 16, 2024 at 07:01:22AM -0700, Otto Moerbeek wrote: > CVSROOT: /cvs > Module name: ports > Changes by: o...@cvs.openbsd.org2024/02/16 07:01:22 > > Modified files: > net/dnsdist: Makefile distinfo > net/dnsdist/patches: patch-configure_ac > net/dnsdist/pkg: DESCR > > Log message: > Update to dnsdist 1.9.0, introducing a -h2 flavor for incoming DoQ and DoH3 ehh -h3 > > This also drops the h2o dependenccy >
remove h2o?
Hi, now that dnsdist is no longer using h2o, its time for it to go away. It's aslo not being maintained in a useful way anymore by upstream. No releases etc... OK? -Otto
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: o...@cvs.openbsd.org2024/02/16 07:02:28 Modified files: net: Makefile Log message: Hook up dnsdist,h3
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: o...@cvs.openbsd.org2024/02/16 07:01:22 Modified files: net/dnsdist: Makefile distinfo net/dnsdist/patches: patch-configure_ac net/dnsdist/pkg: DESCR Log message: Update to dnsdist 1.9.0, introducing a -h2 flavor for incoming DoQ and DoH3 This also drops the h2o dependenccy
Re: Update to dnsdist 1.9.0, with new flavor
On Fri, Feb 16, 2024 at 10:20:49AM +, Stuart Henderson wrote: > Generally the diff looks good to me. I'd have a slight preference to > move SYSCONFDIR/COMPILER/MODULES/NODLUA_VERSION up so that the > flavour-independent variables are grouped together above the .if. > > On 2024/02/16 10:29, Renaud Allard wrote: > > Then maybe the flavor should be named -http3 because DoQ is http3 over non > > 443 port. > > I would prefer this (or just -h3) than -quiche. Also please mention > it in DESCR e.g. something like > > Flavours: > h3: build with libquiche/boringssl, providing HTTP/3 and QUIC support > On a side note, "DoQ is http3 over another port" is not true, there is no http layer involved in DoQ, it's more like TCP over QUIC. So the -h3 name is not ideal, but good enough. -Otto
Re: Update to dnsdist 1.9.0, with new flavor
On Fri, Feb 16, 2024 at 10:20:49AM +, Stuart Henderson wrote: > Generally the diff looks good to me. I'd have a slight preference to > move SYSCONFDIR/COMPILER/MODULES/NODLUA_VERSION up so that the > flavour-independent variables are grouped together above the .if. > > On 2024/02/16 10:29, Renaud Allard wrote: > > Then maybe the flavor should be named -http3 because DoQ is http3 over non > > 443 port. > > I would prefer this (or just -h3) than -quiche. Also please mention > it in DESCR e.g. something like > > Flavours: > h3: build with libquiche/boringssl, providing HTTP/3 and QUIC support > Updated diff, taking the -h3 and other suggesions. -Otto Index: Makefile === RCS file: /home/cvs/ports/net/dnsdist/Makefile,v diff -u -p -r1.36 Makefile --- Makefile15 Dec 2023 13:11:55 - 1.36 +++ Makefile16 Feb 2024 11:12:02 - @@ -1,6 +1,6 @@ COMMENT= highly DNS-, DoS- and abuse-aware loadbalancer -DISTNAME= dnsdist-1.8.3 +DISTNAME= dnsdist-1.9.0 EXTRACT_SUFX= .tar.bz2 CATEGORIES=net @@ -10,18 +10,28 @@ HOMEPAGE= https://dnsdist.org/ MAINTAINER=Otto Moerbeek # GPLv2 only -PERMIT_PACKAGE=Yes +PERMIT_PACKAGE=Yes -WANTLIB += ${MODLUA_WANTLIB} c curses edit m \ - fstrm h2o-evloop>=1 nghttp2 pthread \ - z crypto ssl sodium ${COMPILER_LIBCXX} +FLAVORS= h3 +FLAVOR?= + +WANTLIB += ${MODLUA_WANTLIB} ${COMPILER_LIBCXX} curses crypto edit m \ + fstrm nghttp2 pthread sodium SITES= https://downloads.powerdns.com/releases/ CONFIGURE_STYLE= autoconf -AUTOCONF_VERSION= 2.69 +AUTOCONF_VERSION= 2.71 USE_GMAKE= Yes +SYSCONFDIR=${BASESYSCONFDIR}/dnsdist + +COMPILER = base-clang ports-gcc + +MODULES+= lang/lua +MODLUA_VERSION=5.3 + + CONFIGURE_ARGS+= --without-gnutls \ --disable-hardening \ --with-lua=lua${MODLUA_VERSION:S/.//} \ @@ -34,27 +44,43 @@ CONFIGURE_ARGS+=--without-gnutls \ --enable-dnscrypt \ --enable-dnstap \ --enable-dns-over-https -# -# On at least amd64 using clang13 dnsdist-lua-bindings.cc gets -# miscompiled with -O2, resulting in a segfault on startup -# + CONFIGURE_ENV+=CPPFLAGS="-I${LOCALBASE}/include" \ LDFLAGS="-L${LOCALBASE}/lib" \ - LIBEDIT_LIBS='-ledit -lcurses' \ - LIBEDIT_CFLAGS=' ' -SYSCONFDIR=${BASESYSCONFDIR}/dnsdist - -COMPILER = base-clang ports-gcc - -MODULES+= lang/lua -MODLUA_VERSION=5.3 + LIBEDIT_LIBS="-ledit -lcurses" \ + LIBEDIT_CFLAGS=" " BUILD_DEPENDS= devel/boost LIB_DEPENDS+= ${MODLUA_LIB_DEPENDS} \ security/libsodium \ net/libfstrm \ - www/h2o \ www/nghttp2 + + +.if ${FLAVOR:L:Mh3} + +# Lua module overrides FULLPKGNAME +FULLPKGNAME= ${PKGNAME}-${FLAVOR} +CONFIGURE_ARGS+= --with-libcrypto=${LOCALBASE}/eboringssl \ + --with-quiche \ + --enable-dns-over-quic \ + --enable-dns-over-http3 \ + +# OpenBSD libtool seems to leave out boringssl libs, because they have the same +# name as system libs, so explicitly refer to the .a files. +CONFIGURE_ENV+= LIBSSL_CFLAGS="-I${LOCALBASE}/eboringssl/include \ + -DDISABLE_OCSP_STAPLING" \ + LIBSSL_LIBS="${LOCALBASE}/eboringssl/lib/libssl.a \ + ${LOCALBASE}/eboringssl/lib/libcrypto.a" \ + QUICHE_CFLAGS=" " \ + QUICHE_LIBS='-lquiche' +BUILD_DEPENDS+=net/quiche \ + security/boringssl/head +.else + +WANTLIB+= ssl + +.endif post-install: ${INSTALL_DATA_DIR} ${PREFIX}/share/{doc,examples}/dnsdist Index: distinfo === RCS file: /home/cvs/ports/net/dnsdist/distinfo,v diff -u -p -r1.19 distinfo --- distinfo15 Dec 2023 13:11:55 - 1.19 +++ distinfo16 Feb 2024 11:12:02 - @@ -1,2 +1,2 @@ -SHA256 (dnsdist-1.8.3.tar.bz2) = hYMj8u1RgUiLt1WPv0+E7HGYYAsHCyxTddFdQGlXJ/Q= -SIZE (dnsdist-1.8.3.tar.bz2) = 1484359 +SHA256 (dnsdist-1.9.0.tar.bz2) = FrqxXK2SRVcYBjmKjkpdwyqStrtg5hfBL+lYyUWInHw= +SIZE (dnsdist-1.9.0.tar.bz2) = 1620891 Index: patches/pat
Re: Update to dnsdist 1.9.0, with new flavor
On Fri, Feb 16, 2024 at 10:19:12AM +0100, Renaud Allard wrote: > > > On 2/16/24 10:05, Otto Moerbeek wrote: > > Hi, > > > > dnsdist 1.9.0 will be released today, he tarball is already there. > > > > The port is somewhat involved. dnsdist 1.9.0 includes incoming quick > > and http3 support, which is nice, but requires libquiche which > > requires boringssl. To avoid mixing libs, I link statically with those > > two. > > > > I also introduced a flavor for tha (-quiche), as some platforms get > > knocked out by needing libquiche and borinssl. > > > > So the defaut flavor is without quick and http3 support and builds on > > the same set of platforms as before. > > > > Does this makes all sense? I could use a review of the actual port > > itself as well. > > > > Thanks, > > > > I have been running dnsdist-1.9-alpha3 with DoQ without any issue for > several weeks on amd64 and arm64. > Maybe I would have named the flavor -doq and not -quiche as quiche is "only" > a needed library to allow DoQ, so it's not that meaningful. after alpha3 incoming http3 was added, so doq is also not a good flavor name. -Otto
Update to dnsdist 1.9.0, with new flavor
Hi, dnsdist 1.9.0 will be released today, he tarball is already there. The port is somewhat involved. dnsdist 1.9.0 includes incoming quick and http3 support, which is nice, but requires libquiche which requires boringssl. To avoid mixing libs, I link statically with those two. I also introduced a flavor for tha (-quiche), as some platforms get knocked out by needing libquiche and borinssl. So the defaut flavor is without quick and http3 support and builds on the same set of platforms as before. Does this makes all sense? I could use a review of the actual port itself as well. Thanks, -Otto Index: Makefile === RCS file: /home/cvs/ports/net/dnsdist/Makefile,v diff -u -p -r1.36 Makefile --- Makefile15 Dec 2023 13:11:55 - 1.36 +++ Makefile16 Feb 2024 08:55:53 - @@ -1,6 +1,6 @@ COMMENT= highly DNS-, DoS- and abuse-aware loadbalancer -DISTNAME= dnsdist-1.8.3 +DISTNAME= dnsdist-1.9.0 EXTRACT_SUFX= .tar.bz2 CATEGORIES=net @@ -10,16 +10,18 @@ HOMEPAGE= https://dnsdist.org/ MAINTAINER=Otto Moerbeek # GPLv2 only -PERMIT_PACKAGE=Yes +PERMIT_PACKAGE=Yes -WANTLIB += ${MODLUA_WANTLIB} c curses edit m \ - fstrm h2o-evloop>=1 nghttp2 pthread \ - z crypto ssl sodium ${COMPILER_LIBCXX} +FLAVORS= quiche +FLAVOR?= + +WANTLIB += ${MODLUA_WANTLIB} ${COMPILER_LIBCXX} curses crypto edit m \ + fstrm nghttp2 pthread sodium SITES= https://downloads.powerdns.com/releases/ CONFIGURE_STYLE= autoconf -AUTOCONF_VERSION= 2.69 +AUTOCONF_VERSION= 2.71 USE_GMAKE= Yes CONFIGURE_ARGS+= --without-gnutls \ @@ -34,27 +36,50 @@ CONFIGURE_ARGS+=--without-gnutls \ --enable-dnscrypt \ --enable-dnstap \ --enable-dns-over-https -# -# On at least amd64 using clang13 dnsdist-lua-bindings.cc gets -# miscompiled with -O2, resulting in a segfault on startup -# + CONFIGURE_ENV+=CPPFLAGS="-I${LOCALBASE}/include" \ LDFLAGS="-L${LOCALBASE}/lib" \ - LIBEDIT_LIBS='-ledit -lcurses' \ - LIBEDIT_CFLAGS=' ' -SYSCONFDIR=${BASESYSCONFDIR}/dnsdist - -COMPILER = base-clang ports-gcc - -MODULES+= lang/lua -MODLUA_VERSION=5.3 + LIBEDIT_LIBS="-ledit -lcurses" \ + LIBEDIT_CFLAGS=" " BUILD_DEPENDS= devel/boost LIB_DEPENDS+= ${MODLUA_LIB_DEPENDS} \ security/libsodium \ net/libfstrm \ - www/h2o \ www/nghttp2 + + +.if ${FLAVOR:L:Mquiche} + +# Lua module overrides FULLPKGNAME +FULLPKGNAME= ${PKGNAME}-${FLAVOR} +CONFIGURE_ARGS+= --with-libcrypto=${LOCALBASE}/eboringssl \ + --with-quiche \ + --enable-dns-over-quic \ + --enable-dns-over-http3 \ + +# OpenBSD libtool seems to leave out boringssl libs, because they have the same +# name as system libs, so explicitly refer to the .a files. +CONFIGURE_ENV+= LIBSSL_CFLAGS="-I${LOCALBASE}/eboringssl/include \ + -DDISABLE_OCSP_STAPLING" \ + LIBSSL_LIBS="${LOCALBASE}/eboringssl/lib/libssl.a \ + ${LOCALBASE}/eboringssl/lib/libcrypto.a" \ + QUICHE_CFLAGS=" " \ + QUICHE_LIBS='-lquiche' +BUILD_DEPENDS+=net/quiche \ + security/boringssl/head +.else + +WANTLIB+= ssl + +.endif + +SYSCONFDIR=${BASESYSCONFDIR}/dnsdist + +COMPILER = base-clang ports-gcc + +MODULES+= lang/lua +MODLUA_VERSION=5.3 post-install: ${INSTALL_DATA_DIR} ${PREFIX}/share/{doc,examples}/dnsdist Index: distinfo === RCS file: /home/cvs/ports/net/dnsdist/distinfo,v diff -u -p -r1.19 distinfo --- distinfo15 Dec 2023 13:11:55 - 1.19 +++ distinfo16 Feb 2024 08:55:53 - @@ -1,2 +1,2 @@ -SHA256 (dnsdist-1.8.3.tar.bz2) = hYMj8u1RgUiLt1WPv0+E7HGYYAsHCyxTddFdQGlXJ/Q= -SIZE (dnsdist-1.8.3.tar.bz2) = 1484359 +SHA256 (dnsdist-1.9.0.tar.bz2) = FrqxXK2SRVcYBjmKjkpdwyqStrtg5hfBL+lYyUWInHw= +SIZE (dnsdist-1.9.0.tar.bz2) = 1620891 Index: patches/patch-configure_ac === RCS file: /home/cvs/ports/net/dnsdist/patches/patch-configure_ac,v diff -u -p -r1.8 patch-configure_ac --- patches/patch-configure_a
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: o...@cvs.openbsd.org2024/02/13 06:53:41 Modified files: net/powerdns_recursor: Makefile distinfo Log message: Update to PowerDNS Recursor 5.0.2, fixing CVE-2023-50387 and CVE-2023-50868 https://blog.powerdns.com/2024/02/13/powerdns-recursor-4-8-6-4-9-3-5-0-2-released
Re: boost sparc64 support for context
On Tue, Feb 13, 2024 at 09:03:49AM +0100, Claudio Jeker wrote: > On Thu, Feb 08, 2024 at 06:50:55PM +0100, Claudio Jeker wrote: > > Since I did most of the work for php I decided to also implement the fiber > > / context ASM for sparc64 in boost. > > > > The difference is that boost has a ontop_fcontext() function which is > > not really documented but in the end I figured it out. > > > > I did test this against the tests in boost-context. Will do more tests > > tomorrow but maybe someone else wants to join the context party :) > > I realized that I can save an instruction by using the builtin ADD > instruction in 'restore' to set up %o0. So a context switch on sparc64 > is now 10 instructions :) > > I ran test for context, fiber, coroutine and coroutine2 in boost and all > test pass. The powerdns_recoursor package built but I'm not sure how to > test this properly. If it starts and does not crash you're good. My sparc64s are currently in storage, so I cannot test myself right now as our attic will be under construction soon. Will test when things are back to normal. -Otto > > IMO this is ready for commit. I will start pushing these changes upstream > as well. > -- > :wq Claudio > > Index: Makefile > === > RCS file: /cvs/ports/devel/boost/Makefile,v > diff -u -p -r1.142 Makefile > --- Makefile 4 Jan 2024 08:16:16 - 1.142 > +++ Makefile 12 Feb 2024 19:12:11 - > @@ -1,4 +1,4 @@ > -NOT_FOR_ARCHS-md = alpha hppa sparc64 > +NOT_FOR_ARCHS-md = alpha hppa > > DPB_PROPERTIES= parallel > > @@ -132,7 +132,7 @@ SUBST_VARS+= SO_VERSION > > # Revert back to Boost::Context 1.80.0 ASM code for i386 > post-extract: > - cp -f ${FILESDIR}/*_i386_sysv_elf_gas.S ${WRKSRC}/libs/context/src/asm/ > + cp -f ${FILESDIR}/*_sysv_elf_gas.S ${WRKSRC}/libs/context/src/asm/ > > do-configure: > echo "using ${TOOLSET} : : ${CXX} ;" > >>${WRKSRC}/tools/build/src/user-config.jam > Index: files/jump_sparc64_sysv_elf_gas.S > === > RCS file: files/jump_sparc64_sysv_elf_gas.S > diff -N files/jump_sparc64_sysv_elf_gas.S > --- /dev/null 1 Jan 1970 00:00:00 - > +++ files/jump_sparc64_sysv_elf_gas.S 12 Feb 2024 19:12:11 - > @@ -0,0 +1,51 @@ > +/* > + Copyright Claudio Jeker 2024 > + Distributed under the Boost Software License, Version 1.0. > + (See accompanying file LICENSE_1_0.txt or copy at > + http://www.boost.org/LICENSE_1_0.txt) > +*/ > + > +/* > + * typedef void* fcontext_t; > + * > + * struct transfer_t { > + * fcontext_t fctx; > + * void* data; > + * }; > + * > + * transfer_t jump_fcontext(fcontext_t const to, void *vp); > + */ > +#define CC64FSZ 176 > +#define BIAS 2047 > +#define SP 128 > +#define I7 136 > + > +.file "jump_sparc64_sysv_elf_gas.S" > +.text > +.align 4 > +.global jump_fcontext > +.type jump_fcontext, %function > +jump_fcontext: > + # prepare stack > + save%sp, -CC64FSZ, %sp > + > + # store framepointer and return address in slots reserved > + # for arguments > + stx %fp, [%sp + BIAS + SP] > + stx %i7, [%sp + BIAS + I7] > + mov %sp, %o0 > + # force flush register windows to stack and with that save context > + flushw > + # get SP (pointing to new context-data) from %i0 param > + mov %i0, %sp > + # load framepointer and return address from context > + ldx [%sp + BIAS + SP], %fp > + ldx [%sp + BIAS + I7], %i7 > + > + ret > + restore %o0, %g0, %o0 > + # restore old %sp (pointing to old context-data) in %o0 > + # *data stored in %o1 was not modified > +.sizejump_fcontext,.-jump_fcontext > +# Mark that we don't need executable stack. > +.section .note.GNU-stack,"",%progbits > Index: files/make_sparc64_sysv_elf_gas.S > === > RCS file: files/make_sparc64_sysv_elf_gas.S > diff -N files/make_sparc64_sysv_elf_gas.S > --- /dev/null 1 Jan 1970 00:00:00 - > +++ files/make_sparc64_sysv_elf_gas.S 12 Feb 2024 19:12:11 - > @@ -0,0 +1,68 @@ > +/* > + Copyright Claudio Jeker 2024 > + Distributed under the Boost Software License, Version 1.0. > + (See accompanying file LICENSE_1_0.txt or copy at > + http://www.boost.org/LICENSE_1_0.txt) > +*/ > + > +/* > + * fcontext_t *make_fcontext(void *sp, size_t size, void (*fn)(transfer_t)); > + */ > +#define CC64FSZ 176 > +#define BIAS 2047 > +#define FP 112 > +#define SP 128 > +#define I7 136 > + > +.file "make_sparc64_sysv_elf_gas.S" > +.text > +.align 4 > +.global make_fcontext > +.type make_fcontext, %function > +make_fcontext: > + save%sp, -CC64FSZ, %sp > + > + # shift address in %i0 (allocated stack) to lower 16 byte boundary > + and %i0, -0xf, %i0 > + > + # reserve space for two frames on the stack > + # the first frame is for the call
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: o...@cvs.openbsd.org2024/01/10 05:53:35 Modified files: net/powerdns_recursor: Makefile distinfo Added files: net/powerdns_recursor: crates.inc Removed files: net/powerdns_recursor/patches: patch-mtasker_fcontext_cc Log message: Update to PowerDNS Recursor 5.0.1 Tweaks and ok: tb@ semarie@
Re: Update: PowerDNS Recursor 5.0.1
On Wed, Jan 10, 2024 at 11:52:39AM +0100, Theo Buehler wrote: > On Wed, Jan 10, 2024 at 10:53:33AM +0100, Otto Moerbeek wrote: > > Hi, > > > > this is a somewhat large update as PowerDNS Recursor 5 uses Rust code > > (in addition to C++) > > > > So I would like some extra eyes and OK. > > WANTLIB needs fixing (see 'lib-depends-check' in bsd.port.mk). It should > be: > > WANTLIB += ${MODCARGO_WANTLIB} ${COMPILER_LIBCXX} boost_atomic-mt > WANTLIB += boost_context-mt boost_filesystem-mt boost_system-mt crypto > WANTLIB += curl fstrm m nghttp2 nghttp3 ngtcp2 ngtcp2_crypto_quictls > WANTLIB += sodium ssl z > > ${MODCARGO_WANTLIB} is needed for Rust ports to have correct WANTLIB on > all platforms. Currently only sparc64 is special because ports-gcc > statically links against its C++ library - not really an issue here > since boost-md isn't available there. > > You also need to 'cvs rm -f patches/patch-mtasker_fcontext_cc' because > with your diff I get: > > ===> Applying OpenBSD patch patch-mtasker_fcontext_cc > Hmm... Looks like a unified diff to me... > The text leading up to this was: > -- > |Index: mtasker_fcontext.cc > |--- mtasker_fcontext.cc.orig > |+++ mtasker_fcontext.cc > -- > Patching file mtasker_fcontext.cc using Plan A... > Reversed (or previously applied) patch detected! Assume -R? [y] > > With these two fixed > > ok tb > > The CET_ENDBR handling in that now upstreamed patch needs fixing to have > a chance to work with other operating systems. Using endbr64 is not > enough. Other OS also need .gnu.property notes and there is some > assembly to emit that in the compiler-internal cet.h (present if __CET__ > is defined), so the proper portable way is to use _CET_ENDBR and to do > > #ifdef __CET__ > #include > #else > #define _CET_ENDBR > #endif Thanks, I'll look into that. My previous attempts at using cet.h did not work because it assumes it is included from an assembly file, not a C or C++ file. -Otto
Update: PowerDNS Recursor 5.0.1
Hi, this is a somewhat large update as PowerDNS Recursor 5 uses Rust code (in addition to C++) So I would like some extra eyes and OK. (5.0.0 was never released publically) -Otto Index: Makefile === RCS file: /home/cvs/ports/net/powerdns_recursor/Makefile,v diff -u -p -r1.48 Makefile --- Makefile8 Nov 2023 10:08:06 - 1.48 +++ Makefile10 Jan 2024 09:28:10 - @@ -1,8 +1,8 @@ COMMENT= recursive nameserver -V= 4.9.2 +V= 5.0.1 DISTNAME= pdns-recursor-${V} -EXTRACT_SUFX = .tar.bz2 +EXTRACT_SUFX= .tar.bz2 PKGNAME= powerdns-recursor-${V} CATEGORIES=net @@ -29,6 +29,12 @@ LIB_DEPENDS= devel/boost \ net/libfstrm \ security/libsodium +MODULES+= devel/cargo +MODCARGO_CARGOTOML=${WRKSRC}/settings/rust/Cargo.toml +MODCARGO_BUILD=No +MODCARGO_INSTALL= No +MODCARGO_TEST= No + MODULES+= lang/lua MODLUA_VERSION=5.3 MODLUA_SA= Yes @@ -37,8 +43,8 @@ WANTLIB+= ${MODLUA_WANTLIB} SYSCONFDIR=${BASESYSCONFDIR}/pdns -CONFIGURE_STYLE= autoreconf -AUTOCONF_VERSION= 2.69 +CONFIGURE_STYLE= cargo autoreconf +AUTOCONF_VERSION= 2.71 AUTOMAKE_VERSION= 1.16 USE_GMAKE= Yes @@ -62,5 +68,7 @@ post-install: ${WRKSRC}/recursor.conf # concatenate local and upstream sample configs ${INSTALL_DATA} ${WRKSRC}/recursor.conf ${EXAMPLE_DIR} rm ${WRKINST}${SYSCONFDIR}/recursor.conf-dist + +.include "crates.inc" .include Index: crates.inc === RCS file: crates.inc diff -N crates.inc --- /dev/null 1 Jan 1970 00:00:00 - +++ crates.inc 10 Jan 2024 09:28:10 - @@ -0,0 +1,30 @@ +MODCARGO_CRATES += cc 1.0.84 # MIT OR Apache-2.0 +MODCARGO_CRATES += codespan-reporting 0.11.1 # Apache-2.0 +MODCARGO_CRATES += cxx 1.0.110 # MIT OR Apache-2.0 +MODCARGO_CRATES += cxx-build 1.0.110 # MIT OR Apache-2.0 +MODCARGO_CRATES += cxxbridge-flags 1.0.110 # MIT OR Apache-2.0 +MODCARGO_CRATES += cxxbridge-macro 1.0.110 # MIT OR Apache-2.0 +MODCARGO_CRATES += equivalent 1.0.1 # Apache-2.0 OR MIT +MODCARGO_CRATES += hashbrown 0.14.2 # MIT OR Apache-2.0 +MODCARGO_CRATES += indexmap2.1.0 # Apache-2.0 OR MIT +MODCARGO_CRATES += ipnet 2.9.0 # MIT OR Apache-2.0 +MODCARGO_CRATES += itoa1.0.9 # MIT OR Apache-2.0 +MODCARGO_CRATES += libc0.2.150 # MIT OR Apache-2.0 +MODCARGO_CRATES += link-cplusplus 1.0.9 # MIT OR Apache-2.0 +MODCARGO_CRATES += once_cell 1.18.0 # MIT OR Apache-2.0 +MODCARGO_CRATES += proc-macro2 1.0.69 # MIT OR Apache-2.0 +MODCARGO_CRATES += quote 1.0.33 # MIT OR Apache-2.0 +MODCARGO_CRATES += ryu 1.0.15 # Apache-2.0 OR BSL-1.0 +MODCARGO_CRATES += scratch 1.0.7 # MIT OR Apache-2.0 +MODCARGO_CRATES += serde 1.0.192 # MIT OR Apache-2.0 +MODCARGO_CRATES += serde_derive1.0.192 # MIT OR Apache-2.0 +MODCARGO_CRATES += serde_yaml 0.9.27 # MIT OR Apache-2.0 +MODCARGO_CRATES += syn 2.0.39 # MIT OR Apache-2.0 +MODCARGO_CRATES += termcolor 1.4.0 # Unlicense OR MIT +MODCARGO_CRATES += unicode-ident 1.0.12 # (MIT OR Apache-2.0) AND Unicode-DFS-2016 +MODCARGO_CRATES += unicode-width 0.1.11 # MIT/Apache-2.0 +MODCARGO_CRATES += unsafe-libyaml 0.2.9 # MIT +MODCARGO_CRATES += winapi 0.3.9 # MIT/Apache-2.0 +MODCARGO_CRATES += winapi-i686-pc-windows-gnu 0.4.0 # MIT/Apache-2.0 +MODCARGO_CRATES += winapi-util 0.1.6 # Unlicense/MIT +MODCARGO_CRATES += winapi-x86_64-pc-windows-gnu0.4.0 # MIT/Apache-2.0 Index: distinfo === RCS file: /home/cvs/ports/net/powerdns_recursor/distinfo,v diff -u -p -r1.39 distinfo --- distinfo8 Nov 2023 10:08:06 - 1.39 +++ distinfo10 Jan 2024 09:28:10 - @@ -1,2 +1,62 @@ -SHA256 (pdns-recursor-4.9.2.tar.bz2) = TLgYBFjs+1KKPZo0uihEts0u1pyhxGHd4koOvWaCkUQ= -SIZE (pdns-recursor-4.9.2.tar.bz2) = 1562694 +SHA256 (cargo/cc-1.0.84.tar.gz) = D458kK+tiQSEohZT0ItuIJrjR3D7XuKY+caZ/MHlyFY= +SHA256 (cargo/codespan-reporting-0.11.1.tar.gz) = NTgnDTPMZpZQxLCThIRQ04De8QwzHTjHaONMrIBXbm4= +SHA256 (cargo/cxx-1.0.110.tar.gz) = cSnjQQNOy5QMkHKBfNkAeXTqaWhE/E3VgtwWU6f74ug= +SHA256 (cargo/cxx-build-1.0.110.tar.gz) = oqJPP1+O7XGTbyHlcENvAk9cLiVij3SWqnzNA7kBCdU= +SHA256 (cargo/cxxbridge-flags-1.0.110.tar.gz) = Bv3Rd/xhBQ1j9n9b1jUfrGq1UmaU6o41nNnNO3WFf0Q= +SHA256 (cargo/cxxbridge-macro-1.0.110.tar.gz) = WHZj3V+z0QkyyK7P58hE2xvPCu6T7qsI+sE9wSEsLn8= +SHA256 (cargo/equivalent-1.0.1.tar.gz) =
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: o...@cvs.openbsd.org2023/12/21 03:55:19 Modified files: net/powerdns : Makefile distinfo Log message: Update to PowerDNS Authoritative Server 4.8.4
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: o...@cvs.openbsd.org2023/12/15 06:11:55 Modified files: net/dnsdist: Makefile distinfo Log message: Update to dnsdist 1.8.3
Re: devel/boost syscall(2) removal
On Wed, Nov 15, 2023 at 04:34:01PM -0500, Brad Smith wrote: > On 11/15/2023 2:26 AM, Otto Moerbeek wrote: > > On Tue, Nov 07, 2023 at 11:55:55AM +0100, Otto Moerbeek wrote: > > > > > On Sat, Nov 04, 2023 at 09:45:01PM +0100, Otto Moerbeek wrote: > > > > > > > On Sat, Nov 04, 2023 at 03:50:33PM -0400, Brad Smith wrote: > > > > > > > > > On 2023-11-04 4:07 a.m., Otto Moerbeek wrote: > > > > > > > > > > > On Fri, Nov 03, 2023 at 09:03:20PM -0400, Brad Smith wrote: > > > > > > > > > > > > > On Sun, Oct 29, 2023 at 08:13:45AM -0400, Brad Smith wrote: > > > > > > > > On Sun, Oct 29, 2023 at 10:52:39AM +, Stuart Henderson > > > > > > > > wrote: > > > > > > > > > Doesn't hurt but we probably don't need #ifdef around > > > > > > > > > SYS_getrandom tbh. > > > > > > > > > > > > > > > > > > Has anyone looked at updating boost recently? It would be a > > > > > > > > > good time in our > > > > > > > > > release cycle and I can do an i386 bulk if anyone has a diff > > > > > > > > > handy. > > > > > > > > > > > > > > > > > > -- > > > > > > > > >Sent from a phone, apologies for poor formatting. > > > > > > > > The existing Boost::Context diffs do not apply as is, there > > > > > > > > were some changes there. > > > > > > > > Someone more familiar with this neeeds to look at what needs to > > > > > > > > be done for i386. > > > > > > > > > > > > > > > > https://github.com/boostorg/context/commit/2f0b5034b72cbb858849bd744ffb06df04521e5a > > > > > > > > > > > > > > > > https://github.com/boostorg/context/commit/e286649f08a6312ba3d88aa3eafa179ec3eab163 > > > > > > > > https://github.com/boostorg/context/commit/0eb42637058741bef2456e0fb165e02d47670c8a > > > > > > > Here is an updated diff with the reverted 1.80 ASM. > > > > > > That does not look like the correct verisons of the assmebly. What > > > > > > we > > > > > > need is the 1.80 assembly *after* the patches in the 1.80 port have > > > > > > been applied. > > > > > The original patches are still there. > > > > Ah, right, sorry for the noise. > > > > > > > > -Otto > > > > > > > And tested with PowerDNS Recursor on i386 to work as expected. > > > > > > -Otto > > > > > Any reason to not commit this? > > > > -Otto > > To be clear it's ready for a bulk build. There is fallout with pretty much > every > Boost release. We need to find out what breaks first. With the current LLVM > 16 > work it's probably best to wait at least another week for things to settle > before > doing that. OK, I'll have some patience :-). Do note that one issue in boost::threads that clang16 trips on is solved by 1.83. -Otto
Re: devel/boost syscall(2) removal
On Tue, Nov 07, 2023 at 11:55:55AM +0100, Otto Moerbeek wrote: > On Sat, Nov 04, 2023 at 09:45:01PM +0100, Otto Moerbeek wrote: > > > On Sat, Nov 04, 2023 at 03:50:33PM -0400, Brad Smith wrote: > > > > > On 2023-11-04 4:07 a.m., Otto Moerbeek wrote: > > > > > > > On Fri, Nov 03, 2023 at 09:03:20PM -0400, Brad Smith wrote: > > > > > > > > > On Sun, Oct 29, 2023 at 08:13:45AM -0400, Brad Smith wrote: > > > > > > On Sun, Oct 29, 2023 at 10:52:39AM +, Stuart Henderson wrote: > > > > > > > Doesn't hurt but we probably don't need #ifdef around > > > > > > > SYS_getrandom tbh. > > > > > > > > > > > > > > Has anyone looked at updating boost recently? It would be a good > > > > > > > time in our > > > > > > > release cycle and I can do an i386 bulk if anyone has a diff > > > > > > > handy. > > > > > > > > > > > > > > -- > > > > > > > Sent from a phone, apologies for poor formatting. > > > > > > The existing Boost::Context diffs do not apply as is, there were > > > > > > some changes there. > > > > > > Someone more familiar with this neeeds to look at what needs to be > > > > > > done for i386. > > > > > > > > > > > > https://github.com/boostorg/context/commit/2f0b5034b72cbb858849bd744ffb06df04521e5a > > > > > > > > > > > > https://github.com/boostorg/context/commit/e286649f08a6312ba3d88aa3eafa179ec3eab163 > > > > > > https://github.com/boostorg/context/commit/0eb42637058741bef2456e0fb165e02d47670c8a > > > > > Here is an updated diff with the reverted 1.80 ASM. > > > > That does not look like the correct verisons of the assmebly. What we > > > > need is the 1.80 assembly *after* the patches in the 1.80 port have > > > > been applied. > > > > > > The original patches are still there. > > > > Ah, right, sorry for the noise. > > > > -Otto > > > > And tested with PowerDNS Recursor on i386 to work as expected. > > -Otto > Any reason to not commit this? -Otto
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: o...@cvs.openbsd.org2023/11/08 03:08:06 Modified files: net/powerdns_recursor: Makefile distinfo Log message: Update to PowerDNS Recursor 4.9.2
Re: devel/boost syscall(2) removal
On Sat, Nov 04, 2023 at 09:45:01PM +0100, Otto Moerbeek wrote: > On Sat, Nov 04, 2023 at 03:50:33PM -0400, Brad Smith wrote: > > > On 2023-11-04 4:07 a.m., Otto Moerbeek wrote: > > > > > On Fri, Nov 03, 2023 at 09:03:20PM -0400, Brad Smith wrote: > > > > > > > On Sun, Oct 29, 2023 at 08:13:45AM -0400, Brad Smith wrote: > > > > > On Sun, Oct 29, 2023 at 10:52:39AM +, Stuart Henderson wrote: > > > > > > Doesn't hurt but we probably don't need #ifdef around SYS_getrandom > > > > > > tbh. > > > > > > > > > > > > Has anyone looked at updating boost recently? It would be a good > > > > > > time in our > > > > > > release cycle and I can do an i386 bulk if anyone has a diff handy. > > > > > > > > > > > > -- > > > > > > Sent from a phone, apologies for poor formatting. > > > > > The existing Boost::Context diffs do not apply as is, there were some > > > > > changes there. > > > > > Someone more familiar with this neeeds to look at what needs to be > > > > > done for i386. > > > > > > > > > > https://github.com/boostorg/context/commit/2f0b5034b72cbb858849bd744ffb06df04521e5a > > > > > > > > > > https://github.com/boostorg/context/commit/e286649f08a6312ba3d88aa3eafa179ec3eab163 > > > > > https://github.com/boostorg/context/commit/0eb42637058741bef2456e0fb165e02d47670c8a > > > > Here is an updated diff with the reverted 1.80 ASM. > > > That does not look like the correct verisons of the assmebly. What we > > > need is the 1.80 assembly *after* the patches in the 1.80 port have > > > been applied. > > > > The original patches are still there. > > Ah, right, sorry for the noise. > > -Otto > And tested with PowerDNS Recursor on i386 to work as expected. -Otto
Re: devel/boost syscall(2) removal
On Sat, Nov 04, 2023 at 03:50:33PM -0400, Brad Smith wrote: > On 2023-11-04 4:07 a.m., Otto Moerbeek wrote: > > > On Fri, Nov 03, 2023 at 09:03:20PM -0400, Brad Smith wrote: > > > > > On Sun, Oct 29, 2023 at 08:13:45AM -0400, Brad Smith wrote: > > > > On Sun, Oct 29, 2023 at 10:52:39AM +, Stuart Henderson wrote: > > > > > Doesn't hurt but we probably don't need #ifdef around SYS_getrandom > > > > > tbh. > > > > > > > > > > Has anyone looked at updating boost recently? It would be a good time > > > > > in our > > > > > release cycle and I can do an i386 bulk if anyone has a diff handy. > > > > > > > > > > -- > > > > > Sent from a phone, apologies for poor formatting. > > > > The existing Boost::Context diffs do not apply as is, there were some > > > > changes there. > > > > Someone more familiar with this neeeds to look at what needs to be done > > > > for i386. > > > > > > > > https://github.com/boostorg/context/commit/2f0b5034b72cbb858849bd744ffb06df04521e5a > > > > > > > > https://github.com/boostorg/context/commit/e286649f08a6312ba3d88aa3eafa179ec3eab163 > > > > https://github.com/boostorg/context/commit/0eb42637058741bef2456e0fb165e02d47670c8a > > > Here is an updated diff with the reverted 1.80 ASM. > > That does not look like the correct verisons of the assmebly. What we > > need is the 1.80 assembly *after* the patches in the 1.80 port have > > been applied. > > The original patches are still there. Ah, right, sorry for the noise. -Otto
Re: devel/boost syscall(2) removal
On Sun, Oct 29, 2023 at 08:13:43AM -0400, Brad Smith wrote: > On Sun, Oct 29, 2023 at 10:52:39AM +, Stuart Henderson wrote: > > Doesn't hurt but we probably don't need #ifdef around SYS_getrandom tbh. > > > > Has anyone looked at updating boost recently? It would be a good time in our > > release cycle and I can do an i386 bulk if anyone has a diff handy. > > > > -- > > Sent from a phone, apologies for poor formatting. > > The existing Boost::Context diffs do not apply as is, there were some changes > there. > Someone more familiar with this neeeds to look at what needs to be done for > i386. > > https://github.com/boostorg/context/commit/2f0b5034b72cbb858849bd744ffb06df04521e5a This one is a bit surprising. PowerDNS Recursor has been compiled with -fstack-protector for ages, and has no issues with the existing (patched to conform to out calling conventions) context switch assembly in boost 1.80 on i386. I'm setting up a i386 VM to check things further. I imagine we might be able to take the easy way and continue to use the asm from the 1.80 port in boost 1.83. -Otto > > https://github.com/boostorg/context/commit/e286649f08a6312ba3d88aa3eafa179ec3eab163 > https://github.com/boostorg/context/commit/0eb42637058741bef2456e0fb165e02d47670c8a > > > Index: Makefile > === > RCS file: /home/cvs/ports/devel/boost/Makefile,v > diff -u -p -u -p -r1.136 Makefile > --- Makefile 28 Oct 2023 14:03:31 - 1.136 > +++ Makefile 29 Oct 2023 10:56:23 - > @@ -8,17 +8,16 @@ COMMENT-md= machine-dependent libraries > # if updating >=1.81.0 review ports using boost_context that have USE_NOBTCFI > # set and see if it can be removed (definite problems seen with > pdns_recursor); > # sqlite3 /usr/local/share/sqlports "select * from wantlib where value like > 'boost_context%'" > -VERSION= 1.80.0 > -EPOCH = 0 > +VERSION= 1.83.0 > DISTNAME=boost_${VERSION:S/./_/g} > PKGNAME= boost-${VERSION} > -REVISION-main= 3 > +EPOCH= 0 > CATEGORIES= devel > SITES= > https://boostorg.jfrog.io/artifactory/main/release/${VERSION}/source/ > EXTRACT_SUFX=.tar.bz2 > FIX_EXTRACT_PERMISSIONS =Yes > > -SO_VERSION= 21.0 > +SO_VERSION= 22.0 > BOOST_LIBS= boost_atomic-mt \ > boost_chrono-mt boost_chrono \ > boost_container-mt boost_container \ > @@ -53,6 +52,7 @@ BOOST_LIBS= boost_atomic-mt \ > boost_timer-mt boost_timer \ > boost_type_erasure-mt boost_type_erasure \ > boost_unit_test_framework-mt boost_unit_test_framework \ > + boost_url-mt boost_url \ > boost_wserialization-mt boost_wserialization \ > boost_wave-mt boost_wave > > Index: distinfo > === > RCS file: /home/cvs/ports/devel/boost/distinfo,v > diff -u -p -u -p -r1.21 distinfo > --- distinfo 29 Aug 2022 17:28:18 - 1.21 > +++ distinfo 29 Oct 2023 10:57:07 - > @@ -1,2 +1,2 @@ > -SHA256 (boost_1_80_0.tar.bz2) = HhlWXYLkO8WSCaFo9ayJnTukcdVcdhDGd9TM8snFAMA= > -SIZE (boost_1_80_0.tar.bz2) = 115869904 > +SHA256 (boost_1_83_0.tar.bz2) = ZHjt/i8zBRJ8/+jK9z6gF2xTdp9L8Vhb4jfrMHmMO44= > +SIZE (boost_1_83_0.tar.bz2) = 122892751 > Index: patches/patch-Jamroot > === > RCS file: /home/cvs/ports/devel/boost/patches/patch-Jamroot,v > diff -u -p -u -p -r1.18 patch-Jamroot > --- patches/patch-Jamroot 29 Aug 2022 17:28:18 - 1.18 > +++ patches/patch-Jamroot 8 Aug 2023 00:39:33 - > @@ -3,7 +3,7 @@ Index: Jamroot > +++ Jamroot > @@ -146,11 +146,12 @@ import tools/boost\_install/boost-install ; > path-constant BOOST_ROOT : . ; > - constant BOOST_VERSION : 1.80.0 ; > + constant BOOST_VERSION : 1.83.0 ; > constant BOOST_JAMROOT_MODULE : $(__name__) ; > +constant SO_VERSION : ${SO_VERSION} ; > > Index: patches/patch-boost_config_stdlib_libcpp_hpp > === > RCS file: patches/patch-boost_config_stdlib_libcpp_hpp > diff -N patches/patch-boost_config_stdlib_libcpp_hpp > --- patches/patch-boost_config_stdlib_libcpp_hpp 13 Nov 2022 11:42:23 > - 1.1 > +++ /dev/null 1 Jan 1970 00:00:00 - > @@ -1,20 +0,0 @@ > -The std lib unary/binary_function base classes are deprecated/removed from > libcpp15. > -https://github.com/boostorg/container_hash/issues/24 > - > -Index: boost/config/stdlib/libcpp.hpp > boost/config/stdlib/libcpp.hpp.orig > -+++ boost/config/stdlib/libcpp.hpp > -@@ -168,4 +168,13 @@ > - # define BOOST_NO_CXX14_HDR_SHARED_MUTEX > - #endif > - > -+#if _LIBCPP_VERSION >= 15000 > -+// > -+// Unary function is now deprecated in C++11 and later: > -+// > -+#if __cplusplus >= 201103L > -+#define BOOST_NO_CXX98_FUNCTION_BASE > -+#endif > -+#endif > -+ > - // --- end --- >
net/powerdns_recursor fix boost context switching wrt IBT
Hi, After some digging it turns out that boost 1.80 does have the required endbr64 instructions in its context switch assembly code. But that is not enough, to make PowerDNS Recursor work on my new laptop I had to place three endbr64 instructions after jump_fcontext() calls in pdns_recursor itself. This way I could remove the USE_NOBTCFI marker. So far only lightly tested but I thought I'd share this result as others may benefit from this insight. -Otto Index: Makefile === RCS file: /home/cvs/ports/net/powerdns_recursor/Makefile,v retrieving revision 1.46 diff -u -p -r1.46 Makefile --- Makefile30 Sep 2023 10:41:00 - 1.46 +++ Makefile29 Oct 2023 08:04:58 - @@ -4,7 +4,7 @@ V= 4.9.1 DISTNAME= pdns-recursor-${V} EXTRACT_SUFX = .tar.bz2 PKGNAME= powerdns-recursor-${V} -REVISION= 0 +REVISION= 1 CATEGORIES=net @@ -16,9 +16,6 @@ PERMIT_PACKAGE= Yes WANTLIB+= ${COMPILER_LIBCXX} boost_context-mt boost_filesystem-mt WANTLIB+= boost_system-mt boost_thread-mt c crypto curl fstrm m WANTLIB+= sodium ssl z - -# boost_context 1.80.0 doesn't have landing pads (1.81.0 should fix) -USE_NOBTCFI= Yes SITES= https://downloads.powerdns.com/releases/ MAINTAINER=Otto Moerbeek Index: patches/patch-mtasker_fcontext_cc === RCS file: patches/patch-mtasker_fcontext_cc diff -N patches/patch-mtasker_fcontext_cc --- /dev/null 1 Jan 1970 00:00:00 - +++ patches/patch-mtasker_fcontext_cc 29 Oct 2023 08:04:58 - @@ -0,0 +1,48 @@ +Index: mtasker_fcontext.cc +--- mtasker_fcontext.cc.orig mtasker_fcontext.cc +@@ -32,6 +32,20 @@ using boost::context::make_fcontext; + using boost::context::detail::make_fcontext; + #endif /* BOOST_VERSION < 106100 */ + ++# ifdef __LP64__ ++# if __CET__ & 0x1 ++#define _CET_ENDBR __asm("endbr64") ++# else ++#define _CET_ENDBR ++# endif ++# else ++# if __CET__ & 0x1 ++#define _CET_ENDBR __asm("endbr32") ++# else ++#define _CET_ENDBR ++# endif ++# endif ++ + #ifdef PDNS_USE_VALGRIND + #include + #endif /* PDNS_USE_VALGRIND */ +@@ -132,6 +146,7 @@ extern "C" + static_cast(args->prev_ctx), 0); + #else + transfer_t res = jump_fcontext(t.fctx, 0); ++_CET_ENDBR; + /* we got switched back from pdns_swapcontext() */ + if (res.data) { + /* if res.data is not a nullptr, it holds a pointer to the context +@@ -203,6 +218,7 @@ void pdns_swapcontext(pdns_ucontext_t& __restrict octx + std::rethrow_exception(origctx->exception); + #else + transfer_t res = jump_fcontext(static_cast(ctx.uc_mcontext), _mcontext); ++ _CET_ENDBR; + if (res.data) { + /* if res.data is not a nullptr, it holds a pointer to the context +we just switched from, and we need to fill it to be able to +@@ -235,6 +251,7 @@ void pdns_makecontext(pdns_ucontext_t& ctx, std::funct + #else + transfer_t res = jump_fcontext(static_cast(ctx.uc_mcontext), + ); ++ _CET_ENDBR; + /* back from threadwrapper, updating the context */ + ctx.uc_mcontext = res.fctx; + #endif
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: o...@cvs.openbsd.org2023/10/24 03:51:00 Modified files: net: Makefile Log message: Add quiche
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: o...@cvs.openbsd.org2023/10/24 03:49:05 Log message: Add quiche, a rust crate providing implementations of the QUIC and HTTP/3 protocols; ok and help tb@ Status: Vendor Tag: otto Release Tags: otto_20231024 N ports/net/quiche/Makefile N ports/net/quiche/crates.inc N ports/net/quiche/distinfo N ports/net/quiche/pkg/PLIST N ports/net/quiche/pkg/DESCR No conflicts created by this import
Re: new port: net/quiche (QUIC library)
On Tue, Oct 24, 2023 at 10:58:55AM +0200, Theo Buehler wrote: > On Tue, Oct 24, 2023 at 10:42:56AM +0200, Theo Buehler wrote: > > > Updated tarball, now also portcheck clean. > > > > MOD_ENV += ${MODCARGO_ENV} > > QUICHE_BSSL_PATH=${LOCALBAS}/eboringssl > > > > Typo: LOCALBASE > > > > There's no BUILD_DEPENDS on security/boringssl/head which would > > need to be present for this to work. > > > > Is boringssl actually used by the build? I can't spot anything in the > > build log (there are the boring and boring-sys crate, but they aren't > > built) and the deps directory is also empty. > > > > The port seems to produce the same libquiche.a whether boringssl is > > installed or not and whether the typo above is fixed or not. > > > > Is there something missing or does DoQ already work with dnsdist? > > > > If DoQ already works then we can leave the port as it is modulo removal > > of the above line. > > > > If not, we will probably need the boringssl-boring-crate feature and > > teach it to use the eboringssl from ports. > > > > libquiche.a definitely tries to use BoringSSL API (which libcrypto > doesn't provide), for example > > U EVP_AEAD_CTX_seal_scatter > > the bindings are made in quiche/src/tls.rs and quiche/src/crypto.rs. > > Is the upcoming dnsdist port going to be linked statically against > libquiche and eboringssl? Yes, that's the plan. For now I unconditionally enable DoQ, that may change if I'm going to make a regular dnsdist and a doq flavour. I can send you my wip port if you want (using dnsdist-1.9.0-alpha3). -Otto
Re: new port: net/quiche (QUIC library)
On Tue, Oct 24, 2023 at 10:42:56AM +0200, Theo Buehler wrote: > > Updated tarball, now also portcheck clean. > > MOD_ENV += ${MODCARGO_ENV} > QUICHE_BSSL_PATH=${LOCALBAS}/eboringssl > > Typo: LOCALBASE > > There's no BUILD_DEPENDS on security/boringssl/head which would > need to be present for this to work. > > Is boringssl actually used by the build? I can't spot anything in the > build log (there are the boring and boring-sys crate, but they aren't > built) and the deps directory is also empty. > > The port seems to produce the same libquiche.a whether boringssl is > installed or not and whether the typo above is fixed or not. > > Is there something missing or does DoQ already work with dnsdist? > > If DoQ already works then we can leave the port as it is modulo removal > of the above line. > > If not, we will probably need the boringssl-boring-crate feature and > teach it to use the eboringssl from ports. Hmm, well spotted!. I suppose it works because I'm linking my dnsdist wip port agains boringssl. I have no clue yet why my previous attempts at building quiche without specifying QUICHE_BSSL_PATH did not work. What I observed is it was trying to build its own boringssl, which failed. -Otto quiche.tgz Description: application/tar-gz
Re: new port: net/quiche (QUIC library)
On Mon, Oct 23, 2023 at 09:12:06PM +0200, Theo Buehler wrote: > On Mon, Oct 23, 2023 at 09:06:28PM +0200, Otto Moerbeek wrote: > > On Mon, Oct 23, 2023 at 08:39:51PM +0200, Theo Buehler wrote: > > > > > On Mon, Oct 23, 2023 at 05:15:31PM +0200, Otto Moerbeek wrote: > > > > Hi, > > > > > > > > upcomng dnsdist has incoming DoQ (DNS over QUIC) supoport using the > > > > quiche library. > > > > > > Please add > > > > > > # ring-v0.16.20 does not support those archs > > > NOT_FOR_ARCHS = powerpc64 riscv64 sparc64 > > > > > > at the top of the makefile. ring 0.17 adds powerpc64 support, but that > > > isn't out yet. > > > > > > > So here's a port of that library. Since it links with boringssl, which > > > > is static-only, I made this one static-only as well. That avoid quite > > > > some issues. > > > > > > rust knocks out quite a few architectures, then ring removes powerpc64, > > > riscv64 and sparc64 and then it doesn't really matter anymore that > > > boringssl removes all big-endian architectures. > > > > > > Is it possible to make DoQ optional in dnsdist so that it keeps being > > > available on less mainstream architectures? Or is it not worth it? > > > > Not supporting DoQ in dnsdist is easy from the configure point of > > view. But I'm also building my wip dnsdist post against boringssl as > > I think it is impossible (or at least undesirable) to have two > > instances of lib*ssl in one executable. > > I was going to ask about that when seeing the BoringSSL port. > > > So if we want to have two flavours of dnsdist I think that would mean > > that one flavour is linked against libressl and one against boringssl. > > Don't know if I want to maintain something like that. I have to think > > about it. > > It's your choice. > > > > > I'm not using the crate install command, as the crate does not seem to > > > > be setup in such a way to allow for that. > > > > > > Yes, that seems the right choice, although I think you want > > > > > > MODCARGO_INSTALL = No > > > > > > rather than NO_INSTALL > > > > > > > As I'm still a port-newbie, any feedbakc is very welcome. > > > > > > Apart from the above, the port looks good to me. I haven't managed to > > > figure out how to run tests. > > > > Thanks for the review. > > > > While working on this and a make build with an updated cvs tree I'm now > > seeing > > > > syntax error at /usr/libdata/perl5/OpenBSD/Subst.pm line 54, near "" $v")" > > > > This looks fishy, have to figrue out if this is a broken tree or > > something else. > > Revert this commit: > > https://github.com/openbsd/src/commit/41302357a6f17bff2670a6860ac5dfa34f78e40d > > I'm going to do this a bit later if espie doesn't show up. Updated tarball, now also portcheck clean. -Otto quiche.tgz Description: application/tar-gz
Re: new port: net/quiche (QUIC library)
On Mon, Oct 23, 2023 at 08:39:51PM +0200, Theo Buehler wrote: > On Mon, Oct 23, 2023 at 05:15:31PM +0200, Otto Moerbeek wrote: > > Hi, > > > > upcomng dnsdist has incoming DoQ (DNS over QUIC) supoport using the > > quiche library. > > Please add > > # ring-v0.16.20 does not support those archs > NOT_FOR_ARCHS = powerpc64 riscv64 sparc64 > > at the top of the makefile. ring 0.17 adds powerpc64 support, but that > isn't out yet. > > > So here's a port of that library. Since it links with boringssl, which > > is static-only, I made this one static-only as well. That avoid quite > > some issues. > > rust knocks out quite a few architectures, then ring removes powerpc64, > riscv64 and sparc64 and then it doesn't really matter anymore that > boringssl removes all big-endian architectures. > > Is it possible to make DoQ optional in dnsdist so that it keeps being > available on less mainstream architectures? Or is it not worth it? Not supporting DoQ in dnsdist is easy from the configure point of view. But I'm also building my wip dnsdist post against boringssl as I think it is impossible (or at least undesirable) to have two instances of lib*ssl in one executable. So if we want to have two flavours of dnsdist I think that would mean that one flavour is linked against libressl and one against boringssl. Don't know if I want to maintain something like that. I have to think about it. > > > I'm not using the crate install command, as the crate does not seem to > > be setup in such a way to allow for that. > > Yes, that seems the right choice, although I think you want > > MODCARGO_INSTALL = No > > rather than NO_INSTALL > > > As I'm still a port-newbie, any feedbakc is very welcome. > > Apart from the above, the port looks good to me. I haven't managed to > figure out how to run tests. Thanks for the review. While working on this and a make build with an updated cvs tree I'm now seeing syntax error at /usr/libdata/perl5/OpenBSD/Subst.pm line 54, near "" $v")" This looks fishy, have to figrue out if this is a broken tree or something else. -Otto
new port: net/quiche (QUIC library)
Hi, upcomng dnsdist has incoming DoQ (DNS over QUIC) supoport using the quiche library. So here's a port of that library. Since it links with boringssl, which is static-only, I made this one static-only as well. That avoid quite some issues. I'm not using the crate install command, as the crate does not seem to be setup in such a way to allow for that. As I'm still a port-newbie, any feedbakc is very welcome. -Otto quiche.tgz Description: application/tar-gz
Re: Valgrind: Detect access outside the range of malloc.
On Tue, Oct 17, 2023 at 05:41:06PM +0900, Masato Asou wrote: > From: Otto Moerbeek > Date: Tue, 17 Oct 2023 07:52:12 +0200 > > > On Tue, Oct 17, 2023 at 07:33:47AM +0900, Masato Asou wrote: > > > >> From: Masato Asou > >> Date: Fri, 29 Sep 2023 19:39:12 +0900 (JST) > >> > >> ping > > > > Sorry for not replying earlier. > > No problem :-) > > > The locations are reported > > correctly now. But there are spurious reports of accesses of a 8 byte > > value on each system call. (tested on amd64). > > > > -Otto > > > > ==83949== Use of uninitialised value of size 8 > > ==83949==at 0x4A0CEC2: write (sys/w_write.c:28) > > ==83949==by 0x4A1B322: __sflush (stdio/fflush.c:80) > > ==83949==by 0x49B21E8: __sfvwrite (stdio/fvwrite.c:188) > > ==83949==by 0x4A26FED: __sprint (stdio/vfprintf.c:108) > > ==83949==by 0x4A26FED: __vfprintf (stdio/vfprintf.c:1062) > > ==83949==by 0x4A23CA5: vfprintf (stdio/vfprintf.c:263) > > ==83949==by 0x49EDB54: printf (stdio/printf.c:44) > > ==83949==by 0x109B90: main (m.c:16) > > I know this issue. This is due to the references to the FS (F > segment) registers TCB and TIB made by ENTER_CANCEL_POINT and > LEAVE_CANCEL_POINT during the system call invocation. > > I think this problem should be solved in the `.supp' files (valgrind > --suppressions=). I am working on a `.supp' file for > OpenBSD. However, this cannot be created quickly yet. Wouldn't it be > better to commit the `.supp' file as a separate modification? Yes, a separate commit is fine. Would this suppression file end up in pkg-readmes, or does valgrind have a place to store a system-wide default supression file? -Otto > -- > ASOU Masato > > >> > >> > I have fixed the bug you reported when symbols were not printed. > >> > > >> > As shown below, if the lower 12 bits of Virtual Address are 0 and File > >> > Size is 0, then the a.out symbol will not be printed. > >> > > >> > $ readelf -l a.out > >> > > >> > Program Headers: > >> > Type Offset VirtAddr PhysAddr > >> > FileSizMemSiz Flags Align > >> > > >> > LOAD 0x1000 0x3000 0x3000 > >> > 0x 0x0055 RW 1000 > >> > > >> > From: Otto Moerbeek > >> > Date: Tue, 5 Sep 2023 07:40:18 +0200 > >> > > >> >> On Tue, Sep 05, 2023 at 09:38:40AM +0900, Masato Asou wrote: > >> >> > >> >>> hi, > >> >>> > >> >>> I have fixed a bug in Valgrind. The Valgrind could not detect access > >> >>> outside the range of malloc. > >> >>> > >> >>> comments, ok? > >> >> > >> >> This works much better that before. Thanks for working on this! > >> >> > >> >> It now detects out of bounds read and writes correctly. A double > >> >> free is detected. > >> >> Also, the spurious reports for accesses to errno are gone. > >> >> > >> >> It does not report proper locations though, even if I compile my test > >> >> program with -g: > >> >> > >> >> ==23912== Invalid read of size 1 > >> >> ==23912==at 0x109B5D: ??? (in ./a.out) > >> >> ==23912==by 0x1098D1: ??? (in ./a.out) > >> >> ==23912== Address 0x4a42840 is 0 bytes after a block of size 10,240 > >> >> alloc'd > >> >> ==23912==at 0x493A3A9: malloc (vg_replace_malloc.c:435) > >> >> ==23912==by 0x109B32: ??? (in ./a.out) > >> >> ==23912==by 0x1098D1: ??? (in ./a.out) > >> >> ==23912== > >> >> 0 > >> > > >> > The a.out symbol is now printed as shown below: > >> > > >> > $ cat malloctest.c > >> > #include > >> > #include > >> > > >> > int > >> > main(int argc, char *argv[]) > >> > { > >> > size_t sz = atoi(argv[1]); > >> > unsigned char *p = malloc(sz); > >> > printf("%p\n", p); > >> > p[sz] = 0; > >> > printf("%x\n", p[sz]); > >> > free(p); > >> > free(p); > >> > return 0; > >> > } > >> > $ c
Re: Valgrind: Detect access outside the range of malloc.
On Tue, Oct 17, 2023 at 07:33:47AM +0900, Masato Asou wrote: > From: Masato Asou > Date: Fri, 29 Sep 2023 19:39:12 +0900 (JST) > > ping Sorry for not replying earlier. The locations are reported correctly now. But there are spurious reports of accesses of a 8 byte value on each system call. (tested on amd64). -Otto ==83949== Use of uninitialised value of size 8 ==83949==at 0x4A0CEC2: write (sys/w_write.c:28) ==83949==by 0x4A1B322: __sflush (stdio/fflush.c:80) ==83949==by 0x49B21E8: __sfvwrite (stdio/fvwrite.c:188) ==83949==by 0x4A26FED: __sprint (stdio/vfprintf.c:108) ==83949==by 0x4A26FED: __vfprintf (stdio/vfprintf.c:1062) ==83949==by 0x4A23CA5: vfprintf (stdio/vfprintf.c:263) ==83949==by 0x49EDB54: printf (stdio/printf.c:44) ==83949==by 0x109B90: main (m.c:16) > > > I have fixed the bug you reported when symbols were not printed. > > > > As shown below, if the lower 12 bits of Virtual Address are 0 and File > > Size is 0, then the a.out symbol will not be printed. > > > > $ readelf -l a.out > > > > Program Headers: > > Type Offset VirtAddr PhysAddr > > FileSizMemSiz Flags Align > > > > LOAD 0x1000 0x3000 0x3000 > > 0x 0x0055 RW 1000 > > > > From: Otto Moerbeek > > Date: Tue, 5 Sep 2023 07:40:18 +0200 > > > >> On Tue, Sep 05, 2023 at 09:38:40AM +0900, Masato Asou wrote: > >> > >>> hi, > >>> > >>> I have fixed a bug in Valgrind. The Valgrind could not detect access > >>> outside the range of malloc. > >>> > >>> comments, ok? > >> > >> This works much better that before. Thanks for working on this! > >> > >> It now detects out of bounds read and writes correctly. A double > >> free is detected. > >> Also, the spurious reports for accesses to errno are gone. > >> > >> It does not report proper locations though, even if I compile my test > >> program with -g: > >> > >> ==23912== Invalid read of size 1 > >> ==23912==at 0x109B5D: ??? (in ./a.out) > >> ==23912==by 0x1098D1: ??? (in ./a.out) > >> ==23912== Address 0x4a42840 is 0 bytes after a block of size 10,240 > >> alloc'd > >> ==23912==at 0x493A3A9: malloc (vg_replace_malloc.c:435) > >> ==23912==by 0x109B32: ??? (in ./a.out) > >> ==23912==by 0x1098D1: ??? (in ./a.out) > >> ==23912== > >> 0 > > > > The a.out symbol is now printed as shown below: > > > > $ cat malloctest.c > > #include > > #include > > > > int > > main(int argc, char *argv[]) > > { > > size_t sz = atoi(argv[1]); > > unsigned char *p = malloc(sz); > > printf("%p\n", p); > > p[sz] = 0; > > printf("%x\n", p[sz]); > > free(p); > > free(p); > > return 0; > > } > > $ cc -g malloctest.c > > $ valgrind ./a.out 128 > > ==21074== Memcheck, a memory error detector > > ==21074== Copyright (C) 2002-2022, and GNU GPL'd, by Julian Seward et al. > > ==21074== Using Valgrind-3.21.0 and LibVEX; rerun with -h for copyright info > > ==21074== Command: ./a.out 128 > > ==21074== > > ==21074== Use of uninitialised value of size 8 > > ==21074==at 0x49C34CD: write (sys/w_write.c:26) > > ==21074==by 0x4A150D2: __sflush (stdio/fflush.c:80) > > ==21074==by 0x49B99C8: __sfvwrite (stdio/fvwrite.c:191) > > ==21074==by 0x4979E4D: __sprint (stdio/vfprintf.c:108) > > ==21074==by 0x4979E4D: __vfprintf (stdio/vfprintf.c:1064) > > ==21074==by 0x4976B05: vfprintf (stdio/vfprintf.c:263) > > ==21074==by 0x49D6434: printf (stdio/printf.c:44) > > ==21074==by 0x109B48: main (malloctest.c:9) > > ==21074== > > ==21074== Use of uninitialised value of size 8 > > ==21074==at 0x49C34DE: write (sys/w_write.c:26) > > ==21074==by 0x4A150D2: __sflush (stdio/fflush.c:80) > > ==21074==by 0x49B99C8: __sfvwrite (stdio/fvwrite.c:191) > > ==21074==by 0x4979E4D: __sprint (stdio/vfprintf.c:108) > > ==21074==by 0x4979E4D: __vfprintf (stdio/vfprintf.c:1064) > > ==21074==by 0x4976B05: vfprintf (stdio/vfprintf.c:263) > > ==21074==by 0x49D6434: printf (stdio/printf.c:44) > > ==21074==by 0x109B48: main (malloctest.c:9) > > > >> ==23912== Invalid free() / delete / delete[] / realloc() > >> ==23912==at 0x49
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: o...@cvs.openbsd.org2023/10/12 00:13:38 Modified files: net/dnsdist: Makefile distinfo Log message: Update to dnsdist 1.8.2, using the h2o version with rapid reset fixes. ok and help sthen@
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: o...@cvs.openbsd.org2023/10/12 00:12:33 Modified files: www/h2o: Makefile distinfo Log message: Switch to PowerDNS fork of h2o, it has the rapid reset fixes. ok sthen@
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: o...@cvs.openbsd.org2023/10/11 08:27:56 Modified files: net/powerdns : Makefile distinfo Log message: Update to PowerDNS Authoritative Server 4.8.3
Re: Bought a new Intel laptop? Test packages!
On Sat, Sep 30, 2023 at 12:42:53PM +0200, Theo Buehler wrote: > > I ran into a similar backtrace with Taisei [0]. Most seemingly it needs > > to USE_NOBTCFI=Yes as our Boost sits at 1.80.0, while CET was introduced > > in boost::context at 1.81.0 and updating Boost is not an option at this > > point. > > > > I successfully tested it with USE_NOBTCFI. > > > > [0]: https://github.com/taisei-project/koishi/issues/6#issue-1888944976 > > > > Committed this slightly tweaked diff, thanks. REVISION bump is needed > since the package changed. Maybe USE_NOBTCFI-amd64 would be enough, but > I think at this point in the cycle it's better to play it safe. Great! Please commit. I'll make sure it gets removed on the first update after boost 1.81 is in-tree. -Otto > > Index: Makefile > === > RCS file: /cvs/ports/net/powerdns_recursor/Makefile,v > retrieving revision 1.45 > diff -u -p -r1.45 Makefile > --- Makefile 27 Sep 2023 14:18:29 - 1.45 > +++ Makefile 30 Sep 2023 10:38:58 - > @@ -4,6 +4,7 @@ V=4.9.1 > DISTNAME=pdns-recursor-${V} > EXTRACT_SUFX = .tar.bz2 > PKGNAME= powerdns-recursor-${V} > +REVISION=0 > > CATEGORIES= net > > @@ -11,6 +12,9 @@ HOMEPAGE= https://www.powerdns.com/ > > # GPLv2 only, OpenSSL exemption > PERMIT_PACKAGE= Yes > + > +# SIGILL on startup because of boost::context. Needed until boost >= 1.81. > +USE_NOBTCFI= Yes > > WANTLIB+=${COMPILER_LIBCXX} boost_context-mt boost_filesystem-mt > WANTLIB+=boost_system-mt boost_thread-mt c crypto curl fstrm m
Re: Bought a new Intel laptop? Test packages!
On Fri, Sep 29, 2023 at 09:20:00PM +, Lucas wrote: > Christian Weisgerber wrote: > > If you bought a new Intel laptop this year, chances are its hardware > > is capable of IBT (indirect branch tracking), and OpenBSD now enables > > this security feature by default. > > > > It is very likely that some packages are broken at runtime because > > they fail to handle IBT. This doesn't show up at build time, so > > people need to actually run their favorite packages from a snapshot > > _now_ and need to report SIGILL failures, or those problems will > > not be fixed in time for the release. > > I gave the ports that depend on boost-context a spin, after running into > issues trying out the Taisei Project update [0]. > > I used this command to hopefully find all the programs that potentially > consume boost-context: > > sqlite3 /usr/local/share/sqlports <<'_EOF' > SELECT FullPkgPath FROM PortsQ WHERE > (BUILD_DEPENDS LIKE '%devel/boost,-md%' OR > LIB_DEPENDS LIKE '%devel/boost,-md%' OR > WANTLIB LIKE '%boost_context-mt%') AND > USE_NOBTCFI IS NULL > ORDER BY FullPkgPath; > _EOF > > which returned > > - games/wesnoth: didn't find issues playing the campaign, nor I found > any clear usage of boost::context or boost::coroutine grepping the > code around. > - net/icinga/core2,-main: didn't find issues running it in foreground > nor any clear usage of boost::context or boost::coroutine grepping > the code, other than printing a backtrace (but I get the feeling it > can be hidden in boost::asio) > - net/powerdns_recursor: SIGILLs on start. otto@ in cc. > - sysutils/nix: tbh, I have no clue on how to test it. > > Lucas I can only guess about powerdns_recursor here, ads I do not have a modern Intel machine: it's the context switch assembly in boot-md. I have no clue yet how to fix this, I did not follow the IBT developments closely. Sadly I'll also be leaving for a holiday for a week this weekend. So I'd be really grateful if somebody else could take a look at powerdns_recursor and boost-md. -Otto
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: o...@cvs.openbsd.org2023/09/08 05:52:26 Modified files: net/dnsdist: Makefile distinfo Log message: Update to dnsdist 1.8.1
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: o...@cvs.openbsd.org2023/09/07 03:45:08 Modified files: net/powerdns : Makefile distinfo Log message: Update to PowerDNS Authoritative Server 4.8.2
Re: Valgrind: Detect access outside the range of malloc.
On Tue, Sep 05, 2023 at 03:00:12PM +0900, Masato Asou wrote: > From: Otto Moerbeek > Date: Tue, 5 Sep 2023 07:40:18 +0200 > > > On Tue, Sep 05, 2023 at 09:38:40AM +0900, Masato Asou wrote: > > > >> hi, > >> > >> I have fixed a bug in Valgrind. The Valgrind could not detect access > >> outside the range of malloc. > >> > >> comments, ok? > > > > This works much better that before. Thanks for working on this! > > > > It now detects out of bounds read and writes correctly. A double > > free is detected. > > Also, the spurious reports for accesses to errno are gone. > > Thank you for your comment. > > > It does not report proper locations though, even if I compile my test > > program with -g: > > Could you show me your test program? Here it is: #include #include int main(int argc, char *argv[]) { size_t sz = atoi(argv[1]); unsigned char *p = malloc(sz); printf("%p\n", p); p[sz] = 0; printf("%x\n", p[sz]); free(p); free(p); return 0; } > -- > ASOU Masato > > > ==23912== Invalid read of size 1 > > ==23912==at 0x109B5D: ??? (in ./a.out) > > ==23912==by 0x1098D1: ??? (in ./a.out) > > ==23912== Address 0x4a42840 is 0 bytes after a block of size 10,240 alloc'd > > ==23912==at 0x493A3A9: malloc (vg_replace_malloc.c:435) > > ==23912==by 0x109B32: ??? (in ./a.out) > > ==23912==by 0x1098D1: ??? (in ./a.out) > > ==23912== > > 0 > > ==23912== Invalid free() / delete / delete[] / realloc() > > ==23912==at 0x493C981: free (vg_replace_malloc.c:978) > > ==23912==by 0x109B80: ??? (in ./a.out) > > ==23912==by 0x1098D1: ??? (in ./a.out) > > ==23912== Address 0x4a40040 is 0 bytes inside a block of size 10,240 free'd > > ==23912==at 0x493C981: free (vg_replace_malloc.c:978) > > ==23912==by 0x109B77: ??? (in ./a.out) > > ==23912==by 0x1098D1: ??? (in ./a.out) > > ==23912== Block was alloc'd at > > ==23912==at 0x493A3A9: malloc (vg_replace_malloc.c:435) > > ==23912==by 0x109B32: ??? (in ./a.out) > > ==23912==by 0x1098D1: ??? (in ./a.out) > > > > addr2line -e ./a.out 0x109B80 also does not succeed in translating the > > address. > > > > There are also a few spurious reports of the form: > > > > ==23912== Use of uninitialised value of size 8 > > ==23912==at 0x499518D: write (sys/w_write.c:26) > > ==23912==by 0x49E9FB2: __sflush (stdio/fflush.c:80) > > ==23912==by 0x49DA548: __sfvwrite (stdio/fvwrite.c:191) > > ==23912==by 0x49E42ED: __sprint (stdio/vfprintf.c:108) > > ==23912==by 0x49E42ED: __vfprintf (stdio/vfprintf.c:1064) > > ==23912==by 0x49E0FA5: vfprintf (stdio/vfprintf.c:263) > > ==23912==by 0x49F7C74: printf (stdio/printf.c:44) > > ==23912==by 0x109B48: ??? (in ./a.out) > > ==23912==by 0x1098D1: ??? (in ./a.out) > > > > But it is a great improvement, > > > > -Otto > > > >> -- > >> ASOU Masato > >> > >> Index: devel/valgrind/Makefile > >> === > >> RCS file: /cvs/ports/devel/valgrind/Makefile,v > >> retrieving revision 1.31 > >> diff -u -p -r1.31 Makefile > >> --- devel/valgrind/Makefile5 Aug 2023 03:32:22 - 1.31 > >> +++ devel/valgrind/Makefile4 Sep 2023 23:14:33 - > >> @@ -5,7 +5,7 @@ CATEGORIES = devel > >> > >> V = 3.21.0 > >> DISTNAME =valgrind-${V} > >> -REVISION =0 > >> +REVISION =1 > >> EXTRACT_SUFX =.tar.bz2 > >> > >> MASTER_SITES =https://sourceware.org/pub/valgrind/ > >> Index: > >> devel/valgrind/patches/patch-coregrind_m_replacemalloc_vg_replace_malloc_c > >> === > >> RCS file: > >> devel/valgrind/patches/patch-coregrind_m_replacemalloc_vg_replace_malloc_c > >> diff -N > >> devel/valgrind/patches/patch-coregrind_m_replacemalloc_vg_replace_malloc_c > >> --- /dev/null 1 Jan 1970 00:00:00 - > >> +++ > >> devel/valgrind/patches/patch-coregrind_m_replacemalloc_vg_replace_malloc_c > >> 4 Sep 2023 23:14:33 - > >> @@ -0,0 +1,263 @@ > >> +--- coregrind/m_replacemalloc/vg_replace_malloc.c.orig > >> coregrind/m_replacemalloc/vg_
Re: Valgrind: Detect access outside the range of malloc.
On Tue, Sep 05, 2023 at 09:38:40AM +0900, Masato Asou wrote: > hi, > > I have fixed a bug in Valgrind. The Valgrind could not detect access > outside the range of malloc. > > comments, ok? This works much better that before. Thanks for working on this! It now detects out of bounds read and writes correctly. A double free is detected. Also, the spurious reports for accesses to errno are gone. It does not report proper locations though, even if I compile my test program with -g: ==23912== Invalid read of size 1 ==23912==at 0x109B5D: ??? (in ./a.out) ==23912==by 0x1098D1: ??? (in ./a.out) ==23912== Address 0x4a42840 is 0 bytes after a block of size 10,240 alloc'd ==23912==at 0x493A3A9: malloc (vg_replace_malloc.c:435) ==23912==by 0x109B32: ??? (in ./a.out) ==23912==by 0x1098D1: ??? (in ./a.out) ==23912== 0 ==23912== Invalid free() / delete / delete[] / realloc() ==23912==at 0x493C981: free (vg_replace_malloc.c:978) ==23912==by 0x109B80: ??? (in ./a.out) ==23912==by 0x1098D1: ??? (in ./a.out) ==23912== Address 0x4a40040 is 0 bytes inside a block of size 10,240 free'd ==23912==at 0x493C981: free (vg_replace_malloc.c:978) ==23912==by 0x109B77: ??? (in ./a.out) ==23912==by 0x1098D1: ??? (in ./a.out) ==23912== Block was alloc'd at ==23912==at 0x493A3A9: malloc (vg_replace_malloc.c:435) ==23912==by 0x109B32: ??? (in ./a.out) ==23912==by 0x1098D1: ??? (in ./a.out) addr2line -e ./a.out 0x109B80 also does not succeed in translating the address. There are also a few spurious reports of the form: ==23912== Use of uninitialised value of size 8 ==23912==at 0x499518D: write (sys/w_write.c:26) ==23912==by 0x49E9FB2: __sflush (stdio/fflush.c:80) ==23912==by 0x49DA548: __sfvwrite (stdio/fvwrite.c:191) ==23912==by 0x49E42ED: __sprint (stdio/vfprintf.c:108) ==23912==by 0x49E42ED: __vfprintf (stdio/vfprintf.c:1064) ==23912==by 0x49E0FA5: vfprintf (stdio/vfprintf.c:263) ==23912==by 0x49F7C74: printf (stdio/printf.c:44) ==23912==by 0x109B48: ??? (in ./a.out) ==23912==by 0x1098D1: ??? (in ./a.out) But it is a great improvement, -Otto > -- > ASOU Masato > > Index: devel/valgrind/Makefile > === > RCS file: /cvs/ports/devel/valgrind/Makefile,v > retrieving revision 1.31 > diff -u -p -r1.31 Makefile > --- devel/valgrind/Makefile 5 Aug 2023 03:32:22 - 1.31 > +++ devel/valgrind/Makefile 4 Sep 2023 23:14:33 - > @@ -5,7 +5,7 @@ CATEGORIES = devel > > V = 3.21.0 > DISTNAME = valgrind-${V} > -REVISION = 0 > +REVISION = 1 > EXTRACT_SUFX = .tar.bz2 > > MASTER_SITES = https://sourceware.org/pub/valgrind/ > Index: > devel/valgrind/patches/patch-coregrind_m_replacemalloc_vg_replace_malloc_c > === > RCS file: > devel/valgrind/patches/patch-coregrind_m_replacemalloc_vg_replace_malloc_c > diff -N > devel/valgrind/patches/patch-coregrind_m_replacemalloc_vg_replace_malloc_c > --- /dev/null 1 Jan 1970 00:00:00 - > +++ > devel/valgrind/patches/patch-coregrind_m_replacemalloc_vg_replace_malloc_c > 4 Sep 2023 23:14:33 - > @@ -0,0 +1,263 @@ > +--- coregrind/m_replacemalloc/vg_replace_malloc.c.orig > coregrind/m_replacemalloc/vg_replace_malloc.c > +@@ -222,7 +222,7 @@ > + #define SET_ERRNO_ENOMEM if (__errno_location)\ > + (*__errno_location ()) = VKI_ENOMEM; > + #define SET_ERRNO_EINVAL {} > +-#elif defined(VGO_freebsd) > ++#elif defined(VGO_freebsd) || defined(VGO_openbsd) > + extern int *__error (void) __attribute__((weak)); > + #define SET_ERRNO_ENOMEM if (__error)\ > + (*__error ()) = VKI_ENOMEM; > +@@ -430,7 +430,7 @@ > + ALLOC_or_NULL(VG_Z_LIBC_SONAME, malloc, malloc); > + ALLOC_or_NULL(SO_SYN_MALLOC, malloc, malloc); > + > +-#elif defined(VGO_freebsd) > ++#elif defined(VGO_freebsd) || defined(VGO_openbsd) > + ALLOC_or_NULL(VG_Z_LIBC_SONAME, malloc, malloc); > + ALLOC_or_NULL(SO_SYN_MALLOC, malloc, malloc); > + > +@@ -472,7 +472,7 @@ > + ALLOC_or_BOMB(SO_SYN_MALLOC, _Znwm, __builtin_new); > + #endif > + > +-#elif defined(VGO_freebsd) > ++#elif defined(VGO_freebsd) || defined(VGO_openbsd) > + // operator new(unsigned int) > + #if VG_WORDSIZE == 4 > + ALLOC_or_BOMB(VG_Z_LIBSTDCXX_SONAME, _Znwj, __builtin_new); > +@@ -532,7 +532,7 @@ > + ALLOC_or_BOMB_ALIGNED(SO_SYN_MALLOC, _ZnwmSt11align_val_t, > __builtin_new_aligned); > + #endif > + > +-#elif defined(VGO_freebsd) > ++#elif defined(VGO_freebsd) || defined(VGO_openbsd) > + // operator new(unsigned int) > + #if VG_WORDSIZE == 4 > + ALLOC_or_BOMB_ALIGNED(VG_Z_LIBSTDCXX_SONAME, _ZnwjSt11align_val_t, > __builtin_new_aligned); > +@@ -592,7 +592,7 @@ > + ALLOC_or_NULL(SO_SYN_MALLOC,
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: o...@cvs.openbsd.org2023/08/25 03:01:26 Modified files: net/powerdns_recursor: Makefile distinfo Log message: Update to PowerDNS Recursor 4.9.1
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: o...@cvs.openbsd.org2023/07/07 07:26:29 Modified files: net/powerdns : Makefile distinfo Log message: Update to PowerDNS Authoritative Server 4.8.1
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: o...@cvs.openbsd.org2023/06/30 05:18:24 Modified files: net/powerdns_recursor: Makefile distinfo net/powerdns_recursor/files: recursor.conf net/powerdns_recursor/patches: patch-Makefile_am patch-configure_ac net/powerdns_recursor/pkg: README Log message: Update to PowerDNS Recursor 4.9.0
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: o...@cvs.openbsd.org2023/06/01 08:30:01 Modified files: net/powerdns : Makefile distinfo net/powerdns/patches: patch-configure_ac Log message: Update to PowerDNS Authoritative Server 4.8.0
Re: chessx chunk canary corrupted
On Tue, May 30, 2023 at 11:48:50AM +0200, HAL.gi wrote: > Dear OpenBSD team ! > > I installed chessx-1.5.6 with pkg_add from the ftp[.]openbsd[.]org 7.3 amd64 > package directory, without any errors. > > Starting it the first time in xterm it shows the GUI. > > Popping up a setup question regarding if one like to download the chessx > database. > > Simultaneously it tried to connect to chessx[.]sourceforge[.]net in the > background. At this stage one have not the chance to disable this option to > download the database or any download activity in the first place. > > But that is not the main error. > > If one quits the program over the menu, it shows the "chunk corrupt canary" > error ! > > It produces this error and a core dump every time. > > vm.malloc_conf is set at >S It's almost certainly y a bug in the app or a lib, writing outside an allocation. Either try to hunt it down (use gdb form packages, it is much better than the one in base), switch of S all together or try running without canaries for this app only: MALLOC_OPTIONS=c chessx (malloc options are processed incrementally) -Otto > > I loaded chessx with gdb, used than the run and bt option: > > > GNU gdb 6.3 > Copyright 2004 Free Software Foundation, Inc. > GDB is free software, covered by the GNU General Public License, and you are > welcome to change it and/or distribute copies of it under certain > conditions. > Type "show copying" to see the conditions. > There is absolutely no warranty for GDB. Type "show warranty" for details. > This GDB was configured as "amd64-unknown-openbsd7.3"...(no debugging > symbols found) > > Core was generated by `chessx'. > Program terminated with signal 6, Aborted. > Reading symbols from /usr/lib/libpthread.so.27.0...done. > Loaded symbols for /usr/lib/libpthread.so.27.0 > Loaded symbols for /usr/local/bin/chessx > Reading symbols from /usr/local/lib/libquazip1-qt5.so.0.0...done. > Loaded symbols for /usr/local/lib/libquazip1-qt5.so.0.0 > Reading symbols from /usr/lib/libz.so.7.0...done. > Loaded symbols for /usr/lib/libz.so.7.0 > Reading symbols from /usr/local/lib/qt5/libQt5PrintSupport.so.3.0...done. > Loaded symbols for /usr/local/lib/qt5/./libQt5PrintSupport.so.3.0 > Reading symbols from /usr/local/lib/qt5/libQt5Svg.so.3.0...done. > Loaded symbols for /usr/local/lib/qt5/./libQt5Svg.so.3.0 > Reading symbols from /usr/local/lib/qt5/libQt5Widgets.so.4.0...done. > Loaded symbols for /usr/local/lib/qt5/./libQt5Widgets.so.4.0 > Reading symbols from /usr/local/lib/qt5/libQt5Xml.so.3.0...done. > Loaded symbols for /usr/local/lib/qt5/./libQt5Xml.so.3.0 > Reading symbols from /usr/local/lib/qt5/libQt5Multimedia.so.3.0...done. > Loaded symbols for /usr/local/lib/qt5/./libQt5Multimedia.so.3.0 > Reading symbols from /usr/local/lib/qt5/libQt5Gui.so.3.0...done. > Loaded symbols for /usr/local/lib/qt5/./libQt5Gui.so.3.0 > Reading symbols from /usr/local/lib/qt5/libQt5Network.so.4.0...done. > Loaded symbols for /usr/local/lib/qt5/./libQt5Network.so.4.0 > Reading symbols from /usr/local/lib/libpulse.so.7.2...done. > Loaded symbols for /usr/local/lib/libpulse.so.7.2 > Reading symbols from /usr/local/lib/libpulse-mainloop-glib.so.1.2...done. > Loaded symbols for /usr/local/lib/libpulse-mainloop-glib.so.1.2 > Reading symbols from /usr/local/lib/qt5/libQt5TextToSpeech.so.1.0...done. > Loaded symbols for /usr/local/lib/qt5/./libQt5TextToSpeech.so.1.0 > Reading symbols from /usr/local/lib/qt5/libQt5Core.so.4.0...done. > Loaded symbols for /usr/local/lib/qt5/./libQt5Core.so.4.0 > Reading symbols from /usr/lib/libc++.so.9.0...done. > Loaded symbols for /usr/lib/libc++.so.9.0 > Reading symbols from /usr/lib/libc++abi.so.6.0...done. > Loaded symbols for /usr/lib/libc++abi.so.6.0 > Symbols already loaded for /usr/lib/libpthread.so.27.0 > Reading symbols from /usr/lib/libm.so.10.1...done. > Loaded symbols for /usr/lib/libm.so.10.1 > Reading symbols from /usr/lib/libc.so.97.0...done. > Loaded symbols for /usr/lib/libc.so.97.0 > Reading symbols from /usr/X11R6/lib/libGL.so.18.0...done. > Loaded symbols for /usr/X11R6/lib/libGL.so.18.0 > Reading symbols from /usr/local/lib/libglib-2.0.so.4201.9...done. > Loaded symbols for /usr/local/lib/libglib-2.0.so.4201.9 > Reading symbols from /usr/local/lib/libintl.so.7.0...done. > Loaded symbols for /usr/local/lib/libintl.so.7.0 > Reading symbols from /usr/local/lib/libpng.so.18.0...done. > Loaded symbols for /usr/local/lib/libpng.so.18.0 > Reading symbols from /usr/local/lib/libharfbuzz.so.18.1...done. > Loaded symbols for /usr/local/lib/libharfbuzz.so.18.1 > Reading symbols from /usr/lib/libssl.so.53.2...done. > Loaded symbols for /usr/lib/libssl.so.53.2 > Reading symbols from /usr/lib/libcrypto.so.50.2...done. > Loaded symbols for /usr/lib/libcrypto.so.50.2 > Reading symbols from > /usr/local/lib/pulseaudio/libpulsecommon-16.1.so...done. > Loaded symbols for /usr/local/lib/pulseaudio/libpulsecommon-16.1.so > Reading symbols from
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: o...@cvs.openbsd.org2023/05/01 02:51:43 Modified files: net/powerdns : Makefile distinfo Log message: Update to PowerDNS Authoritative Server 4.7.4
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: o...@cvs.openbsd.org2023/03/30 23:38:03 Modified files: net/dnsdist: Makefile distinfo net/dnsdist/patches: patch-configure_ac net/dnsdist/pkg: PLIST Log message: Update to dnsdist 1.8.0; ok sthen@
Re: update to dnsdist 1.8.0
On Thu, Mar 30, 2023 at 05:21:16PM +0100, Stuart Henderson wrote: > On 2023/03/30 15:50, Otto Moerbeek wrote: > > Hi, > > > > Did I do the PLIST correctly? The dnsdistconf.lua conf file was the > > only one in share/examples/dnsdist > > > > OK? > > > > -Otto > > > > Index: Makefile > > === > > RCS file: /home/cvs/ports/net/dnsdist/Makefile,v > > retrieving revision 1.31 > > diff -u -p -r1.31 Makefile > > --- Makefile8 Nov 2022 11:14:54 - 1.31 > > +++ Makefile30 Mar 2023 13:47:51 - > > @@ -1,8 +1,7 @@ > > COMMENT= highly DNS-, DoS- and abuse-aware loadbalancer > > > > -DISTNAME= dnsdist-1.7.3 > > +DISTNAME= dnsdist-1.8.0 > > EXTRACT_SUFX= .tar.bz2 > > -REVISION= 0 > > > > CATEGORIES=net > > > > @@ -58,8 +57,7 @@ LIB_DEPENDS+= ${MODLUA_LIB_DEPENDS} \ > > www/nghttp2 > > > > post-install: > > - ${INSTALL_DATA_DIR} ${PREFIX}/share/{doc,examples}/dnsdist > > - cd ${WRKSRC}; ${INSTALL_DATA} README.md ${PREFIX}/share/doc/dnsdist/; \ > > - ${INSTALL_DATA} dnsdistconf.lua ${PREFIX}/share/examples/dnsdist/ > > How about using dnsdist.conf-dist instead? It seems helpful to have an > example file. Like this? -Otto Index: Makefile === RCS file: /home/cvs/ports/net/dnsdist/Makefile,v retrieving revision 1.31 diff -u -p -r1.31 Makefile --- Makefile8 Nov 2022 11:14:54 - 1.31 +++ Makefile30 Mar 2023 18:01:43 - @@ -1,8 +1,7 @@ COMMENT= highly DNS-, DoS- and abuse-aware loadbalancer -DISTNAME= dnsdist-1.7.3 +DISTNAME= dnsdist-1.8.0 EXTRACT_SUFX= .tar.bz2 -REVISION= 0 CATEGORIES=net @@ -60,6 +59,6 @@ LIB_DEPENDS+= ${MODLUA_LIB_DEPENDS} \ post-install: ${INSTALL_DATA_DIR} ${PREFIX}/share/{doc,examples}/dnsdist cd ${WRKSRC}; ${INSTALL_DATA} README.md ${PREFIX}/share/doc/dnsdist/; \ - ${INSTALL_DATA} dnsdistconf.lua ${PREFIX}/share/examples/dnsdist/ + ${INSTALL_DATA} dnsdist.conf-dist ${PREFIX}/share/examples/dnsdist/ .include Index: distinfo === RCS file: /home/cvs/ports/net/dnsdist/distinfo,v retrieving revision 1.15 diff -u -p -r1.15 distinfo --- distinfo2 Nov 2022 15:10:05 - 1.15 +++ distinfo30 Mar 2023 18:01:43 - @@ -1,2 +1,2 @@ -SHA256 (dnsdist-1.7.3.tar.bz2) = fq9vrC8mVlxdhljUKiE3meBfTTvGjnxxbnF030ExWIY= -SIZE (dnsdist-1.7.3.tar.bz2) = 1379011 +SHA256 (dnsdist-1.8.0.tar.bz2) = HA03XCVFPTSbiOA/9YmqJgPKhpL8mDZMBo6tNygEcE8= +SIZE (dnsdist-1.8.0.tar.bz2) = 1480500 Index: patches/patch-configure_ac === RCS file: /home/cvs/ports/net/dnsdist/patches/patch-configure_ac,v retrieving revision 1.7 diff -u -p -r1.7 patch-configure_ac --- patches/patch-configure_ac 11 Mar 2022 19:45:54 - 1.7 +++ patches/patch-configure_ac 30 Mar 2023 18:01:43 - @@ -1,14 +1,14 @@ Index: configure.ac --- configure.ac.orig +++ configure.ac -@@ -18,8 +18,8 @@ AC_DEFINE([DNSDIST], [1], +@@ -16,8 +16,8 @@ AC_DEFINE([DNSDIST], [1], LT_PREREQ([2.2.2]) LT_INIT([disable-static]) --CFLAGS="-g -O3 -Wall -Wextra -Wshadow -Wno-unused-parameter $CFLAGS" --CXXFLAGS="-g -O3 -Wall -Wextra -Wshadow -Wno-unused-parameter -Wmissing-declarations -Wredundant-decls $CXXFLAGS" -+CFLAGS="-g -Wall -Wextra -Wshadow -Wno-unused-parameter $CFLAGS" -+CXXFLAGS="-g -Wall -Wextra -Wshadow -Wno-unused-parameter -Wmissing-declarations -Wredundant-decls $CXXFLAGS" +-CFLAGS="-g -O3 -Wall -Wextra -Wshadow -Wno-unused-parameter -fvisibility=hidden $CFLAGS" +-CXXFLAGS="-g -O3 -Wall -Wextra -Wshadow -Wno-unused-parameter -Wmissing-declarations -Wredundant-decls -fvisibility=hidden $CXXFLAGS" ++CFLAGS="-g -Wall -Wextra -Wshadow -Wno-unused-parameter -fvisibility=hidden $CFLAGS" ++CXXFLAGS="-g -Wall -Wextra -Wshadow -Wno-unused-parameter -Wmissing-declarations -Wredundant-decls -fvisibility=hidden $CXXFLAGS" PDNS_WITH_LIBSODIUM PDNS_CHECK_DNSTAP([auto]) Index: pkg/PLIST === RCS file: /home/cvs/ports/net/dnsdist/pkg/PLIST,v retrieving revision 1.7 diff -u -p -r1.7 PLIST --- pkg/PLIST 8 Nov 2022 11:14:54 - 1.7 +++ pkg/PLIST 30 Mar 2023 18:01:43 - @@ -7,7 +7,7 @@ share/doc/dnsdist/ share/doc/dnsdist/README.md share/examples/dnsdist/ @sample ${SYSCONFDIR}/ -share/examples/dnsdist/dnsdistconf.lua +share/examples/dnsdist/dnsdist.conf-dist @mode 640 @group _dnsdist @sample ${SYSCONFDIR}/dnsdist.conf
update to dnsdist 1.8.0
Hi, Did I do the PLIST correctly? The dnsdistconf.lua conf file was the only one in share/examples/dnsdist OK? -Otto Index: Makefile === RCS file: /home/cvs/ports/net/dnsdist/Makefile,v retrieving revision 1.31 diff -u -p -r1.31 Makefile --- Makefile8 Nov 2022 11:14:54 - 1.31 +++ Makefile30 Mar 2023 13:47:51 - @@ -1,8 +1,7 @@ COMMENT= highly DNS-, DoS- and abuse-aware loadbalancer -DISTNAME= dnsdist-1.7.3 +DISTNAME= dnsdist-1.8.0 EXTRACT_SUFX= .tar.bz2 -REVISION= 0 CATEGORIES=net @@ -58,8 +57,7 @@ LIB_DEPENDS+= ${MODLUA_LIB_DEPENDS} \ www/nghttp2 post-install: - ${INSTALL_DATA_DIR} ${PREFIX}/share/{doc,examples}/dnsdist - cd ${WRKSRC}; ${INSTALL_DATA} README.md ${PREFIX}/share/doc/dnsdist/; \ - ${INSTALL_DATA} dnsdistconf.lua ${PREFIX}/share/examples/dnsdist/ + ${INSTALL_DATA_DIR} ${PREFIX}/share/doc/dnsdist + cd ${WRKSRC}; ${INSTALL_DATA} README.md ${PREFIX}/share/doc/dnsdist/ .include Index: distinfo === RCS file: /home/cvs/ports/net/dnsdist/distinfo,v retrieving revision 1.15 diff -u -p -r1.15 distinfo --- distinfo2 Nov 2022 15:10:05 - 1.15 +++ distinfo30 Mar 2023 13:47:51 - @@ -1,2 +1,2 @@ -SHA256 (dnsdist-1.7.3.tar.bz2) = fq9vrC8mVlxdhljUKiE3meBfTTvGjnxxbnF030ExWIY= -SIZE (dnsdist-1.7.3.tar.bz2) = 1379011 +SHA256 (dnsdist-1.8.0.tar.bz2) = HA03XCVFPTSbiOA/9YmqJgPKhpL8mDZMBo6tNygEcE8= +SIZE (dnsdist-1.8.0.tar.bz2) = 1480500 Index: patches/patch-configure_ac === RCS file: /home/cvs/ports/net/dnsdist/patches/patch-configure_ac,v retrieving revision 1.7 diff -u -p -r1.7 patch-configure_ac --- patches/patch-configure_ac 11 Mar 2022 19:45:54 - 1.7 +++ patches/patch-configure_ac 30 Mar 2023 13:47:51 - @@ -1,14 +1,14 @@ Index: configure.ac --- configure.ac.orig +++ configure.ac -@@ -18,8 +18,8 @@ AC_DEFINE([DNSDIST], [1], +@@ -16,8 +16,8 @@ AC_DEFINE([DNSDIST], [1], LT_PREREQ([2.2.2]) LT_INIT([disable-static]) --CFLAGS="-g -O3 -Wall -Wextra -Wshadow -Wno-unused-parameter $CFLAGS" --CXXFLAGS="-g -O3 -Wall -Wextra -Wshadow -Wno-unused-parameter -Wmissing-declarations -Wredundant-decls $CXXFLAGS" -+CFLAGS="-g -Wall -Wextra -Wshadow -Wno-unused-parameter $CFLAGS" -+CXXFLAGS="-g -Wall -Wextra -Wshadow -Wno-unused-parameter -Wmissing-declarations -Wredundant-decls $CXXFLAGS" +-CFLAGS="-g -O3 -Wall -Wextra -Wshadow -Wno-unused-parameter -fvisibility=hidden $CFLAGS" +-CXXFLAGS="-g -O3 -Wall -Wextra -Wshadow -Wno-unused-parameter -Wmissing-declarations -Wredundant-decls -fvisibility=hidden $CXXFLAGS" ++CFLAGS="-g -Wall -Wextra -Wshadow -Wno-unused-parameter -fvisibility=hidden $CFLAGS" ++CXXFLAGS="-g -Wall -Wextra -Wshadow -Wno-unused-parameter -Wmissing-declarations -Wredundant-decls -fvisibility=hidden $CXXFLAGS" PDNS_WITH_LIBSODIUM PDNS_CHECK_DNSTAP([auto]) Index: pkg/PLIST === RCS file: /home/cvs/ports/net/dnsdist/pkg/PLIST,v retrieving revision 1.7 diff -u -p -r1.7 PLIST --- pkg/PLIST 8 Nov 2022 11:14:54 - 1.7 +++ pkg/PLIST 30 Mar 2023 13:47:51 - @@ -5,9 +5,7 @@ @man man/man1/dnsdist.1 share/doc/dnsdist/ share/doc/dnsdist/README.md -share/examples/dnsdist/ @sample ${SYSCONFDIR}/ -share/examples/dnsdist/dnsdistconf.lua @mode 640 @group _dnsdist @sample ${SYSCONFDIR}/dnsdist.conf
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: o...@cvs.openbsd.org2023/03/30 01:28:57 Modified files: net/powerdns_recursor: Makefile distinfo Log message: Update to PowerDNS Recursor 4.8.4 Low severity security release, see https://blog.powerdns.com/2023/03/29/security-advisory-2023-02-for-powerdns-recursor-up-to-and-including-4-6-5-4-7-4-and-4-8-3/
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: o...@cvs.openbsd.org2023/03/07 02:49:21 Modified files: net/powerdns_recursor: Makefile distinfo Log message: Update to PowerDNS Recursor 4.8.3
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: o...@cvs.openbsd.org2023/02/28 02:13:29 Modified files: www: Makefile www/h2o: Makefile www/h2o/patches: patch-CMakeLists_txt www/h2o/pkg: DESCR PLIST Removed files: www/h2o/patches: patch-deps_mruby_Rakefile www/h2o/pkg: PFRAG.mruby Log message: Remove mruby flavor of h2o, it is broken with the upcoming mruby update ok sthen@
Re: mruby 3.2.0
On Tue, Feb 28, 2023 at 07:17:50AM +0100, Otto Moerbeek wrote: > I have no problem removing the mruby flavor of h2o. I hope to be able > to do that soon. Here you go. -Otto Index: Makefile === RCS file: /home/cvs/ports/www/h2o/Makefile,v retrieving revision 1.14 diff -u -p -r1.14 Makefile --- Makefile7 Jan 2023 20:05:13 - 1.14 +++ Makefile28 Feb 2023 07:17:36 - @@ -21,13 +21,6 @@ WANTLIB += ${COMPILER_LIBCXX} c crypto m # uses TLS COMPILER = base-clang ports-gcc -FLAVORS = mruby -FLAVOR ?= - -.if ${FLAVOR:Mmruby} -BROKEN-arm = SIGBUS in mruby build -.endif - MODULES = devel/cmake BUILD_DEPENDS =devel/bison @@ -37,22 +30,6 @@ CONFIGURE_ARGS = -DBUILD_SHARED_LIBS=ON -DCMAKE_INSTALL_SYSCONFDIR=${SYSCONFDIR}/h2o \ -DWITH_BUNDLED_SSL=OFF \ -DWITH_PICOTLS=OFF - -.if ${FLAVOR:Mmruby} -MODULES += lang/ruby -MODRUBY_RUNDEP = No -CONFIGURE_ARGS += -DWITH_MRUBY=ON -# mruby build picks up the static oniguruma library if present during build -# otherwise uses a bundled older version -BUILD_DEPENDS += textproc/oniguruma -MAKE_ENV +=AWK=/usr/bin/awk \ - SED=/usr/bin/sed \ - GREP=/usr/bin/grep \ - ac_cv_path_mkdir=/bin/mkdir - -pre-configure: - ln -sf ${RUBY} ${WRKDIR}/bin/ruby -.endif post-install: cd ${WRKSRC}/doc; pax -rw [a-z]* ${PREFIX}/share/doc/h2o/ Index: patches/patch-CMakeLists_txt === RCS file: /home/cvs/ports/www/h2o/patches/patch-CMakeLists_txt,v retrieving revision 1.3 diff -u -p -r1.3 patch-CMakeLists_txt --- patches/patch-CMakeLists_txt2 Sep 2022 13:18:16 - 1.3 +++ patches/patch-CMakeLists_txt28 Feb 2023 07:17:36 - @@ -45,27 +45,3 @@ Index: CMakeLists.txt IF (ZLIB_FOUND) INCLUDE_DIRECTORIES(${ZLIB_INCLUDE_DIRS}) -@@ -460,14 +460,21 @@ IF (WITH_MRUBY) - ELSE () - SET(MRUBY_TOOLCHAIN "gcc") - ENDIF () --ADD_CUSTOM_TARGET(mruby MRUBY_TOOLCHAIN=${MRUBY_TOOLCHAIN} MRUBY_CONFIG=${CMAKE_CURRENT_SOURCE_DIR}/misc/mruby_config.rb MRUBY_BUILD_DIR=${CMAKE_CURRENT_BINARY_DIR}/mruby ruby minirake --WORKING_DIRECTORY ${CMAKE_CURRENT_SOURCE_DIR}/deps/mruby) -+ADD_CUSTOM_TARGET(mruby -+# deps/mruby/tasks/toolchains/clang.rake looks for CC, CXX and LD. -+# There are no C++ files in deps/mruby, use the C compiler for linking. -+MRUBY_TOOLCHAIN=${MRUBY_TOOLCHAIN} CC=${CMAKE_C_COMPILER} CXX=${CMAKE_CXX_COMPILER} LD=${CMAKE_C_COMPILER} MRUBY_CONFIG=${CMAKE_CURRENT_SOURCE_DIR}/misc/mruby_config.rb MRUBY_BUILD_DIR=${CMAKE_CURRENT_BINARY_DIR}/mruby MRUBY_ADDITIONAL_CONFIG=${MRUBY_ADDITIONAL_CONFIG} ruby minirake -+WORKING_DIRECTORY ${CMAKE_CURRENT_SOURCE_DIR}/deps/mruby -+BYPRODUCTS "${CMAKE_CURRENT_BINARY_DIR}/mruby/host/lib/libmruby.a" -+ "${CMAKE_CURRENT_BINARY_DIR}/mruby/host/mrbgems/mruby-onig-regexp/onigmo-6.1.1/.libs/libonigmo.a" -+) - LIST(APPEND STANDALONE_SOURCE_FILES - lib/handler/mruby.c - lib/handler/mruby/chunked.c - lib/handler/mruby/http_request.c - lib/handler/configurator/mruby.c) - SET(STANDALONE_COMPILE_FLAGS "${STANDALONE_COMPILE_FLAGS} -DH2O_USE_MRUBY=1") -+SET_PROPERTY(DIRECTORY PROPERTY ADDITIONAL_MAKE_CLEAN_FILES ${CMAKE_CURRENT_BINARY_DIR}/mruby/) - ENDIF (WITH_MRUBY) - IF (WITH_PICOTLS) - LIST(APPEND STANDALONE_SOURCE_FILES ${PICOTLS_SOURCE_FILES}) Index: patches/patch-deps_mruby_Rakefile === RCS file: patches/patch-deps_mruby_Rakefile diff -N patches/patch-deps_mruby_Rakefile --- patches/patch-deps_mruby_Rakefile 11 Mar 2022 20:09:54 - 1.2 +++ /dev/null 1 Jan 1970 00:00:00 - @@ -1,66 +0,0 @@ -Fix build with Ruby 3.0. - -Index: deps/mruby/Rakefile deps/mruby/Rakefile.orig -+++ deps/mruby/Rakefile -@@ -37,15 +37,15 @@ load "#{MRUBY_ROOT}/tasks/gitlab.rake" - task :default => :all - - bin_path = ENV['INSTALL_DIR'] || "#{MRUBY_ROOT}/bin" --FileUtils.mkdir_p bin_path, { :verbose => $verbose } -+FileUtils.mkdir_p bin_path, :verbose => $verbose - - depfiles = MRuby.targets['host'].bins.map do |bin| - install_path = MRuby.targets['host'].exefile("#{bin_path}/#{bin}") - source_path = MRuby.targets['host'].exefile("#{MRuby.targets['host'].build_dir}/bin/#{bin}") - - file install_path => source_path do |t| --FileUtils.rm_f t.name, { :verbose => $verbose } --FileUtils.cp t.prerequisites.first, t.name, { :verbose => $verbose } -+FileUtils.rm_f t.name, :verbose => $verbose -+F
Re: mruby 3.2.0
I have no problem removing the mruby flavor of h2o. I hope to be able to do that soon. -Otto On Mon, Feb 27, 2023 at 03:34:11PM +, Jeremy Evans wrote: > On 02/27 01:42, Stuart Henderson wrote: > > On 2023/02/25 21:42, Jeremy Evans wrote: > > > Simple update to the latest version of mruby. Tests pass on amd64, and > > > nothing depends on it. Looks like the 3.1.0 update was missed. > > > > www/h2o FLAVOR=mruby fails during link with this update. There are > > rather a lot of commits in h2o upstream relating to mruby including at > > least some to do with mruby 3.1.0 API, but there has been no upstream > > release in a long time > > (https://github.com/h2o/h2o/issues/2317#issuecomment-874828988) > > > > Might be simplest to remove the mruby flavour of h2o if nobody is > > actually using it... > > This is odd, because h20 does not declare a dependency on lang/mruby. > Might be fixable by adjusting the h20 build system to not pick up an > installed copy. > > However, there is no pressing need for the mruby update, it can wait for > a bit for otto@ (MAINTAINER) to fix. > > Thanks, > Jeremy > > > LD../../../build-amd64-mruby/mruby/host/bin/mirb > > mirb.c(/usr/obj/ports/h2o-2.2.6-mruby/build-amd64-mruby/mruby/host/mrbgems/mruby-bin-mirb/tools/mirb/mirb.o:(main)): > > warning: strcpy() is almost always misused, please use strlcpy() > > mirb.c(/usr/obj/ports/h2o-2.2.6-mruby/build-amd64-mruby/mruby/host/mrbgems/mruby-bin-mirb/tools/mirb/mirb.o:(main)): > > warning: strcat() is almost always misused, please use strlcat() > > ld: error: undefined symbol: mrb_exc_get_id > > >>> referenced by mruby_onig_regexp.c > > >>> mruby_onig_regexp.o:(onig_regexp_initialize) in archive > > >>> /usr/obj/ports/h2o-2.2.6-mruby/build-amd64-mruby/mruby/host/lib/libmruby.a > > >>> referenced by mruby_onig_regexp.c > > >>> mruby_onig_regexp.o:(onig_regexp_initialize) in archive > > >>> /usr/obj/ports/h2o-2.2.6-mruby/build-amd64-mruby/mruby/host/lib/libmruby.a > > >>> referenced by mruby_onig_regexp.c > > >>> mruby_onig_regexp.o:(onig_regexp_equal) in archive > > >>> /usr/obj/ports/h2o-2.2.6-mruby/build-amd64-mruby/mruby/host/lib/libmruby.a > > >>> referenced 8 more times > > > > ld: error: undefined symbol: mrb_ensure_string_type > > >>> referenced by mruby_onig_regexp.c > > >>> mruby_onig_regexp.o:(onig_regexp_match) in archive > > >>> /usr/obj/ports/h2o-2.2.6-mruby/build-amd64-mruby/mruby/host/lib/libmruby.a > > > > ld: error: undefined symbol: mrb_sym_name_len > > >>> referenced by mruby_onig_regexp.c > > >>> mruby_onig_regexp.o:(match_data_actual_index) in archive > > >>> /usr/obj/ports/h2o-2.2.6-mruby/build-amd64-mruby/mruby/host/lib/libmruby.a > > cc: error: linker command failed with exit code 1 (use -v to see invocation) > > mirb.c(/usr/obj/ports/h2o-2.2.6-mruby/build-amd64-mruby/mruby/host/mrbgems/mruby-bin-mirb/tools/mirb/mirb.o:(main)): > > warning: strcpy() is almost always misused, please use strlcpy() > > mirb.c(/usr/obj/ports/h2o-2.2.6-mruby/build-amd64-mruby/mruby/host/mrbgems/mruby-bin-mirb/tools/mirb/mirb.o:(main)): > > warning: strcat() is almost always misused, please use strlcat() > > ld: error: undefined symbol: mrb_exc_get_id > > >>> referenced by mruby_onig_regexp.c > > >>> mruby_onig_regexp.o:(onig_regexp_initialize) in archive > > >>> /usr/obj/ports/h2o-2.2.6-mruby/build-amd64-mruby/mruby/host/lib/libmruby.a > > >>> referenced by mruby_onig_regexp.c > > >>> mruby_onig_regexp.o:(onig_regexp_initialize) in archive > > >>> /usr/obj/ports/h2o-2.2.6-mruby/build-amd64-mruby/mruby/host/lib/libmruby.a > > >>> referenced by mruby_onig_regexp.c > > >>> mruby_onig_regexp.o:(onig_regexp_equal) in archive > > >>> /usr/obj/ports/h2o-2.2.6-mruby/build-amd64-mruby/mruby/host/lib/libmruby.a > > >>> referenced 8 more times > > > > ld: error: undefined symbol: mrb_ensure_string_type > > >>> referenced by mruby_onig_regexp.c > > >>> mruby_onig_regexp.o:(onig_regexp_match) in archive > > >>> /usr/obj/ports/h2o-2.2.6-mruby/build-amd64-mruby/mruby/host/lib/libmruby.a > > > > ld: error: undefined symbol: mrb_sym_name_len > > >>> referenced by mruby_onig_regexp.c > > >>> mruby_onig_regexp.o:(match_data_actual_index) in archive > > >>> /usr/obj/ports/h2o-2.2.6-mruby/build-amd64-mruby/mruby/host/lib/libmruby.a > > cc: error: linker command failed with exit code 1 (use -v to see invocation) > > rake aborted!
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: o...@cvs.openbsd.org2023/01/31 02:57:08 Modified files: net/powerdns_recursor: Makefile distinfo Log message: Update to PowerDNS Recursor 4.8.2
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: o...@cvs.openbsd.org2023/01/24 07:20:29 Modified files: security/john-jumbo: Makefile Added files: security/john-jumbo/patches: patch-src_nsec3_fmt_plug_c Log message: Add patch for https://github.com/openwall/john/pull/5185, already merged in upstream, allowing for empty NSEC3 salt ok sebastia@
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: o...@cvs.openbsd.org2023/01/20 05:23:01 Modified files: net/powerdns_recursor: Makefile distinfo Log message: Update to PowerDNS Recursor 4.8.1 https://docs.powerdns.com/recursor/security-advisories/powerdns-advisory-2023-01.html
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: o...@cvs.openbsd.org2022/12/12 06:00:26 Modified files: net/powerdns_recursor: Makefile distinfo Log message: Update to PowerDNS Recursor 4.8.0
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: o...@cvs.openbsd.org2022/12/09 03:23:42 Modified files: net/powerdns : Makefile distinfo Log message: Update to PowerDNS Authoritative Server 4.7.3
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: o...@cvs.openbsd.org2022/11/25 06:23:17 Modified files: net/powerdns_recursor: Makefile distinfo Log message: Update to PowerDNS Recursor 4.7.4
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: o...@cvs.openbsd.org2022/11/02 09:10:05 Modified files: net/dnsdist: Makefile distinfo Log message: Update to dnsdist 1.7.3
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: o...@cvs.openbsd.org2022/11/01 07:58:15 Modified files: net/powerdns : Makefile distinfo Log message: Update to Authoritative Server 4.7.2
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: o...@cvs.openbsd.org2022/10/31 09:35:07 Modified files: net/powerdns : Makefile distinfo net/powerdns/pkg: PLIST-main PLIST-mysql PLIST-pgsql Log message: Update to PowerDNS Authoritative Server 4.7.1
Re: powerdns auth: mandatory schema upgrade
Indeed 4.7.0 was released without the schema upgrade files. 4.7.1 will be released very soon with that fixed. I'll make sure to adapt the packaging after that. -Otto On Fri, Oct 28, 2022 at 09:18:04AM +0200, Florian Obser wrote: > Hi, > > there is a mandatory schema upgrade when upgrading from 4.6.x to 4.7.0 > > https://doc.powerdns.com/authoritative/upgrading.html > > | The new Catalog Zones feature comes with a mandatory schema change for > | the gsql database backends. See files named > | 4.3.x_to_4.7.0_schema.X.sql > > Without it, it fails to start with: > > pdns[17775]: Exiting because communicator thread died with error: virtual > void GSQLBackend::getUpdatedMasters(vector &, > std::unordered_set &, CatalogHashMap &) unable to retrieve list of > master domains: Fatal error during prePQpreparepare: select domains.id, > domains.name, domains.type, domains.notified_serial, domains.options, > domains.catalog, records.content from records join domains on > records.domain_id=domains.id and records.name=domains.name where > records.type='SOA' and records.disabled=false and domains.type in ('MASTER', > 'PRODUCER'): ERROR: column domains.options does not exist LINE 1: > ...ains.name, domains.type, domains.notified_serial, domains.op... > > Unfortunately the upstream tarball is missing > 4.3.0_to_4.7.0_schema.pgsql.sql and 4.3.0_to_4.7.0_schema.mysql.sql so > we also don't package it. I have reported this upstream: > https://mailman.powerdns.com/pipermail/pdns-users/2022-October/027898.html > > The files are here: > https://raw.githubusercontent.com/PowerDNS/pdns/master/modules/gpgsqlbackend/4.3.0_to_4.7.0_schema.pgsql.sql > https://raw.githubusercontent.com/PowerDNS/pdns/master/modules/gmysqlbackend/4.3.0_to_4.7.0_schema.mysql.sql > > No idea how much effort it would be to carry them as a local patch to > have them installed in /usr/local/share/doc/pdns/ > > Thanks, > Florian > > -- > I'm not entirely sure you are real.
Re: CVS: cvs.openbsd.org: ports
Woops, that was supposed to be: Upgrade to PowerDNS Auhthoritative Server 4.7.0 Mon, Oct 24, 2022 at 01:53:47AM -0600, Otto Moerbeek wrote: > CVSROOT: /cvs > Module name: ports > Changes by: o...@cvs.openbsd.org2022/10/24 01:53:47 > > Modified files: > net/powerdns : Makefile distinfo > > Log message: > >
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: o...@cvs.openbsd.org2022/10/24 01:53:47 Modified files: net/powerdns : Makefile distinfo Log message:
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: o...@cvs.openbsd.org2022/10/04 02:10:06 Modified files: net/libfstrm : Makefile distinfo net/libfstrm/pkg: PLIST Log message: Update to libfstrm 0.6.1; ok tb@
update libfstrm
Hi, AFAIK, this does not change the ABI or API. Is there a way to easily check that? Can this be committed as such? Thanks, -Otto Index: Makefile === RCS file: /cvs/ports/net/libfstrm/Makefile,v retrieving revision 1.3 diff -u -p -r1.3 Makefile --- Makefile11 Mar 2022 19:46:13 - 1.3 +++ Makefile4 Oct 2022 06:27:00 - @@ -1,6 +1,6 @@ COMMENT= C implementation of the Frame Streams protocol -DISTNAME= fstrm-0.5.0 +DISTNAME= fstrm-0.6.1 CATEGORIES=net devel SHARED_LIBS += fstrm0.0 Index: distinfo === RCS file: /cvs/ports/net/libfstrm/distinfo,v retrieving revision 1.1.1.1 diff -u -p -r1.1.1.1 distinfo --- distinfo2 Jun 2019 06:19:38 - 1.1.1.1 +++ distinfo4 Oct 2022 06:27:00 - @@ -1,2 +1,2 @@ -SHA256 (fstrm-0.5.0.tar.gz) = EO53kqhvrOHSJx3FkWUquMevaXaIOIfGn9sR8Q2hNfw= -SIZE (fstrm-0.5.0.tar.gz) = 470558 +SHA256 (fstrm-0.6.1.tar.gz) = vKSsHpgqLZI8zSTM4smPTO7tUAlpRDD3P8Dc68qPCY8= +SIZE (fstrm-0.6.1.tar.gz) = 471218
Re: Repeated gimp crashes?
On Tue, Sep 20, 2022 at 10:02:17AM +0100, Stuart Henderson wrote: > On 2022/09/19 12:40, Nam Nguyen wrote: > > >> 2. shift+s to resize > > >> 3. resize image with corners > > >> 4. gimp crashes > > > > > > I'm using the default datasize-cur limit for "staff" (4GB) and haven't > > > been able to provoke it into crashing by doing this yet (tried it with a > > > 8590x6000 TIFF). Wonder if something else is involved. > > > > > > If you do the above from "/bin/time -l gimp" (with raised limits so it > > > doesn't crash), what's the maximum rss that you see? > > > > $ /usr/bin/time -l gimp > >20.74 real 5.85 user 1.33 sys > > 796744 maximum resident set size > > hm. So how does raising datasize help if this is <800MB? > gimp could be allocating more than it uses in the end? -Otto
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: o...@cvs.openbsd.org2022/09/20 03:05:32 Modified files: net/powerdns_recursor: Makefile distinfo Log message: Update to PowerDNS Recursor 4.7.3
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: o...@cvs.openbsd.org2022/08/23 04:31:41 Modified files: net/powerdns_recursor: Makefile distinfo Log message: Update to PowerDNS Recursor 4.7.2. Including a medium severity security issue: https://docs.powerdns.com/recursor/security-advisories/powerdns-advisory-2022-02.html only relevant to people runinng with protobuf logging enabled.
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: o...@cvs.openbsd.org2022/07/13 06:26:30 Modified files: net/powerdns : Makefile distinfo Log message: Update to PowerDNS Authoritative Server 4.6.3
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: o...@cvs.openbsd.org2022/07/08 03:33:20 Modified files: net/powerdns_recursor: Makefile distinfo Log message: Update to PowerDNS Recursor 4.7.1
Re: CVS: cvs.openbsd.org: ports
On Mon, Jul 04, 2022 at 01:25:04PM +0200, Sol??ne Rapenne wrote: > Le Mon, 4 Jul 2022 05:22:00 -0600 (MDT), > Solene Rapenne a ??crit : > > > CVSROOT:/cvs > > Module name:ports > > Changes by: sol...@cvs.openbsd.org 2022/07/04 05:22:00 > > > > Modified files: > > sysutils/appmanager: Makefile distinfo > > sysutils/appmanager/pkg: DESCR PLIST > > > > Log message: > > > > > > cvs did commit while I thought I removed all the text in my commit > message... > > I updated it to 1.1.1, database is now built at runtime to prevent > discrepancy between the generated database (which wasn't updated except > if the package was rebuilt) and the actual package database. > > I wanted to abort the commit because I wanted to ask a review about the > changes, so if someone want to take a look, please go ahead, we can > still commit changes later. > Empty commit message convention is git specific. cvs only aborts the commit if you exit the editor session witout saving. e.g. :q! in vi. -Otto
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: o...@cvs.openbsd.org2022/06/14 04:08:58 Modified files: net/dnsdist: Makefile distinfo Log message: Update to dnsdist 1.7.2
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: o...@cvs.openbsd.org2022/05/30 08:21:04 Modified files: net/powerdns_recursor: Makefile distinfo net/powerdns_recursor/patches: patch-configure_ac Log message: Update to PowerDNS Recursor 4.7.0
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: o...@cvs.openbsd.org2022/05/02 00:14:55 Modified files: net/dnsdist: Makefile distinfo Removed files: net/dnsdist/patches: patch-dnsdist-lua-bindings_cc Log message: Update to dnsdist 1.7.1