summaryrefslogtreecommitdiff
AgeCommit message (Collapse)AuthorFilesLines
2023-03-31system/perl-test-fatal: upgrade to 0.017Síle Ekaterin Liszka1-3/+2
2023-03-31system/perl-libwww: upgrade to 6.68Síle Ekaterin Liszka1-2/+2
2023-03-31system/perl-io-socket-ssl: upgrade to 2.081Síle Ekaterin Liszka1-2/+2
2023-03-31system/perl-html-parser: upgrade to 3.81Síle Ekaterin Liszka1-2/+2
2023-03-31system/perl-cpanel-json-xs: upgrade to 4.36Síle Ekaterin Liszka1-2/+2
2023-03-31system/sed: upgrade to 4.9Síle Ekaterin Liszka1-2/+2
2023-03-31system/file: upgrade to 5.44Síle Ekaterin Liszka1-2/+2
2023-03-31user/apache-httpd: upgrade to 2.4.56 [CVEs]Síle Ekaterin Liszka1-2/+5
2023-03-31user/mbedtls: upgrade to 3.3.0 [CVEs]Síle Ekaterin Liszka1-2/+9
2023-03-31user/gvim: upgrade to 9.0.1385 [CVE]Síle Ekaterin Liszka1-3/+10
2023-03-31system/vim: upgrade to 9.0.1385 [CVE]Síle Ekaterin Liszka1-2/+9
2023-03-31system/rsync: upgrade to 3.2.7 [CVE]Síle Ekaterin Liszka1-2/+4
2023-03-31system/ncurses: upgrade to 6.4 [CVE]Síle Ekaterin Liszka28-232334/+7
2023-03-23user/swig: patch broken test. fixes #470.Zach van Rijn2-4/+29
This patch is in upstream v4.10.0 and up.
2023-03-23user/ol: patch out sys/cdefs.h check. fixes #974.Zach van Rijn2-2/+26
2023-03-22user/ol: update checksum. fixes #973.Zach van Rijn1-1/+1
2023-03-22user/ol: New packageYuriy Chumak1-0/+27
2023-03-21user/node: re-enable flaky tests.Zach van Rijn2-15/+1
These tests fail when /etc/hosts does not contain an IPv6 entry for localhost ('::1 localhost'). It isn't immediately clear to me whether the tests are therefore buggy, or if we want to put the blame on a non-conforming build environment. This was initially discovered on a system that only had an IPv4 entry ('127.0.0.1 localhost'), which occurred because the autobuilder tool copied /etc/{hosts,hostname,resolv.conf} from the host into the rootfs for each build environment, and each host may be running a different Linux distribution. This is being addressed in adelie-infra/autobuilder#33, which will ensure build environments are consistent in this way.
2023-03-21user/node: bump { 18.12.1 --> 18.15.0 }. temporarily disable flaky tests.Zach van Rijn5-109/+30
See #971. It is not clear whether these tests ever passed in their current form. Upstream changed these tests (and relevant code) since v16.15.0: https://github.com/nodejs/node/pull/41431 https://github.com/nodejs/node/issues/43014 In v16.15.0, these tests passed on this platform, but they have been modified in subsequent releases, and we've bumped twice since then: b282640c5353f37b706d3395718e80db244644b2 to v16.19.0 5d1083c1c688a496fbf9565046a1e22309cd9ad5 to v18.12.1 I don't see any binaries for aarch64 for either of these versions, suggesting we never rebuilt node on aarch64 after going to v18.12.1, likely due to the ARM builder being out of service between around 2022-12-27 and 2023-01-03. I think we just didn't catch this since bumping to v18.12.1. Temporarily disabling these tests to unblock progress; the bump to v18.15.0 is for some CVEs and expected to be safe.
2023-03-03user/gobject-introspection: disable parallel check. closes #776.Zach van Rijn1-1/+1
2023-02-28system/cmake: bump { 3.23.1 --> 3.23.5 }. add patch for 'git' file protocol.Zach van Rijn2-3/+27
CVE-2022-39253 is mitigated in 'git' >= v2.30.6, v2.31.5, v2.32.4, v2.33.5, v2.34.5, v2.35.5, v2.36.3, v2.37.4, v2.38.1 by changing some default protocol permissions. Two CMake tests fail without git config changes: * ExternalProject * CTest.UpdateGIT We backport CMake commit 79ce0f434e916684d734e136b92e14f472a9d14a to enable these tests to pass when using a more recent version of 'git'. CMake is also bumped to include minor internal bugfixes. Fixes #967.
2023-02-25user/s6-linux-utils: upgrade to 2.6.1.2, disable multicallLaurent Bercot1-4/+3
2023-02-25user/s6-portable-utils: upgrade to 2.3.0.1, disable multicallLaurent Bercot1-4/+3
2023-02-25system/s6-rc: upgrade to 0.5.4.1Laurent Bercot1-2/+2
2023-02-25system/execline: upgrade to 2.9.2.1, disable multicallLaurent Bercot1-3/+2
2023-02-24system/sysvinit: Update to 3.06, remove logsaveA. Wilcox1-3/+9
The `logsave` command is now provided by e2fsprogs. Fixes: #938
2023-02-24user/gtk4: Update to 4.8.3A. Wilcox1-2/+3
This, combined with Harfbuzz 2.6.8, builds properly on ppc. Fixes: #962
2023-02-24user/harfbuzz: Update to 2.6.8A. Wilcox1-3/+3
2023-02-24user/wayland: Update to 1.21.0A. Wilcox1-4/+4
2023-02-23user/dovecot: Fix cache flush on 32-bit big endianA. Wilcox2-0/+15
They already had a workaround for 64-bit BE. Since time_t is 64-bit on all musl platforms, we need it active at all times. Fixes: #966
2023-02-17user/s6-linux-utils: upgrade to 2.6.1.0Laurent Bercot1-4/+5
2023-02-17user/s6-portable-utils: upgrade to 2.3.0.0Laurent Bercot1-4/+5
2023-02-17user/s6-networking: upgrade to 2.5.1.3Laurent Bercot1-2/+2
2023-02-17system/mdevd: upgrade to 0.1.6.2Laurent Bercot1-2/+2
2023-02-17system/s6-linux-init: upgrade to 1.1.0.0Laurent Bercot1-4/+4
2023-02-17system/s6-rc: upgrade to 0.5.4.0Laurent Bercot1-2/+2
2023-02-17system/s6: upgrade to 2.11.3.0Laurent Bercot1-3/+3
2023-02-17system/execline: upgrade to 2.9.2.0Laurent Bercot1-2/+3
2023-02-17system/utmps: move utmpd to stage 2Laurent Bercot2-12/+14
2023-02-17system/skalibs: upgrade to 2.13.1.0Laurent Bercot1-2/+2
2023-02-15system/git: bump { 2.36.0 --> 2.39.2 }.Zach van Rijn1-2/+2
2023-02-12user/clucene: Retire to legacy/A. Wilcox11-0/+0
No upstream commits in years. No response to bug reports from other teams. Really weird test failures on 32-bit architectures, likely caused by interactions with signedness and 64-bit integers. Ref: #873 #880
2023-02-12user/dovecot: Remove CLucene FTS supportA. Wilcox1-14/+6
We still have solr FTS support, so that should be enough. Didn't bump pkgrel because it hasn't been built/uploaded anywhere yet.
2023-02-12user/sword: Update to 1.9.0, fix tests, use XapianA. Wilcox1-4/+9
* Tests were not actually being run. Fix that. * Port away from CLucene, which is broken on multiple architectures, to the Xapian full-text search (FTS) engine. See #873.
2023-02-11user/clucene: More fixesA. Wilcox7-7/+315
This doesn't fix the test failure, but is more correct.
2023-02-10user/perl-type-tiny: add patch for 32-bit regression. fixes #964.Zach van Rijn2-3/+41
2023-02-09dovecot 2.3.20Lee Starnes1-5/+5
2023-02-09system/libuv: apply upstream tty size patch. actually fixes #695.Zach van Rijn3-29/+31
2023-02-08system/perl: use different test server. fixes #960.Zach van Rijn2-1/+17
2023-02-08system/libuv: add patch for nonsense test assumption. fixes #695.Zach van Rijn2-2/+32