CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: gkoeh...@cvs.openbsd.org2024/04/13 22:21:30 Modified files: sysutils/u-boot/rk356x: Makefile sysutils/u-boot/rk356x/files: rkbinpatch.c sysutils/u-boot/rk3588: Makefile sysutils/u-boot/rk3588/files: rkbinpatch.c Added files: sysutils/u-boot/rk356x/patches: patch-tools_rkcommon_c sysutils/u-boot/rk3588/patches: patch-tools_rkcommon_c Log message: Unbreak sysutils/u-boot/{rk356x,rk3588} when big-endian Add some swaps between little-endian and host byte order, which allows my big-endian powerpc to slowly build arm64 rockchip firmware. kurt@ checked that a rock5b boots with firmware from big-endian powerpc and from little-endian amd64. ok kurt@ kettenis@
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: gkoeh...@cvs.openbsd.org2024/04/01 23:11:50 Modified files: graphics/decker: Makefile distinfo Log message: update graphics/decker to 1.41, from Jag Talon (maintainer) https://internet-janitor.itch.io/decker/devlog/698201/decker-141
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: gkoeh...@cvs.openbsd.org2024/04/01 22:45:17 Modified files: math/gbc : Makefile Added files: math/gbc/patches: patch-bc_scan_l Log message: patch math/gbc for write overflow bc/scan.c wrote 8 bytes to a 4-byte int on LP64_ARCHS. This broke the build on big-endian powerpc64, which was reading the wrong end of the oversize value. It might have caused a bus error on sparc64. ok Ivan Ivanov (maintainer)
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: gkoeh...@cvs.openbsd.org2024/03/25 16:31:05 Modified files: devel/arm-none-eabi/gcc: Makefile devel/riscv-elf/gcc: Makefile devel/riscv32-esp-elf/gcc-bootstrap: Makefile devel/xtensa-esp32s2-elf/gcc-bootstrap: Makefile devel/xtensa-esp32s3-elf/gcc-bootstrap: Makefile devel/xtensa-lx106-elf/gcc-bootstrap: Makefile Log message: Mark some cross-gcc ports BROKEN-powerpc64 These are failing to apply patches/vecstep-*; either the patch is rejected, or the patch file is missing. See the powerpc64 bulk build report at https://marc.info/?l=openbsd-ports=171109047629941=2
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: gkoeh...@cvs.openbsd.org2024/03/25 13:12:41 Modified files: lang/gambit: Makefile Added files: lang/gambit/patches: patch-include_gambit_h_in Log message: lang/gambit: disable musttail on mips64, powerpc, powerpc64 gambit-4.9.5 has a fatal error in clang at __attribute__((musttail)) on these platforms. Disable musttail if the compiler defines __mips__ or __powerpc__, then I can build gambit and run gsi(1) on powerpc and powerpc64. https://github.com/gambit/gambit/issues/898 ok juanfra@ (maintainer)
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: gkoeh...@cvs.openbsd.org2024/03/04 18:52:34 Modified files: infrastructure/mk: arch-defines.mk Log message: bump _SYSTEM_VERISON-powerpc for clang fix: CFI for cr2, cr3, cr4 The macppc snap with the Mar 4 kernel has a fixed base-clang. The bump affects only powerpc packages, and will cause pkg_add -u to update to fixed packages, where the CFI might unbreak a C++ exception.
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: gkoeh...@cvs.openbsd.org2024/03/04 13:31:16 Modified files: devel/llvm/13 : Makefile devel/llvm/13/patches: patch-llvm_lib_Target_PowerPC_PPCFrameLowering_cpp devel/llvm/16 : Makefile devel/llvm/16/patches: patch-llvm_lib_Target_PowerPC_PPCFrameLowering_cpp devel/llvm/17 : Makefile devel/llvm/17/patches: patch-llvm_lib_Target_PowerPC_PPCFrameLowering_cpp Log message: sync with base: For macppc, add missing CFI for cr2, cr3, cr4 clang -S was missing a line like ".cfi_offset cr2, -16" in functions that spill cr2 (or cr3, cr4) to the stack. This was breaking a few C++ exceptions. This fix adds the missing CFI. https://github.com/llvm/llvm-project/pull/83098 ok robert@ (maintainer)
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: gkoeh...@cvs.openbsd.org2024/02/18 14:55:41 Modified files: graphics/decker: Makefile distinfo graphics/decker/pkg: PLIST Log message: Update to decker 1.39 from Jag Talon (maintainer)
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: gkoeh...@cvs.openbsd.org2023/12/23 17:26:12 Modified files: devel/boost: Makefile Added files: devel/boost/patches: patch-boost_fiber_detail_cpu_relax_hpp Log message: Unbreak devel/boost on powerpc, powerpc64 was picking the wrong syntax for PowerPC asm. ok tb@
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: gkoeh...@cvs.openbsd.org2023/10/22 12:23:42 Modified files: devel/swig : Makefile devel/swig/pkg : PLIST Log message: Configure swig without ruby; stop installing examples This prevents a dependency cycle (devel/swig -> lang/ruby/3.2 -> lang/rust -> devel/llvm/16 -> swig) if ruby enables yjit and rust uses our llvm. Our swig now skips ruby during "make test", but can still generate ruby interfaces (as in devel/subversion, textproc/libmarisa, textproc/redland-bindings). Keep guile, perl, python, tcl; so "make test" tries those languages. Configure without any other language. This would affect the Makefiles under /usr/local/share/examples/swig, but they were already broken (by failing to run WRKSRC/swig), so stop installing the examples. --with-php=${MODPHP_BIN} didn't find php because lang/php wasn't in MODULES; change to --without-php. bulk build by sthen@ ok sthen@ tb@ semarie@
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: gkoeh...@cvs.openbsd.org2023/08/29 12:25:16 Modified files: graphics/imlib2: Makefile graphics/imlib2/pkg: PLIST Log message: Unbreak graphics/imlib2 for arch without rust If there's no rust compiler, then x11/gnome/librsvg is an old version and fails imlib2's check for librsvg-2.0 >= 2.46; the built imlib2 is missing svg.so and was failing to package. Unbreak with a conditional comment in PLIST, so we can package imlib2 without svg.so if we don't have rust, or with svg.so if we do. Also add sharpyuv to WANTLIB. I redid WANTLIB on a rust arch; the trick is that "make RUST_ARCHS=" pretends to not have rust, so I can twice run port-lib-depends-check (without then with rust). help from semarie@ ok sdk@ older diff ok op@
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: gkoeh...@cvs.openbsd.org2023/08/08 14:40:55 Modified files: www/firefox-esr: Makefile www/firefox-esr/files: unveil.content www/mozilla-firefox: Makefile www/mozilla-firefox/files: unveil.content Log message: Unveil /etc/localtime r in firefox's content process This fixes dst (daylight saving time, summer time) in JavaScript. Firefox turns off dst in libc, then applies its dst rules. It fails to apply dst if it can't realpath(3) /etc/localtime. This failure might cause local times to be off by an hour. ok landry@ (maintainer)
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: gkoeh...@cvs.openbsd.org2023/08/02 21:50:23 Modified files: games/openttd : Makefile distinfo Log message: Update games/openttd to 13.4 Florian Viehweger (maintainer) changed MASTER_SITES. sthen@ changed WANTLIB and COMPILER. solene@ and I both added sthen's changes to Florian's diff and got a similar result. ok solene@
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: gkoeh...@cvs.openbsd.org2023/07/16 18:39:22 Modified files: misc : Makefile Log message: +p5-Spreadsheet-XLSX
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: gkoeh...@cvs.openbsd.org2023/07/16 18:36:59 Log message: Import misc/p5-Spreadsheet-XLSX from Wen Heping ok sthen@ This Perl module is a emulation of Spreadsheet::ParseExcel for Excel 2007 (.xlsx) file format. It supports styles and many of Excel's quirks, but not all. It populates the classes from Spreadsheet::ParseExcel for interoperability; including Workbook, Worksheet, and Cell. Status: Vendor Tag: gkoehler Release Tags: gkoehler_20230716 N ports/misc/p5-Spreadsheet-XLSX/Makefile N ports/misc/p5-Spreadsheet-XLSX/distinfo N ports/misc/p5-Spreadsheet-XLSX/pkg/DESCR N ports/misc/p5-Spreadsheet-XLSX/pkg/PLIST No conflicts created by this import
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: gkoeh...@cvs.openbsd.org2023/07/16 18:30:08 Modified files: textproc : Makefile Log message: +p5-String-Util
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: gkoeh...@cvs.openbsd.org2023/07/16 18:24:15 Log message: Import textproc/p5-String-Util from Wen Heping ok sthen@ String::Util provides a collection of small, handy functions for processing strings in various ways. Status: Vendor Tag: gkoehler Release Tags: gkoehler_20230716 N ports/textproc/p5-String-Util/Makefile N ports/textproc/p5-String-Util/distinfo N ports/textproc/p5-String-Util/pkg/DESCR N ports/textproc/p5-String-Util/pkg/PLIST No conflicts created by this import
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: gkoeh...@cvs.openbsd.org2023/07/16 18:10:06 Modified files: www: Makefile Log message: +p5-HTML-Selector-XPath +p5-Web-Scraper
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: gkoeh...@cvs.openbsd.org2023/07/16 18:03:59 Log message: Import www/p5-Web-Scraper from Wen Heping ok sthen@ Web::Scraper is a web scraper toolkit, inspired by Ruby's equivalent Scrapi. It provides a DSL-ish interface for traversing HTML documents and returning a neatly arranged Perl data structure. Status: Vendor Tag: gkoehler Release Tags: gkoehler_20230716 N ports/www/p5-Web-Scraper/Makefile N ports/www/p5-Web-Scraper/distinfo N ports/www/p5-Web-Scraper/pkg/DESCR N ports/www/p5-Web-Scraper/pkg/PLIST No conflicts created by this import
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: gkoeh...@cvs.openbsd.org2023/07/16 17:58:51 Log message: Import www/p5-HTML-Selector-XPath from Wen Heping ok sthen@ HTML::Selector::XPath is a utility function to compile full set of CSS2 and partial CSS3 selectors to the equivalent XPath expression. Status: Vendor Tag: gkoehler Release Tags: gkoehler_20230716 N ports/www/p5-HTML-Selector-XPath/Makefile N ports/www/p5-HTML-Selector-XPath/distinfo N ports/www/p5-HTML-Selector-XPath/pkg/DESCR N ports/www/p5-HTML-Selector-XPath/pkg/PLIST No conflicts created by this import
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: gkoeh...@cvs.openbsd.org2023/07/14 20:28:41 Modified files: x11/gnome/librsvg: distinfo.old Log message: Allow dpb to build librsvg 2.40.21 again If you ran dpb on an arch without rust, it got E=x11/gnome/librsvg "Missing info for gnome/librsvg-2.56.2.tar.xz in .../distinfo.old", because the info was for 2.56.1. It checks both files (2.40.21 and 2.56.2), though the build extracts only 2.40.21. ok phessler@
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: gkoeh...@cvs.openbsd.org2023/07/08 20:10:08 Modified files: graphics/libvips: Makefile Log message: Add -Dfontconfig=disabled to unbreak build without pango In libvips, our -Dpangocairo=disabled also disabled fontconfig, but only after meson.build required both fontconfig and pangoft2. If you didn't install devel/pango, then "make configure" failed at pangoft2. If you did have pango, then pangoft2 got found but never used. Diff from Brad Smith
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: gkoeh...@cvs.openbsd.org2023/07/05 13:08:24 Modified files: infrastructure/bin: clean-old-distfiles Log message: Don't parse undef as a date ajacoutot@ got a fatal error from running clean-old-distfiles without a timestamp argument. ok espie@ Cvs: --
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: gkoeh...@cvs.openbsd.org2023/05/26 19:10:48 Modified files: editors/poke : Makefile Log message: Don't detect emacs or xemacs in configure The build broke if I had installed editors/xemacs (and not editors/emacs), when it tried to run xemacs to compile some Lisp. ok fcambus@ (maintainer)
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: gkoeh...@cvs.openbsd.org2023/03/18 22:02:17 Modified files: lang/node : Makefile lang/node/patches: patch-common_gypi patch-include_node_common_gypi Log message: Unbreak node for powerpc64 by dropping -mminimal-toc from Volker Schlecht (maintainer) checked by mbuhl@ ok sthen@ naddy@
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: gkoeh...@cvs.openbsd.org2022/12/29 14:20:25 Modified files: lang/ruby : Makefile.inc Log message: Configure lang/ruby --without-baseruby This prevents picking /usr/local/bin/ruby from devel/ruby-shims as baseruby, and erroring in "make fake" when baseruby can't require tool/downloader.rb. We don't need baseruby, because we are building ruby-*.tar.gz, not a git checkout of ruby. ok jeremy@ (maintainer)
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: gkoeh...@cvs.openbsd.org2022/11/20 17:04:04 Removed files: devel/libdsm/patches: patch-configure Log message: Forgot a patch when removing devel/libdsm ok tb@
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: gkoeh...@cvs.openbsd.org2022/11/06 13:36:12 Modified files: sysutils/broot : Makefile crates.inc distinfo Log message: Update to broot 1.16.2, unbreaking powerpc64 Crate exr < 1.5.0 failed to build on big-endian arch like powerpc64. This update moves from exr 1.4.2 to 1.5.2. https://github.com/Canop/broot/blob/master/CHANGELOG.md ok landry@
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: gkoeh...@cvs.openbsd.org2022/10/28 14:47:06 Modified files: lang/pcc/pcc : Makefile lang/pcc/pcc/patches: patch-cc_cc_cc_c lang/pcc/pcc-libs: Makefile Added files: lang/pcc/pcc/patches: patch-arch_powerpc_local2_c patch-arch_powerpc_local_c patch-arch_powerpc_macdefs_h patch-arch_powerpc_table_c lang/pcc/pcc-libs/patches: patch-libpcc_cxmuldiv_c patch-libsoftfloat_arch_powerpc_powerpc-gcc_h Log message: Minimally fix pcc for powerpc This unbreaks the powerpc package of lang/pcc/pcc-libs and puts enough fixes in lang/pcc/pcc for powerpc to compile the simplest C program, $ cat hello.c #include int main(void) { puts("Hello, BSD!"); return 0; } $ pcc -O2 -o hello hello.c It can't compile much more; pcc's powerpc code generator still fails with static function calls and with floating-point literals. ok sthen@
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: gkoeh...@cvs.openbsd.org2022/09/25 10:34:20 Modified files: games/mvdsv: Makefile games/mvdsv/patches: patch-CMakeLists_txt Log message: Unbreak games/mvdsv for big-endian platforms Add a missing PRIVATE keyword to CMakeLists.txt; this allows my powerpc and kmos's sparc64 to package mvdsv. ok kmos@ jca@ sthen@
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: gkoeh...@cvs.openbsd.org2022/09/22 20:28:27 Modified files: plan9/plan9port: Makefile Log message: Build plan9port on i386, powerpc, powerpc64 Diff from Kyle Milz, who confirms that plan9port works on i386 and powerpc, but doesn't build on sparc64. I add powerpc64.
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: gkoeh...@cvs.openbsd.org2022/09/22 20:03:30 Modified files: graphics/webp-pixbuf-loader: Makefile Added files: graphics/webp-pixbuf-loader/patches: patch-io-webp_c Log message: Unbreak webp-pixbuf-loader for big-endian platforms The patch calls le32toh in , because OpenBSD doesn't have . This prevents a fatal error on BE_ARCHS when byteswap.h didn't exist. ok jsg@ jca@
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: gkoeh...@cvs.openbsd.org2022/08/11 14:50:12 Modified files: graphics/freeimage: Makefile Added files: graphics/freeimage/patches: patch-Source_FreeImage_PluginBMP_cpp patch-Source_FreeImage_PluginDDS_cpp Log message: Unbreak freeimage on BE_ARCHS PluginBMP.cpp had compiler errors before I added this patch. fixed PluginBMP, PluginDDS for compilation under Big Endian OS (see FREEIMAGE_BIGENDIAN) https://sourceforge.net/p/freeimage/svn/1809/ ok tb@
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: gkoeh...@cvs.openbsd.org2022/08/07 11:00:43 Modified files: lang/python: Makefile.inc Log message: Don't --enable-optimizations on powerpc This unbreaks python 3.* on powerpc. The build had enabled PGO (Profile Guided Optimization), then got "Segmentation fault" from the ./python binary. This seems to be a bug in clang or lld specific to 32-bit powerpc. We continue to --enable-optimizations on powerpc64. ok kmos@, sthen@
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: gkoeh...@cvs.openbsd.org2022/08/01 09:00:04 Modified files: graphics/png : Makefile Log message: Unbreak png on powerpc64 by passing -mvsx Diff from Brad Smith, edited by me
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: gkoeh...@cvs.openbsd.org2022/02/16 15:37:31 Modified files: www/webkitgtk4 : Makefile Log message: Stop using -DUSE_OPENGL_OR_ES=OFF on powerpc This arg breaks the build of webkitgtk4, because this arg disables the coordinated graphics, but https://commits.webkit.org/238823@main added code that requires the coordinated graphics. After unbreaking webkitgtk4, I can browse a few websites in www/midori on macppc. Most websites crash their tabs, in the same way that glxgears crashes, "LLVM ERROR: Relocation type not implemented yet!" ok ajacoutot@ (maintainer)
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: gkoeh...@cvs.openbsd.org2021/10/17 18:25:49 Modified files: lang/gcc/11: Makefile lang/gcc/11/patches: patch-gcc_config_gcc patch-libphobos_libdruntime_core_sys_posix_setjmp_d lang/gcc/11/pkg: PLIST-main Added files: lang/gcc/11/patches: patch-libphobos_configure patch-libphobos_libdruntime_config_powerpc_switchcontext_S patch-libphobos_libdruntime_core_sys_posix_config_d patch-libphobos_libdruntime_core_sys_posix_sys_mman_d patch-libphobos_libdruntime_core_sys_posix_ucontext_d patch-libphobos_libdruntime_core_thread_fiber_d lang/gcc/11/pkg: PFRAG.powerpc64-main Log message: Fix fibers in dlang; unbreak powerpc64; add D for powerpc64 Unbreak gcc 11 on powerpc64 by adding a missing header file (gnu-user.h in patch-gcc_config_gcc). Add PFRAG.powerpc64-main. The other changes are for D. Add powerpc64 to ONLY_FOR_ARCHS-dlang, and add enough PPC64 code to build libphobos. While adding fibers for powerpc64, also fix fibers for 32-bit powerpc, and pass MAP_STACK to mmap(2) for fibers on all archs. ok pascal@ (maintainer)
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: gkoeh...@cvs.openbsd.org2021/09/10 14:29:04 Modified files: devel/gmp : Makefile Added files: devel/gmp/patches: patch-mpn_asm-defs_m4 Log message: Fix libgmp.a on powerpc; text relocation error from ld.lld This was a mistake in asm-defs.m4: `PIC_ALWAYS' should not have quotes. Delete the quotes to fix the powerpc asm. ok naddy@ (maintainer)
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: gkoeh...@cvs.openbsd.org2021/08/24 12:13:19 Modified files: lang/gcc/11: Makefile lang/gcc/8 : Makefile Log message: Unbreak lang/gcc/11 on powerpc with USE_LLD = No It broke when ld.lld became the default powerpc linker. USE_LLD = No is the same workaround as in gcc 8. Add a comment to both 8 and 11, # powerpc lld can't link C++ code from gcc, "ld: error: relocation # refers to a discarded section:" .data.rel.ro.* or # .sdata.DW.ref.__gxx_personality_v0 USE_LLD = No ok pascal@ (maintainer)
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: gkoeh...@cvs.openbsd.org2021/08/21 11:18:28 Modified files: games/cataclysm-dda: Makefile Log message: Drop -Wl,-relax for powerpc, now that lld is the default linker. Thanks to Brad Smith for noticing that I missed -Wl,-relax here, and providing this diff. With -Wl,-relax, there is no chance of a powerpc package of cataclysm-dda after the switch to lld.
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: gkoeh...@cvs.openbsd.org2021/08/20 21:25:05 Modified files: devel/clang-tools-extra: Makefile devel/llvm : Makefile games/godot: Makefile games/scummvm : Makefile infrastructure/mk: arch-defines.mk lang/gcc/8 : Makefile net/wireshark : Makefile www/webkitgtk4 : Makefile x11/qt5: Makefile.inc Log message: Switch macppc to ld.lld Add powerpc to LLD_ARCH, and bump _SYSTEM_VERSION-powerpc so people will upgrade to lld-linked packages when they exist. Remove -Wl,--relax from several ports; ld.bfd needed this flag, but ld.lld can't take it. lang/gcc/8 needs USE_LLD = No, because LLD can't link C++ code from gcc (error: relocation refers to a discarded section: ...), and also, the adastrap is too old. To avoid linker errors on macppc, you need a snapshot where /usr/bin/ld is LLD. ok sthen@
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: gkoeh...@cvs.openbsd.org2021/08/09 14:13:02 Modified files: lang/ruby : Makefile.inc lang/ruby/2.6 : Makefile lang/ruby/2.7 : Makefile lang/ruby/3.0 : Makefile Log message: Don't use __builtin_setjmp on powerpc, powerpc64 clang-11's __builtin_setjmp is broken, has chance of SIGSEGV during "make build" on powerpc with ld.lld, or when passing a wrong option (like "ruby -e" with no -e code) on powerpc64. ok jeremy@ (maintainer)
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: gkoeh...@cvs.openbsd.org2021/03/10 23:29:21 Modified files: math/libtommath: Makefile distinfo math/libtommath/pkg: PLIST Added files: math/libtommath/patches: patch-bn_mp_set_double_c Removed files: math/libtommath/patches: patch-makefile_include_mk Log message: Update libtommath to 1.2.0 There are new functions in ; some of them are needed to update lang/moarvm. The headers tommath_{class,superclass}.h are no longer installed. ok pascal@
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: gkoeh...@cvs.openbsd.org2021/02/19 20:46:49 Modified files: misc/hfsplus : Makefile misc/hfsplus/patches: patch-libhfsp_src_btreecheck_c patch-libhfsp_src_fscheck_c patch-libhfsp_src_record_c patch-libhfsp_src_volume_c patch-src_hpls_c Added files: misc/hfsplus/patches: patch-libhfsp_src_apple_h patch-libhfsp_src_fscheck_h patch-src_copyout_c patch-src_hpcache_c Log message: fix -fno-common, LP64_ARCHS, HFS+ without wrapper misc/hfsplus is old code from 2002; failed to build with -fno-common; assumed sizeof(long) == 4, which is wrong on LP64_ARCHS; failed when the HFS+ has no HFS wrapper. Thaison Nguyen reported that hpmount always gave an error. - Simplify Makefile by setting AUTORECONF. - Fix LP64_ARCHS by changing UInt32 and like types from long to int. Change many "%ld" to "%d" in printf()s to match. - Apply a patch from Gentoo to fix -fno-common. - Fix HFS+ with no HFS wrapper by setting vol->maxblocks. - In src/copyout.c, fix passing to int *lenptr. ok tb@
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: gkoeh...@cvs.openbsd.org2021/02/08 19:12:43 Modified files: lang/gcc/8 : Makefile lang/gcc/8/patches: patch-gcc_config_rs6000_openbsd64_h Log message: Set powerpc64 LINK_SPEC to same as powerpc ok pascal@ (maintainer) ok kettenis@
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: gkoeh...@cvs.openbsd.org2021/01/23 21:41:51 Modified files: lang/gcc/8 : Makefile lang/gcc/8/patches: patch-gcc_config_aarch64_openbsd_h patch-gcc_config_alpha_openbsd_h patch-gcc_config_arm_openbsd_h patch-gcc_config_i386_openbsdelf_h patch-gcc_config_mips_openbsd_h patch-gcc_config_pa_pa-openbsd_h patch-gcc_config_rs6000_openbsd_h patch-gcc_config_sparc_openbsd64_h Log message: Edit LINK_SPEC, fix static linking with ports-gcc egcc -static was broken on most arches; the static binary wrongly requested ld.so(1), but /usr/lib/rcrt0.o wants to do the relocation without ld.so(1). The fix is to drop ld.so, %{!static:-dynamic-linker /usr/libexec/ld.so} Remove "%{!dynamic-linker:...}" because gcc -dynamic-linker is not an option. Correct "%{!r*:...}" to "%{!r:...}" because gcc -r doesn't take an option argument. Use "%{!nostdlib:-L/usr/lib}" on lld arches, like base-gcc. Edit the powerpc LINK_SPEC in rs6000/openbsd.h to be like the others. Add "%{!nostdlib:-L/usr/lib}" so I can try lld on powerpc. For now, don't edit the powerpc64 LINK_SPEC. ok pascal@ (maintainer)
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: gkoeh...@cvs.openbsd.org2021/01/15 21:44:30 Modified files: devel : Makefile Log message: +ruby-sync
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: gkoeh...@cvs.openbsd.org2021/01/15 21:27:07 Log message: two-phase lock with a counter needed by sysutils/ruby-puppet/5 on ruby 2.7 ok kn@ giovanni@ Status: Vendor Tag: gkoehler Release Tags: gkoehler_20210115 N ports/devel/ruby-sync/Makefile N ports/devel/ruby-sync/distinfo N ports/devel/ruby-sync/pkg/DESCR N ports/devel/ruby-sync/pkg/PLIST No conflicts created by this import
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: gkoeh...@cvs.openbsd.org2020/12/12 19:42:17 Modified files: lang/squeak: Makefile.inc lang/squeak/funsqueak: Makefile lang/squeak/image: Makefile lang/squeak/sources: Makefile lang/squeak/vm : Makefile lang/squeak/vm/patches: patch-platforms_unix_config_acinclude_m4 patch-platforms_unix_vm_sqUnixMain_c Added files: lang/squeak/vm/patches: patch-platforms_unix_config_Makefile_in patch-platforms_unix_plugins_FloatMathPlugin_acinclude_m4 Log message: Unbreak squeak-vm on powerpc, sparc64 Most of this diff is from Anthony Richardby, who provided the patches to unbreak macppc by removing the -mno-fused-madd flag, fixing an asm error, and changing the order of the linker arguments. kmos@ confirmed that the patches also unbreak sparc64. cwen@ moved the HOMEPAGE to https. I removed flags like -O3. ok cwen@
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: gkoeh...@cvs.openbsd.org2020/11/15 15:05:10 Modified files: games/godot: Makefile www/webkitgtk4 : Makefile Log message: Remove last uses of -mlongcall on powerpc -Wl,--relax solves linker errors of "relocation truncated to fit", -mlongcall isn't necessary. diff from Brad Smith, ok cwen@
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: gkoeh...@cvs.openbsd.org2020/09/20 22:02:00 Modified files: lang/tinyscheme: Makefile distinfo Log message: Simple update to tinyscheme-1.42; from wen heping.
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: gkoeh...@cvs.openbsd.org2020/08/24 15:41:21 Modified files: net/clic : Makefile distinfo Log message: Update to clic-1.1.5 Committing this for solene@, who did all the work. The update is necessary to use clic with ecl-20.4.24.
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: gkoeh...@cvs.openbsd.org2020/08/24 15:38:06 Modified files: lang/ecl : Makefile distinfo lang/ecl/patches: patch-src_Makefile_in patch-src_configure lang/ecl/pkg : PLIST Added files: lang/ecl/patches: patch-src_c_package_d Removed files: lang/ecl/patches: patch-src_c_ffi_d patch-src_c_numbers_atan_d Log message: Update to ecl-20.4.24 >From Josh Elsasser; thank you! I'm adding patch-src_c_package_d (from Daniel Kochmaski) to work around a SIGSEGV on 64-bit big-endian platforms, so I can have Common Lisp on powerpc64. Josh Elsasser built ecl on mips64 and sparc64.
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: gkoeh...@cvs.openbsd.org2020/08/24 15:31:13 Modified files: devel/boehm-gc : Makefile devel/boehm-gc/patches: patch-include_private_gcconfig_h patch-libatomic_ops_src_atomic_ops_sysdeps_gcc_sparc_h Log message: Patch boehm-gc to avoid sbrk, fix powerpc64, help sparc64. Add AO_fetch_compare_and_swap for sparc64, because lang/ecl 20.4.24 will need it. Thanks to kmos@ for discovering this problem, and to Josh Elsasser for building this diff on a sparc64.
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: gkoeh...@cvs.openbsd.org2020/08/03 14:36:29 Modified files: lang/crystal : Makefile Log message: Use llvm-config from ports, not from base Base has /usr/bin/llvm-config after the switch to LLVM 10, so crystal would run the wrong llvm-config and fail to build (with a linker error). The bootstrap crystal.o needs LLVM 8. Using llvm-config from ports works for now, because ports devel/llvm is still LLVM 8. ok sthen@
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: gkoeh...@cvs.openbsd.org2020/06/15 15:34:58 Modified files: devel/llvm : Makefile devel/llvm/patches: patch-lib_Target_Mips_MipsReturnProtectorLowering_cpp patch-lib_Target_PowerPC_PPCISelLowering_cpp patch-tools_clang_lib_Driver_ToolChains_Clang_cpp Added files: devel/llvm/patches: patch-lib_Target_PowerPC_PPCTargetMachine_cpp patch-tools_clang_lib_Basic_Targets_OSTargets_h patch-tools_clang_lib_Basic_Targets_cpp Log message: Mips, PowerPC changes from base-clang Mips: Use a distinct trapcode with retguard on mips64. PowerPC: - Set max atomic size for AtomicExpandPass. - For OpenBSD/powerpc64: switch to ELFv2, use _mcount profiling, and add our OpenBSD specific defines. - Optimize away the frame pointer. ok jca@ (maintainer), cwen@
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: gkoeh...@cvs.openbsd.org2020/04/29 13:04:02 Modified files: lang/clisp : Makefile Log message: Override some configure tests for mmap, shm. sthen@ reported that clisp sometimes fails to build, with an error at MAP_ANON. Some tests, including MAP_ANON, might give a random 'no' when their fixed addresses conflict with ASLR. Override to 'yes'. ok sthen@
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: gkoeh...@cvs.openbsd.org2020/03/04 17:50:00 Modified files: devel/llvm : Makefile devel/llvm/patches: patch-tools_clang_include_clang_Driver_Options_td patch-tools_clang_lib_Driver_ToolChains_Clang_cpp patch-tools_clang_lib_Frontend_CompilerInvocation_cpp Added files: devel/llvm/patches: patch-tools_clang_lib_CodeGen_TargetInfo_cpp Log message: Add -msvr4-struct-return to ports-clang for powerpc. This is the same change as in base-clang. ok jca@ (maintainer)
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: gkoeh...@cvs.openbsd.org2020/02/18 11:11:33 Modified files: graphics/babl : Makefile Added files: graphics/babl/patches: patch-babl_babl-fish-path_c Log message: Fix crash on powerpc Some code was not portable to powerpc because it passed the wrong type of value to a va_arg(3) function; upstream commit https://gitlab.gnome.org/GNOME/babl/commit/a585692 "Looks good" ajacoutot@ (maintainer) ok cwen@
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: gkoeh...@cvs.openbsd.org2020/02/04 15:50:04 Modified files: devel/p5-SDL : Makefile devel/p5-SDL/patches: patch-Build_PL patch-src_SDL_xs Log message: Unbreak p5-SDL on amd64, i386. Patch src/SDL.xs: it had been calling boot_SDL() without the required arguments, which doesn't work on all platforms. Also remove a dead HOMEPAGE and add c to WANTLIB. ok cwen@ afresh1@
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: gkoeh...@cvs.openbsd.org2020/01/27 14:05:19 Modified files: mail/sylpheed : Makefile Added files: mail/sylpheed/patches: patch-libsylph_ssl_c Log message: Use SNI; fixes TLSv1.3 to imap.gmail.com ok kn@ tb@ sthen@ "go ahead" Amit Kulkarni (maintainer)