In D54529#1246403, @bapt wrote:I have readded the macos CI and fix the code that needs fixing, I still consider we should use the installed libcurl when building on macos so I haven't fixed the libcurl part.
- Feed Queries
- All Stories
- Search
- Feed Search
- Transactions
- Transaction Logs
Feed Search
Today
Today
Thu, Jan 22
Thu, Jan 22
Tue, Jan 20
Tue, Jan 20
Mon, Jan 19
Mon, Jan 19
Wed, Jan 14
Wed, Jan 14
Mon, Jan 5
Mon, Jan 5
Sat, Jan 3
Sat, Jan 3
Fri, Jan 2
Fri, Jan 2
Sat, Dec 27
Sat, Dec 27
Dec 3 2025
Dec 3 2025
wosch retitled D52330: clang 19/20 needs the real linker name from clang 20 needs the real linker name
Dec 2 2025
Dec 2 2025
I was wondering why `make src.txz' is so slow now, from 2.5min to 9min.
Dec 1 2025
Dec 1 2025
Nov 30 2025
Nov 30 2025
Nov 24 2025
Nov 24 2025
Nov 11 2025
Nov 11 2025
wosch updated the diff for D51443: release/Makefile should not use the host /usr/bin/install command, but the freshly built one.
Shorter line and use the :H modifier instead ../
wosch added a comment to D51443: release/Makefile should not use the host /usr/bin/install command, but the freshly built one.
In D51443#1178659, @emaste wrote:Why would we not do this unconditionally?
Nov 10 2025
Nov 10 2025
wosch updated the diff for D51087: cross-build-macos.sh - build FreeBSD/arm64 from source on MacOS for MacOS 14.x and later.
- support multiple LLVM versions (clang19 and clang20)
- supports Apple M1 arm64 and Apple/Intel
- supports concurrent cross building (e.g. current and stable at the same time with same MAKEOBJDIRPREFIX)
- supported make targets: buildworld; buildkernel; packages; -C./release ftp
- by default it runs fast (no toolchain etc.) for buildworld + buildkernel
- add a help message: ./cross-build-macos.sh help
- keep up to 32 logfiles
- add a debug mode (debug=1)
- keep track of good/bad git commits in .cross-build-macos-git-*
- supported branches: main, stable/15 (buildworld + buildkernel)
- tested on macOS 14.x, 15.x, 26.x
wosch added a comment to D51087: cross-build-macos.sh - build FreeBSD/arm64 from source on MacOS for MacOS 14.x and later.
In D51087#1166093, @olivier wrote:What the advantage of this tools vs the existing tools/build/make.py ?
Nov 9 2025
Nov 9 2025
Nov 8 2025
Nov 8 2025
Oct 25 2025
Oct 25 2025
Oct 2 2025
Oct 2 2025
Sep 20 2025
Sep 20 2025
Sep 16 2025
Sep 16 2025
Sep 13 2025
Sep 13 2025
Sep 6 2025
Sep 6 2025
Sep 3 2025
Sep 3 2025
Ok, it is very easy to reproduce the issue on FreeBSD itself:
fix make keyword typo (.elseif <-> .elif)
In D52330#1195007, @jrtc27 wrote:And what happens if you ls /opt/homebrew/opt/lld@20/bin?
Sep 2 2025
Sep 2 2025
In D52330#1194788, @jrtc27 wrote:This looks odd. How were you trying to build? Have you seen https://github.com/freebsd/freebsd-src/commit/81fc74fd26da2f75c3010595db7a706c3e28d95b and its parent?
wosch added reviewers for D52330: clang 19/20 needs the real linker name: arichardson, emaste, brooks, jrtc27.
Aug 10 2025
Aug 10 2025
Jul 26 2025
Jul 26 2025
wosch added a comment to D51275: Makefile.inc1: in _pkgbootstrap, use freshly built pkg, not host version.
In D51275#1176964, @ivy wrote:i may be missing something but pkg doesn't ever exist in WSTAGEDIR, does it? it's not part of world since it comes from ports.
edit: or is this meant to call the pkg-bootstrap binary, but that will never work on non-FreeBSD platforms...
Jul 25 2025
Jul 25 2025
Jul 23 2025
Jul 23 2025
Jul 22 2025
Jul 22 2025
Jul 21 2025
Jul 21 2025
add forgotten check for .MAKE.JOBS
Jul 20 2025
Jul 20 2025
Fix possible symlink problem for ports.txz as well.
Jul 18 2025
Jul 18 2025
Increase .so include limit to 32, and display a warning if we reach the limit in debug mode.
In D51412#1173928, @imp wrote:10 might be too small. But if so, it's easy to bump to 20
Simplify while loop & fix an indent issue.
Jul 16 2025
Jul 16 2025
In D51335#1172765, @emaste wrote:I think ports.txz has the same issue and should get the same fix. (Aside, please upload diffs with full context (e.g. with -U999999))
Jul 15 2025
Jul 15 2025
Jul 2 2025
Jul 2 2025
wosch added a comment to D51119: Revert "src/Makefile: log real/user/system time for build targets".
For the proposed workaround "write time logging to stderr instead of stdout" see