https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=270421
throwaway_vthg...@protonmail.com changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |throwaway_vthgwq4@protonmai
| |l.com
--- Comment #4 from throwaway_vthg...@protonmail.com ---
I have similar here with -CURRENT (base LLVM 15) and Rust using LLVM 16:
[...]
cargo --locked build --verbose --release --bin rsvg-convert
libtool: link: /usr/bin/nm -B ./.libs/librsvg_c_api.a | /usr/bin/sed -n -e
's/^.*[ ]\([ABCDGIRSTW][ABCDGIRSTW]*\)[ ][ ]*\([_A-Za-
z][_A-Za-z0-9]*\)$/\1 \2 \2/p' | /usr/bin/sed '/ __gnu_lto/d' | /usr/bin/sed
's/.* //' | sort | uniq > .libs/librsvg-2.exp
/usr/bin/nm: error:
./.libs/librsvg_c_api.a(rsvg-6aae7815bd5a5a95.rsvg.f7c38f4e-cgu.0.rcgu.o):
Unknown attribute kind (86) (Producer: 'LLVM16.0.3' Reader: 'LLVM 15.0.7')
[...]
/usr/bin/nm is detected in configure script:
[...]
checking if the linker (/usr/bin/ld) is GNU ld... yes
checking for BSD- or MS-compatible name lister (nm)... /usr/bin/nm -B
checking the name lister (/usr/bin/nm -B) interface... BSD nm
[...]
To avoid Rust LLVM and nm version match, try (adjust 16 for Rust LLVM version
in use):
diff --git a/graphics/librsvg2-rust/Makefile b/graphics/librsvg2-rust/Makefile
--- a/graphics/librsvg2-rust/Makefile
+++ b/graphics/librsvg2-rust/Makefile
@@ -13,7 +13,8 @@ WWW= https://live.gnome.org/LibRsvg
LICENSE= LGPL20
-BUILD_DEPENDS= rst2man:textproc/py-docutils
+BUILD_DEPENDS= rst2man:textproc/py-docutils \
+ llvm-nm16:devel/llvm16
LIB_DEPENDS= libfontconfig.so:x11-fonts/fontconfig \
libfreetype.so:print/freetype2 \
libharfbuzz.so:print/harfbuzz \
@@ -30,7 +31,7 @@ CONFIGURE_ARGS= --disable-Bsymbolic \
--disable-dependency-tracking \
--disable-static \
--enable-vala
-CONFIGURE_ENV= ${CARGO_ENV}
+CONFIGURE_ENV= ${CARGO_ENV} NM=/usr/local/bin/llvm-nm16
MAKE_ENV= ${CARGO_ENV}
CARGO_ENV= RUST_TARGET=${CARGO_BUILD_TARGET}
CARGO_BUILD= no
Alternatively, investigate if parts of librsvg2-rust invoking cargo directly
(rg 'CARGO_RELEASE_ARGS' in WRKDIR) could opt-out either from LTO (maybe
ThinLTO is partly used, and switching to fat LTO would help) or build.
--
You are receiving this mail because:
You are the assignee for the bug.
