amd64 live building 14.1p2 found this happening in the past, with the same errors I found, don't think this was an issue in 3.101 at least didn't seem to come up for me outside of maybe that october 2023 commit in the link below. temporarily removing mpich and nss updated fine. not sure how to resolve. https://cgit.freebsd.org/ports/commit/?id=7fbe5fd00a138a8198b5093a5cde96b938db48ef gmake[5]: Leaving directory '/usr/ports/security/nss/work/nss-3.102/nss/lib/libpkix/pkix/util' cc -B/usr/local/bin -o FreeBSD14.1_OPT.OBJ/pkix_pl_socket.o -c -std=c99 -O2 -pipe -march=westmere -I/usr/local/include/nspr -Wno-error=unused-function -fstack-protector-strong -fno-strict-aliasing -fPIC -Wall -Wno-switch -DFREEBSD -DHAVE_STRERROR -DHAVE_BSD_FLOCK -Wall -Wshadow -Qunused-arguments -Wno-parentheses-equality -Wno-array-bounds -Wno-unevaluated-expression -DNSS_NO_GCC48 -DXP_UNIX -UDEBUG -DNDEBUG -D_THREAD_SAFE -D_REENTRANT -DNSS_DISABLE_SSE3 -DNSS_NO_INIT_SUPPORT -DUSE_UTIL_DIRECTLY -DNO_NSPR_10_SUPPORT -DSSL_DISABLE_DEPRECATED_CIPHER_SUITE_NAMES -I/usr/local/include -I./../dist/FreeBSD14.1_OPT.OBJ/include -I./../dist/public/ -I./../dist/private/ -fPIC -Wall -Wno-switch -DFREEBSD -DHAVE_STRERROR -DHAVE_BSD_FLOCK -Wall -Wshadow -Qunused-arguments -Wno-parentheses-equality -Wno-array-bounds -Wno-unevaluated-expression -DNSS_NO_GCC48 -DXP_UNIX -UDEBUG -DNDEBUG -D_THREAD_SAFE -D_REENTRANT -DNSS_DISABLE_SSE3 -DNSS_NO_INIT_SUPPORT -DUSE_UTIL_DIRECTLY -DNO_NSPR_10_SUPPORT -DSSL_DISABLE_DEPRECATED_CIPHER_SUITE_NAMES -I/usr/local/include -I./../dist/FreeBSD14.1_OPT.OBJ/include -I./../dist/public/ -I./../dist/private/ -fPIC -Wall -Wno-switch -DFREEBSD -DHAVE_STRERROR -DHAVE_BSD_FLOCK -Wall -Wshadow -Qunused-arguments -Wno-parentheses-equality -Wno-array-bounds -Wno-unevaluated-expression -DNSS_NO_GCC48 -DXP_UNIX -UDEBUG -DNDEBUG -D_THREAD_SAFE -D_REENTRANT -DNSS_DISABLE_SSE3 -DNSS_NO_INIT_SUPPORT -DUSE_UTIL_DIRECTLY -DNO_NSPR_10_SUPPORT -DSSL_DISABLE_DEPRECATED_CIPHER_SUITE_NAMES -I/usr/local/include -I../../dist/FreeBSD14.1_OPT.OBJ/include -I../../dist/public/ -I../../dist/private/ -fPIC -Wall -Wno-switch -DFREEBSD -DHAVE_STRERROR -DHAVE_BSD_FLOCK -Wall -Wshadow -Qunused-arguments -Wno-parentheses-equality -Wno-array-bounds -Wno-unevaluated-expression -DNSS_NO_GCC48 -DXP_UNIX -UDEBUG -DNDEBUG -D_THREAD_SAFE -D_REENTRANT -DNSS_DISABLE_SSE3 -DNSS_NO_INIT_SUPPORT -DUSE_UTIL_DIRECTLY -DNO_NSPR_10_SUPPORT -DSSL_DISABLE_DEPRECATED_CIPHER_SUITE_NAMES -I/usr/local/include -I../../../dist/FreeBSD14.1_OPT.OBJ/include -I../../../dist/public/ -I../../../dist/private/ -fPIC -Wall -Wno-switch -DFREEBSD -DHAVE_STRERROR -DHAVE_BSD_FLOCK -Wall -Wshadow -Qunused-arguments -Wno-parentheses-equality -Wno-array-bounds -Wno-unevaluated-expression -DNSS_NO_GCC48 -DXP_UNIX -UDEBUG -DNDEBUG -D_THREAD_SAFE -D_REENTRANT -DNSS_DISABLE_SSE3 -DNSS_NO_INIT_SUPPORT -DUSE_UTIL_DIRECTLY -DNO_NSPR_10_SUPPORT -DSSL_DISABLE_DEPRECATED_CIPHER_SUITE_NAMES -I/usr/local/include -I../../../../dist/FreeBSD14.1_OPT.OBJ/include -I../../../../dist/public/ -I../../../../dist/private/ -fPIC -Wall -Wno-switch -DFREEBSD -DHAVE_STRERROR -DHAVE_BSD_FLOCK -Wall -Wshadow -Qunused-arguments -Wno-parentheses-equality -Wno-array-bounds -Wno-unevaluated-expression -DNSS_NO_GCC48 -DXP_UNIX -DSHLIB_SUFFIX=\"so\" -DSHLIB_PREFIX=\"lib\" -DSHLIB_VERSION=\"\" -UDEBUG -DNDEBUG -D_THREAD_SAFE -D_REENTRANT -DNSS_DISABLE_SSE3 -DNSS_NO_INIT_SUPPORT -DUSE_UTIL_DIRECTLY -DNO_NSPR_10_SUPPORT -DSSL_DISABLE_DEPRECATED_CIPHER_SUITE_NAMES -I/usr/local/include -I../../../../../dist/FreeBSD14.1_OPT.OBJ/include -I../../../../../dist/public/nss -I../../../../../dist/private/nss pkix_pl_socket.c In file included from sha_fast.c:11: In file included from ./sha_fast.h:9: ./blapii.h:84:26: error: unknown type name 'mp_int'; did you mean 'u_int'? 84 | SECStatus generate_prime(mp_int *prime, int primeLen); | ^~~~~~ | u_int /usr/include/sys/types.h:53:22: note: 'u_int' declared here 53 | typedef unsigned int u_int; | ^ 1 error generated. gmake[4]: *** [../../coreconf/rules.mk:292: FreeBSD14.1_OPT.OBJ/FreeBSD_SINGLE_SHLIB/sha_fast.o] Error 1 gmake[4]: Leaving directory '/usr/ports/security/nss/work/nss-3.102/nss/lib/freebl' gmake[3]: *** [Makefile:712: freebl_FREEBL_BUILD_SINGLE_SHLIB] Error 2 gmake[3]: Leaving directory '/usr/ports/security/nss/work/nss-3.102/nss/lib/freebl' gmake[2]: *** [../coreconf/rules.mk:44: freebl] Error 2
Created attachment 252652 [details] Patch to force mpi.h from nss to be found first Can you check if this fixes your problem?
minor white space errors with the patch when applied /usr/ports]$ sudo git apply ~userman/Desktop/Patches/nss/0001-security-nss-Fix-build-failure-when-mpich-is-install.patch Password: /home/userman/Desktop/Patches/nss/0001-security-nss-Fix-build-failure-when-mpich-is-install.patch:46: trailing whitespace. /home/userman/Desktop/Patches/nss/0001-security-nss-Fix-build-failure-when-mpich-is-install.patch:47: trailing whitespace. but it builds and stages, no orphans, so I think this does resolve it. didn't install, but no reason to think this won't be fine. /usr/ports/security/nss]$ sudo make check-orphans ====> Checking for pkg-plist issues (check-plist) ===> Parsing plist ===> Checking for items in STAGEDIR missing from pkg-plist ===> Checking for items in pkg-plist which are not in STAGEDIR ===> No pkg-plist issues found (check-plist) thanks!
Hi I've tried the proposed patch on stable/14-n268482-46c4fc50d301-dirty with mpich-4.2.2 installed. The patch works for me too, for the following version. PORTNAME= nss PORTVERSION= 3.103 Thank you
Created attachment 253653 [details] yet-another way to build successfully Hi 'security/nss: update to 3.104' still has the same build-time issue when mpich is installed. The same workaround still work well. In addition, the '-quote' trick in INCLUDE directive can placed in the top makefile of this port, as I attached with this comment. Please check it out. Cheers
Comment on attachment 253653 [details] yet-another way to build successfully +1. This patch works well - ensures that the bundled mpi (mpi = Multi-precision Integer) header in the work/ tree is intentionally used rather than unintentionally using installed mpich header (mpi = Message-Passing Interface).
Created attachment 254543 [details] ports tree applyable variant of "another way" patch the first patch no longer seems to work, it applies but breaks other patches in ./files from applying so build completely breaks before configure. the alternate patch method posted works, but wasn't prefixed against ports tree, so just cleaned that up and posted this to directly apply. I hope this gets fixed, it tripped me up last night, and broke firefox so had to do the manually method of removing mpich temporarily so i could look into it again this morning. tweaking the ports Makefile rather then patching source files seems to be a more direct and maintainable way to fix this hopefully permanently. I hope this makes it into the tree.
The latest patch attached in this bug was applied in ports ba18dc61ff94a5af7381080dc58a5a37d17d1da9 But the commit did not acknowledge this bug 280580, so no entry referencing that commit was automatically attached to this bug. Since the patch in this bug has been applied, please close marking that the patch was applied. Committer cc'd. Thanks.
Thanks. I forgot to link this PR. Committed in ports ba18dc61ff94a5af7381080dc58a5a37d17d1da9.