Re: dhcpcd5: FTBFS on Hurd-i386
On Fri, Jul 29, 2022 at 11:00 AM Samuel Thibault wrote: > > Martin-Éric Racine, le ven. 29 juil. 2022 10:35:38 +0300, a ecrit: > > "Hurd is not a supported OS. At a minimum it would need an equivalent > > of if-bsd.c being ported." > > > > Does that ring any bell? > > The Hurd has a very BSD-ish interface, so starting from if-bsd.c is > probably a good idea. See https://github.com/NetworkConfiguration/dhcpcd/issues/109 Feel free to comment and send patches. Martin-Éric
Re: dhcpcd5: FTBFS on Hurd-i386
Martin-Éric Racine, le ven. 29 juil. 2022 10:35:38 +0300, a ecrit: > "Hurd is not a supported OS. At a minimum it would need an equivalent > of if-bsd.c being ported." > > Does that ring any bell? The Hurd has a very BSD-ish interface, so starting from if-bsd.c is probably a good idea. Samuel
Re: dhcpcd5: FTBFS on Hurd-i386
On Fri, Jul 29, 2022 at 9:27 AM Jeffrey Walton wrote: > > On Fri, Jul 29, 2022 at 2:09 AM Martin-Éric Racine > wrote: > > > > On Thu, Jul 28, 2022 at 10:38 PM Samuel Thibault > > wrote: > > > > > > Martin-Éric Racine, le jeu. 28 juil. 2022 21:30:42 +0300, a ecrit: > > > > Actually, where it barfs is here: > > > > > > > > #if !defined(__linux__) && !defined(__QNX__) && !defined(__sun) > > > > #include > > > > #else > > > > #include "endian.h" > > > > #endif > > > > > > > > What is the equivalent define for Hurd? > > > > > > It's just the same on the Hurd as on Linux, since that's just coming > > > from glibc. > > > > Yet it FTBFS on Hurd at that precise point. That suggets we need to > > patch that spot. Not knowing Hurd, I have no idea what the correct > > fragment would be. > > If you want the Mach kernel, use __MACH__. If you want GNU Hurd, use > __GNU__ or __gnu_hurd__. > > The Pre-defined Operating System Macros list at Sourceforge is usually > reliable. Also see > https://sourceforge.net/p/predef/wiki/OperatingSystems/ . Upstream says: "Hurd is not a supported OS. At a minimum it would need an equivalent of if-bsd.c being ported." Does that ring any bell? Martin-Éric
Re: dhcpcd5: FTBFS on Hurd-i386
On Fri, Jul 29, 2022 at 2:09 AM Martin-Éric Racine wrote: > > On Thu, Jul 28, 2022 at 10:38 PM Samuel Thibault wrote: > > > > Martin-Éric Racine, le jeu. 28 juil. 2022 21:30:42 +0300, a ecrit: > > > Actually, where it barfs is here: > > > > > > #if !defined(__linux__) && !defined(__QNX__) && !defined(__sun) > > > #include > > > #else > > > #include "endian.h" > > > #endif > > > > > > What is the equivalent define for Hurd? > > > > It's just the same on the Hurd as on Linux, since that's just coming > > from glibc. > > Yet it FTBFS on Hurd at that precise point. That suggets we need to > patch that spot. Not knowing Hurd, I have no idea what the correct > fragment would be. If you want the Mach kernel, use __MACH__. If you want GNU Hurd, use __GNU__ or __gnu_hurd__. The Pre-defined Operating System Macros list at Sourceforge is usually reliable. Also see https://sourceforge.net/p/predef/wiki/OperatingSystems/ . Jeff
Re: dhcpcd5: FTBFS on Hurd-i386
On Thu, Jul 28, 2022 at 10:38 PM Samuel Thibault wrote: > > Martin-Éric Racine, le jeu. 28 juil. 2022 21:30:42 +0300, a ecrit: > > Actually, where it barfs is here: > > > > #if !defined(__linux__) && !defined(__QNX__) && !defined(__sun) > > #include > > #else > > #include "endian.h" > > #endif > > > > What is the equivalent define for Hurd? > > It's just the same on the Hurd as on Linux, since that's just coming > from glibc. Yet it FTBFS on Hurd at that precise point. That suggets we need to patch that spot. Not knowing Hurd, I have no idea what the correct fragment would be. Martin-Éric
Re: dhcpcd5: FTBFS on Hurd-i386
Martin-Éric Racine, le jeu. 28 juil. 2022 21:30:42 +0300, a ecrit: > Actually, where it barfs is here: > > #if !defined(__linux__) && !defined(__QNX__) && !defined(__sun) > #include > #else > #include "endian.h" > #endif > > What is the equivalent define for Hurd? It's just the same on the Hurd as on Linux, since that's just coming from glibc. Samuel
Re: dhcpcd5: FTBFS on Hurd-i386
On Fri, Jul 8, 2022 at 3:11 PM Martin-Éric Racine wrote: > > (non-subscriber; please keep me in CC) > > On Fri, Jul 1, 2022 at 6:31 PM Martin-Éric Racine > wrote: > > > > > ../compat/rbtree.h:50:10: fatal error: sys/endian.h: No such file or > > > directory > > > 50 | #include > > > Probably can be used instead? > > > > Would you have a patch I could test on Debian and later submit to upstream? > > This seems to be the relevant ./configure fragment: > > * > if [ -z "$BE64ENC" ]; then > printf "Testing for be64enc ... " > cat <_be64enc.c > #include > #include > int main(void) { > be64enc(NULL, 0); > return 0; > } > EOF > if $XCC _be64enc.c -o _be64enc 2>&3; then > BE64ENC=yes > else > BE64ENC=no > fi > echo "$BE64ENC" > rm -f _be64enc.c _be64enc > fi > if [ "$BE64ENC" = no ]; then > echo "#include \"compat/endian.h\"" >>$CONFIG_H > fi > * Actually, where it barfs is here: #if !defined(__linux__) && !defined(__QNX__) && !defined(__sun) #include #else #include "endian.h" #endif What is the equivalent define for Hurd? >From what I could google, __GNU__ or __MACH__ might be possible choices, but I'm not sure which one applies here. Martin-Éric
Re: dhcpcd5: FTBFS on Hurd-i386
(non-subscriber; please keep me in CC) On Fri, Jul 1, 2022 at 6:31 PM Martin-Éric Racine wrote: > > > ../compat/rbtree.h:50:10: fatal error: sys/endian.h: No such file or > > directory > > 50 | #include > > Probably can be used instead? > > Would you have a patch I could test on Debian and later submit to upstream? This seems to be the relevant ./configure fragment: * if [ -z "$BE64ENC" ]; then printf "Testing for be64enc ... " cat <_be64enc.c #include #include int main(void) { be64enc(NULL, 0); return 0; } EOF if $XCC _be64enc.c -o _be64enc 2>&3; then BE64ENC=yes else BE64ENC=no fi echo "$BE64ENC" rm -f _be64enc.c _be64enc fi if [ "$BE64ENC" = no ]; then echo "#include \"compat/endian.h\"" >>$CONFIG_H fi * Martin-Éric
Re: dhcpcd5: FTBFS on Hurd-i386
> ../compat/rbtree.h:50:10: fatal error: sys/endian.h: No such file or directory > 50 | #include > Probably can be used instead? Would you have a patch I could test on Debian and later submit to upstream? Martin-Éric
Re: dhcpcd5: FTBFS on Hurd-i386
Hello, Martin-Éric Racine, le ven. 01 juil. 2022 12:01:24 +0300, a ecrit: > ../compat/rbtree.h:50:10: fatal error: sys/endian.h: No such file or directory >50 | #include > | ^~ Probably can be used instead? Samuel
dhcpcd5: FTBFS on Hurd-i386
Greetings, dhcpcd5 FTBFS on Hurd-i386: Makefile:55: warning: ignoring prerequisites on suffix rule definition cc -g -O2 -ffile-prefix-map=/<>=. -fstack-protector-strong -Wformat -Werror=format-security -std=c99 -Wdate-time -D_FORTIFY_SOURCE=2 -I/include -DHAVE_CONFIG_H -DNDEBUG -DINET -DARP -DARPING -DIPV4LL -DINET6 -DDHCP6 -DAUTH -DPRIVSEP -I.. -I../src -I./crypt -c common.c -o common.o In file included from ../config.h:24, from dhcpcd.h:37, from common.c:40: ../compat/rbtree.h:50:10: fatal error: sys/endian.h: No such file or directory 50 | #include | ^~ compilation terminated. make[2]: *** [Makefile:55: common.o] Error 1 make[2]: Leaving directory '/<>/src' make[1]: *** [Makefile:24: all] Error 2 make[1]: Leaving directory '/<>' dh_auto_build: error: make -j1 returned exit code 2 make: *** [debian/rules:6: binary-arch] Error 255 Would anyone have a fix for this? Thanks! Martin-Éric