Tried to merge sys-apps/sdparm into my uclibc root, but the dependencies changed since the last time I did so. sdparm depends on sg3_utils, which depend on rescan-scsi-bus, which require 'usleep', which is provided by app-admin/killproc. It looks like when compiling killproc under uclibc, the 'usleep' binary builds successfully, but the overall build fails because startproc.c includes the argz.h header, which isn't available under uclibc. It's only provided by glibc. I'm not sure of a good workaround for this w/o diving into the code, but I did discover that atftp copied glibc's argz.h definition for itself so that it can compile under BSD, uclibc, and other environments: https://github.com/seveas/atftp/blob/master/argz.h Not sure if the same approach is valid here (doubtful). I can't find anything solid on Google that suggests a uclibc-alternative to argz.h, either. Apparently, it's something that needs porting: http://infobot.rikers.org/%23uclibc/20050815.html.gz 21:24.48 waiting4_0928 Hi, how do I get argz.h to build gettext 21:26.02 solar you really do not want gettext. But I would look at the gentoo or buildroot patches/build systems for that pkg 21:26.32 waiting4_0928 solar, thx 21:36.13 andersee_sleep bummer 21:36.44 andersee last night's buildroot runs failed 21:42.23 nitinkg solar, whats the story behind missing argz.h? 21:50.14 mjn3-work what story? we simply never included (glibc) argz support 21:51.59 nitinkg oh ok, is this something work-in-progress? 21:52.48 mjn3-work should be easy enough to port. are you volunteering? 21:54.25 nitinkg mjn3-work, I will take a look. I have to get a patch ready for gettext first 21:56.07 mjn3-work what are you trying to do with gettext that requires a patch? 21:56.31 nitinkg cleaner inclusion of argz.h Build errors: >>> Emerging (1 of 4) app-admin/killproc-2.13-r1::gentoo * killproc-2.13.tar.gz SHA256 SHA512 WHIRLPOOL size ;-) ... [ ok ] >>> Unpacking source... >>> Unpacking killproc-2.13.tar.gz to /var/tmp/portage/app-admin/killproc-2.13-r1/work >>> Source unpacked in /var/tmp/portage/app-admin/killproc-2.13-r1/work >>> Preparing source in /var/tmp/portage/app-admin/killproc-2.13-r1/work/killproc-2.13 ... * Applying killproc-2.13-makefile.patch ... [ ok ] >>> Source prepared. >>> Configuring source in /var/tmp/portage/app-admin/killproc-2.13-r1/work/killproc-2.13 ... >>> Source configured. >>> Compiling source in /var/tmp/portage/app-admin/killproc-2.13-r1/work/killproc-2.13 ... make -j3 mips-unknown-linux-uclibc-gcc -Os -pipe -march=mips2 -mtune=mips2 -mplt -D_GNU_SOURCE -Wall -DINITDIR=\"/etc/init.d\" -c libinit.c mips-unknown-linux-uclibc-gcc -Os -pipe -march=mips2 -mtune=mips2 -mplt -D_GNU_SOURCE -Wall -Wl,-O1 -Wl,--as-needed -o usleep usleep.c mips-unknown-linux-uclibc-gcc -Os -pipe -march=mips2 -mtune=mips2 -mplt -D_GNU_SOURCE -Wall -Wl,-O1 -Wl,--as-needed -o fsync fsync.c libinit.c: In function 'check4nfs': libinit.c:1556:15: warning: variable 'l' set but not used [-Wunused-but-set-variable] NFS *p, *n, *l; ^ mips-unknown-linux-uclibc-gcc -Os -pipe -march=mips2 -mtune=mips2 -mplt -D_GNU_SOURCE -Wall -Wl,-O1 -Wl,--as-needed -o killproc killproc.c libinit.o mips-unknown-linux-uclibc-gcc -Os -pipe -march=mips2 -mtune=mips2 -mplt -D_GNU_SOURCE -Wall -Wl,-O1 -Wl,--as-needed -o startproc startproc.c libinit.o mips-unknown-linux-uclibc-gcc -Os -pipe -march=mips2 -mtune=mips2 -mplt -D_GNU_SOURCE -Wall -Wl,-O1 -Wl,--as-needed -o checkproc checkproc.c libinit.o startproc.c:28:18: fatal error: argz.h: No such file or directory compilation terminated. Makefile:66: recipe for target 'startproc' failed make: *** [startproc] Error 1 make: *** Waiting for unfinished jobs....
We don't use argz.h there at all, so that was an easy fix. Fixed without revision bump.