On Sun, Aug 05, 2018 at 07:31:55AM +0200, Helmut Grohne wrote: > libtextwrap fails to build from source on amd64 in unstable. A build log > ends with: > > | touch configure-stamp > | dh_testdir > | /usr/bin/make > | make[1]: Entering directory '/<<PKGBUILDDIR>>' > | /usr/bin/make all-am > | make[2]: Entering directory '/<<PKGBUILDDIR>>' > | gcc -DHAVE_CONFIG_H -I. -Wdate-time -D_FORTIFY_SOURCE=2 -g -O2 > -fdebug-prefix-map=/<<PKGBUILDDIR>>=. -fstack-protector-strong -Wformat > -Werror=format-security -MT dotextwrap.o -MD -MP -MF .deps/dotextwrap.Tpo -c > -o dotextwrap.o dotextwrap.c > | mv -f .deps/dotextwrap.Tpo .deps/dotextwrap.Po > | /bin/bash ./libtool --tag=CC --mode=link gcc -g -O2 > -fdebug-prefix-map=/<<PKGBUILDDIR>>=. -fstack-protector-strong -Wformat > -Werror=format-security -Wl,-z,relro -o dotextwrap dotextwrap.o -ltextwrap > | libtool: link: gcc -g -O2 -fdebug-prefix-map=/<<PKGBUILDDIR>>=. > -fstack-protector-strong -Wformat -Werror=format-security -Wl,-z -Wl,relro -o > dotextwrap dotextwrap.o -ltextwrap > | /usr/bin/ld: cannot find -ltextwrap > | collect2: error: ld returned 1 exit status > | make[2]: *** [Makefile:515: dotextwrap] Error 1 > | make[2]: Leaving directory '/<<PKGBUILDDIR>>' > | make[1]: *** [Makefile:373: all] Error 2 > | make[1]: Leaving directory '/<<PKGBUILDDIR>>' > | make: *** [debian/rules:56: build-stamp] Error 2 > | dpkg-buildpackage: error: debian/rules build subprocess returned exit > status 2 > > This looks very strange and more of a breakage of makefile dependencies. > It didn't even try building -ltextwrap and links it already. The issue > is reproducible without parallelism. I haven't seen it until "recently" > though since meson and then pcre3 have fucked up bootstrap qa lately, > "recently" means like two weeks here though reproducible builds didn't > encounter the issue (yet?). The make-dfsg upload looks a bit suspicious, > so I'm Ccing Ben here.
make seems to be a wrong guess here. After downgrading make, it fails the same way. Sorry, Ben, for the noise. Also thank you for having taken care of it. Scanning the recent upload history, automake-1.16 looks suspicious now. I also rescheduled libtextwrap on reproducible.d.n and it fails there as well. Another data point for a fairly recent regression. Helmut
