multi-arch cross-build toolchain has some issues still, but kyotocabinet now supports cross-building https://wiki.linaro.org/Platform/DevPlatform/CrossCompile/UsingMultiArch particularly: export DEB_LDFLAGS_APPEND="-L/usr/lib/arm-linux-gnueabi -L/lib/arm-linux-gnueabi -L/usr/lib -Wl,-rpath-link=/usr/lib/arm-linux-gnueabi:/lib/arm-linux-gnueabi:/usr/lib
This reverts commit eb9eb07. I guess this should really be set elsewhere. The package already obeys S4.9.1 in supporting the "nocheck" option, so tools should set it.
… script these can be very annoying when testing as dpkg-source will abort the compile when it sees these, even though they are harmless
no need to duplicate it.
…ition if ABI changes, will only then require binNMU
also, moving towards support dpkg-buildpackage cross-compiling
I'd like to get this to work with cross-compiling in the long term.
using xz -2 for debs, keep default gzip for source packages (set in debian/source/options#compression) results (kB): gzip -9 (default): 455 kyotocabinet-doc_1.2.76-4_all.deb 1896 kyotocabinet-utils_1.2.76-4_amd64.deb 354 libkyotocabinet16_1.2.76-4_amd64.deb 44 libkyotocabinet-dbg_1.2.76-4_amd64.deb 570 libkyotocabinet-dev_1.2.76-4_amd64.deb xz -6 (8MB dict) 209 kyotocabinet-doc_1.2.76-4_all.deb 872 kyotocabinet-utils_1.2.76-4_amd64.deb 273 libkyotocabinet16_1.2.76-4_amd64.deb 34 libkyotocabinet-dbg_1.2.76-4_amd64.deb 359 libkyotocabinet-dev_1.2.76-4_amd64.deb xz -4 (4MB dict) 231 kyotocabinet-doc_1.2.76-4_all.deb 903 kyotocabinet-utils_1.2.76-4_amd64.deb 277 libkyotocabinet16_1.2.76-4_amd64.deb 37 libkyotocabinet-dbg_1.2.76-4_amd64.deb 391 libkyotocabinet-dev_1.2.76-4_amd64.deb xz -2 (2MB dict) 236 kyotocabinet-doc_1.2.76-4_all.deb 968 kyotocabinet-utils_1.2.76-4_amd64.deb 289 libkyotocabinet16_1.2.76-4_amd64.deb 38 libkyotocabinet-dbg_1.2.76-4_amd64.deb 401 libkyotocabinet-dev_1.2.76-4_amd64.deb
switch from quilt to using gbp-pq, worried that while maybe a bit more convenient, will produce noise in the gitlog cause it doesn't generate the quilt patches at build time like I would like more
specifically on the buildds I cannot reproduce the problems the buildds are having locally
This reverts commit 9c4ea76. just doesn't seem neccicary
…r some reason
fix up configure test to handle lack of 8 byte atomics correctly as is the case with all ARM targets currently