Revert the addition of build dependency in commit 2d1694ff7
to a non-existent host build of zlib.
The host build of zlib was removed already in April 2018 by
8dcd941d8b (diff-1ed408c61d79f9c6c5d197333e94ce8d)
which made zlib a build tool defined in /tools
The newly introduced build dependency causes always a warning like:
WARNING: Makefile 'package/feeds/packages/postgresql/Makefile'
has a build dependency on 'zlib/host', which does not exist
Not sure what was the error that 2d1694ff7 tried to fix,
but reference to a non-existent host build is not the solution.
Signed-off-by: Hannu Nyman <hannu.nyman@iki.fi>
spotted on buildbot trying postgresql/host build:
configure: error: zlib library not found
Fix this by adding zlib/host to HOST_BUILD_DEPENDS.
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
This update includes fixes for the following CVEs:
- CVE-2018-1115
- CVE-2018-10925
- CVE-2018-10915
Signed-off-by: Sebastian Kemper <sebastian_ml@gmx.net>
When building for ARC the compile fails with:
/var/lib/buildbot/slaves/dave-builder/arc_arc700/build/sdk/build_dir/target-arc_arc700_uClibc/postgresql-9.6.5/src/include/storage/s_lock.h:899:2: error: #error PostgreSQL does not have native spinlock support on this platform. To continue the compilation, rerun configure using --disable-spinlocks. However, performance will be poor. Please report this to pgsql-bugs@postgresql.org.
#error PostgreSQL does not have native spinlock support on this platform. To continue the compilation, rerun configure using --disable-spinlocks. However, performance will be poor. Please report this to pgsql-bugs@postgresql.org.
So disable spinlocks when compiling for this arch. This was done
likewise for the avr32 target, which is not supported anymore, so this
can be deleted.
Signed-off-by: Sebastian Kemper <sebastian_ml@gmx.net>
- The extra include is removed; the postgresql build system adds this
include by itself.
- '-lpthread' is removed; the postgresql build system is able to detect
how to access pthread features by itself.
- '$(FPIC)' is removed; adding this to CFLAGS is not recommendable:
a) The postgresql build system adds its own PIC flag behind OpenWrt's
CFLAGS when compiling shared objects, overriding OpenWrt's setting.
b) Forcing applications into PIC mode will just slow them down.
c) This leads to build failures on aarch64:
access/transam/commit_ts.o: In function `CommitTsPagePrecedes':
commit_ts.c:(.text+0x14): relocation truncated to fit: R_AARCH64_LD64_GOTPAGE_LO15 against symbol `TransactionIdPrecedes' defined in .text section in access/transam/transam.o
commit_ts.c:(.text+0x14): warning: Too many GOT entries for -fpic, please recompile with -fPIC
/var/lib/buildbot/slaves/dave-builder/aarch64_generic/build/sdk/staging_dir/toolchain-aarch64_generic_gcc-7.3.0_musl/bin/../lib/gcc/aarch64-openwrt-linux-musl/7.3.0/../../../../aarch64-openwrt-linux-musl/bin/ld: final link failed: Symbol needs debug section which does not exist
collect2: error: ld returned 1 exit status
Signed-off-by: Sebastian Kemper <sebastian_ml@gmx.net>
Extend UCI mechanics to allow pre-populating a newly created
database by executing SQL statements from a scripts file.
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
* make sure contrib stuff gets built and installed
* refresh patches and improve Makefile while at it
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
As both LEDE and OpenWrt have STAGING_DIR_HOSTPKG now, we can start to rely
on it. See 73b7f55424 for more information on
STAGING_DIR_HOSTPKG.
STAGING_DIR_HOSTPKG won't actually be changed before the first LEDE release
(it is equivalent to $(STAGING_DIR)/host), so this simple search/replace
cleanup is safe to apply. Doing this cleanup now will be useful for the
Gluon project (an OpenWrt/LEDE based firmware framework) for experimenting
with modifying STAGING_DIR_HOSTPKG before doing this in the LEDE upstream.
Also fixes a typo in the dbus Makefile ("STAGIND_DIR").
Signed-off-by: Matthias Schiffer <mschiffer@universe-factory.net>
Both postgresql-server and postgresql-cli install /usr/bin/psql which
thus creates a conflict when installing both packages.
Fix this by removing duplicate and useless binaries from the
postgresql-server package.
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
* convert package build to use host-build for ecpg, pg_config and zic
* introduce /lib/functions/postgresql.sh to be used by packages
requiring a postgres database to exist as well as postgres' init
* no longer require shadow-su, patch pg_ctl to setuid() ifself instead
* auto-create database directory if there is enough free space
* auto-create databases configured in UCI
* remove some dead uci config options
* grab maintainership
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
Setting the additional include path for host-build of ecpg is no
longer required and started breaking the build (PostgeSQL BUG-14033),
remove it.
Also, pgsql-server no longer requires shadow-su since procd can take
care of starting it running under the appropriate user.
Moving the dependency to pgsql-cli makes sense as the 'su' command is
still needed to boot-strap a new data directory and also commonly used
for other administration tasks. However, in that way the CLI and
shadow-su can be removed during production once pgsql-server has been
setup.
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
$(CP) was defined as `cp -fpR' since the very begining of OpenWrt build
system (2006-06-22). The -R option should be enough and base packages
use only $(CP) for the same purposes just fine and BSD manual of cp also
discourages the use of `-r' option. So let's just tidy up the usage now.
Signed-off-by: Yousong Zhou <yszhou4tech@gmail.com>