Restore host building on request from a user. Apparently, packages
outside of openwrt/packages feed rely on Freetype 2 building on host.
While at it, remove sed-editing of freetype-config that is no longer
necessary. The current version of freetype-config correctly identifies
prefix, exec_prefix, includedir and libdir.
Signed-off-by: Val Kulkov <val.kulkov@gmail.com>
Update to the latest stable release.
Fix check for file descriptor passing that causes libmilter's runtime
error when libmilter/OpenDKIM are used with Postfix:
"unix_send_fd: your system has no support for file descriptor passing"
The issue with the file descriptor passing has been fixed upstream in
20170618 update to Postfix v3.3 experimental release. The patch must
therefore be removed before compiling Postfix v3.3.*
Signed-off-by: Val Kulkov <val.kulkov@gmail.com>
Add Freetype 2 support to php7-mod-gd. Introduce a configuration
parameter to disable Freetype 2 support if the increased package
size is a concern.
Signed-off-by: Val Kulkov <val.kulkov@gmail.com>
On buildbots, the building directory seems to be not a subdir of
a git repo, so a variable expands into an empty string which in turn
results in a syntax error:
--snip--
-- Looking for in6addr_any - found
-- Found Git: /data/bowl-builder/arm_cortex-a5/build/sdk/staging_dir/host/bin/git (found version "2.1.4")
fatal: Not a git repository (or any parent up to mount point /data)
Stopping at filesystem boundary (GIT_DISCOVERY_ACROSS_FILESYSTEM not set).
CMake Error at CMakeLists.txt:114 (if):
if given arguments:
"STREQUAL" "/data/bowl-builder/arm_cortex-a5/build/sdk/build_dir/target-arm_cortex-a5_musl_eabi/libiio-0.11"
Unknown arguments specified
-- Configuring incomplete, errors occurred!
--snap--
The patch quotes the if arguments and thus fixes the build.
Signed-off-by: Michael Heimpold <mhei@heimpold.de>
The release tar does not contain the srcdir/guess-rev.sh.
If it's not existant the configure scripts uses different
variables which removes the build timestamp.
Fixes reproducible builds.
Signed-off-by: Alexander Couzens <lynxis@fe80.eu>
Previously, configuration of emailrelay was done by fiddling around
with the commandline inside the initscript. Introduce a config file
in /etc/config for basic configuration and at the same time switch
to a procd-style initscript.
Signed-off-by: Paul Wassi <p.wassi@gmx.at>
This is just a cosmetic thing, therefore no increase of PKG_RELEASE.
- Trailing spaces and tabs
- Add newline to end of file
- Add License info to Makefile
- Move maintainer definition to PKG_MAINTAINER
Signed-off-by: Paul Wassi <p.wassi@gmx.at>
If the date is changed by ntp the age value of mwan3 on ubus could jitter.
Use instead the uptime value from /proc/uptime which will not change during
system run.
Signed-off-by: Florian Eckert <fe@dev.tdt.de>
This part of the Makefile was commented out during update from
PHP 5.x to 7.x and not re-enabled in the meanswhile, so fix this finally.
Reported-by: Val Kulkov <val.kulkov@gmail.com>
Signed-off-by: Michael Heimpold <mhei@heimpold.de>
During 4.2.1 version update support for subordinate IDs has been
disabled. It was handled by:
1) Adding --disable-subordinate-ids to avoid:
configure: error: cannot run test program while cross compiling
2) Adding patch 003-fix-disabling-subids.patch to avoid:
usermod.c: In function 'process_flags':
usermod.c:1364:10: error: 'vflg' undeclared (first use in this function)
if ( (vflg || Vflg)
^
This commit adds a patch with a proper configure.in fix. We don't need
to disable subordinate IDs anymore.
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
Fix uid/gid conflict with package icecast (uid=87, gid=87) and
package pianod (uid=88, gid=88) by claiming uid/gid 25 for postfix
instead of 87 and uid/gid 26 for postconf instead of 88.
Remove main.cf.default that is not used by any Postfix module.
The default configuration parameters are hardcoded into postconf
and they are available via "postconf -d <parameter>".
main.cf.default is therefore unnecessary.
"mydomain" parameter should be deduced from the domain configuration
setting in /e/c/dhcp rather than from "hostname" in /e/c/system
which is often just a hostname without the domain name component.
"mynetworks" parameter is unnecessary where "mynetworks_style" is
defined.
The change of the location of data_directory, queue_directory and
mail_spool_directory from /var/lib/postfix, /var/spool/postfix and
/var/mail to /data/postfix/lib, /data/postfix/spool and
/data/postfix/mail has been postponed until the author of this
commit and the maintainer of this package has had a meaningful
opportunity to consult with LEDE/OpenWrt developers on the preferred
mount point for the persistent data storage device.
Change the criteria for executing the post-installation script.
Invoke postinst() where the value of "myhostname" parameter in main.cf
is not defined rather than where "$config_directory"/opkg_postinst is
present with the value of "$config_directory" hardcoded to
"$IPKG_INSTROOT"/etc/postfix in /etc/init.d/postfix.
Enable PCRE support by default.
Remove the code making entries in /etc/sysupgrade.conf and instead
define configuration files to be saved during a sysupgrade in
Makefile's Package/postfix/conffiles section.
Move the creation of mail_spool_directory from host to target in
order to enable change of directory ownership to 'postfix' when
postinit() is executed on the target.
Avoid postfix restart in postinst script that results in an
error message during opkg install process.
Enable Postfix autostart by creating a link in /etc/rc.d/ on
installation.
Make the code less verbose.
Signed-off-by: Val Kulkov <val.kulkov@gmail.com>
This commit updates Freetype 2 to the latest upstream version and fixes
the issue where freetype-config was incorrectly installed into the
host, rather than the target, staging directory, therefore causing
compilation issues with other packages that depend on libfreetype.
Signed-off-by: Val Kulkov <val.kulkov@gmail.com>
This should hopefully fix the Python3 build on buildbot.
For a while I assumed it may be a build-bot issue, but
then looking through the packages repo [and finding
the bluez package] it looks like, if you try
to build all packages, Python3 detects the bluetooth
headers installed by bluez.
It looks like Python's bluetooth support was somewhat
broken ; it was not detecting the <bluetooth/bluetooth.h>
header, so a backport from Python3 to Python fixed that.
Signed-off-by: Alexandru Ardelean <ardeleanalex@gmail.com>
Report SHA256 checksums in addition to the MD5 ones to make cgi-io suitable
for sysupgrade image verification.
Also allow stat(), md5sum and/or sha256sum to fail and respond with a JSON
null value instead, leaving it to the frontend to handle errors as needed.
Fixes#4790.
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
Val Kulkov
Ted Hess
Nikos Mavrogiannopoulos
Vladimir Berezhnoy
Rosen Penev
Michael Heimpold
champtar
Mirko Vogt
mike mcginty
Alexander Couzens
Paul Wassi
John Crispin
Hirokazu MORIKAWA
Florian Eckert
Rafał Miłecki
Hannu Nyman
Alexandru Ardelean
Lucian Cristian
Jo-Philipp Wich
W. van den Akker