Java support is no longer feasible in openwrt. It's outdated, and can't
be usd without classpath, which has been removed.
Signed-off-by: Eneas U de Queiroz <cotequeiroz@gmail.com>
gperf is a basic tool used when building packages. It's not available on
all hosts, in which case having a host package like this one is
required. For instance the CircleCI environment does not contain gperf,
in which case building of software requiring gperf fails.
This commit reverts the gperf removal.
Signed-off-by: Sebastian Kemper <sebastian_ml@gmx.net>
This changes the package to download from PyPI (using pypi.mk) and
removes the incorrect PKG_CPE_ID.
The CPE id was for Debian's python-dns package which contains PyDNS
("pydns" on PyPI). This package contains "dnspython" from PyPI.
Signed-off-by: Jeffery To <jeffery.to@gmail.com>
Add possibility for user to provide setup and cleanup scripts for
additional flexibility. Setup-script takes precedence over the built-in
behavior of uacme.
This helps users with more complex use-cases to utilize uacme to update
certificates without adding complexity to the provided run.sh script.
Signed-off-by: Antti Seppälä <a.seppala@gmail.com>
Add possibility for user to provide setup and cleanup scripts for
additional flexibility. Setup-script takes precedence over the built-in
behavior of acme.
This helps users with more complex use-cases to utilize acme to update
certificates without adding complexity to the provided run.sh script.
Signed-off-by: Antti Seppälä <a.seppala@gmail.com>
Host build was added in 2972cc98e ("Add host build") because classpath
required a host java vm. Now that classpath is gone, host build can be
removed as well.
Signed-off-by: Eneas U de Queiroz <cotequeiroz@gmail.com>
An interface can have both a /64 and a /128 from a provider.
In such a case, use the address from the /64 to do the ping check, not
the /128.
Signed-off-by: Brian J. Murrell <brian@interlinx.bc.ca>
The recent update to ffmpeg broke this.
Instead, use the relevant libraries separately.
Added AAC support. AAC is very common and was recently fixed to be
compatible with BUILD_PATENTED.
Explicitly pass all configure arguments, to avoid future breakage.
Remove autoreconf. autotools files are not being patched.
Add PKG_BUILD_PARALLEL for speed.
Other minor cleanups.
Signed-off-by: Rosen Penev <rosenp@gmail.com>
The package Makefile contains a PKG_BUILD_DEPENDS=libiconv
line, which apart from being incorrect if libiconv-full is
specified in the build configuration, is also unnecessary,
since the package Makefile already includes nls.mk which
sets PKG_BUILD_DEPENDS appropriately.
Signed-off-by: Ian Cooper <iancooper@hotmail.com>
Since 19.07, netifd will not try to bring a device up
after someone brought it down. This way, there is no need
to rename the device.
It also fixes a rename error on 19.07 when the *_down device
was being in use (by netifd?).
Signed-off-by: Luiz Angelo Daros de Luca <luizluca@gmail.com>
perl fails to compile on x86_64 with glibc and stack smashing
protection enabled due to libssp not being specified in the
ldflags.
Signed-off-by: Ian Cooper <iancooper@hotmail.com>
The package Makefile contains an unnecessary PKG_BUILD_DEPENDS
line that builds gettext and libiconv even in cases where the
build has gettext-full and libiconv-full selected.
This behaviour can and does mask errors in other package Makefiles
that are dependent on libiconv-full if it is compiled before
these packages by causing libiconv-stub to be created and put in
the staging tree and potentially linked against by these
dependent packages when they omit to specify an appropriate
PKG_BUILD_DEPENDS line.
procps-ng does not require gettext or libiconv to be built in
order to compile correctly.
Signed-off-by: Ian Cooper <iancooper@hotmail.com>