Josef Schlehofer authored 4 years ago and Jan Pavlinec committed 4 years ago

We have our own package turris-webapps-transmission, so we don't need to
patch transmission package.

It should check if we can apply patches to
openwrt or feeds branches before merging it to HB*

Josef Schlehofer authored 4 years ago and Jan Pavlinec committed 4 years ago

In this pull commit, I backported hwinfo which will help us to provide system information, monitor hardware information and diagnostics.
hwinfo is already in packages repository, but it will be part of OpenWrt
21.xx release.

This resets uboot environment. We of course do not want to do it and
even if so we do not plan to do it right now or possibly this way.

While building node packages on Ubuntu 20.04 LTS, node fails because of:

In file included from /usr/include/c++/9/deque:60,
                 from /usr/include/c++/9/queue:60,
                 from ../deps/v8/src/libplatform/default-foreground-task-runner.h:8,
                 from ../deps/v8/src/libplatform/default-foreground-task-runner.cc:5:
/usr/include/c++/9/bits/stl_algobase.h:59:10: fatal error: bits/c++config.h: No such file or directory
   59 | #include <bits/c++config.h>
      |          ^~~~~~~~~~~~~~~~~~
compilation terminated.

This can be solved by installing g++-multilib

These packages are necessary for running Gitlab CI on Omnia/Mox.
This should help knot-resolver team to test kresd on
different architectures.

Note:
tor-hs package helps with creating tor hidden services
on router. Tor hidden service can be used to access router
services from Tor network.

This allows primarily to include all drivers in medkit. The idea is that
some devices should work out of the box so we ensure they can be used to
connect to the Internet. Once connected any unneeded drivers are
removed.

This can be considered as interesting feature from end user's
perspective so we should mention it in NEWS.

These contain database of USB and PCI device mapping to
drivers/packages. Idea is that updater with additional Lua script on
router can install this way drives automatically depending on what is
actually connected to router.

It also allows installtion of all or just some group of drivers. That is
to include them optionally also in medkit.

Because commit subject is too long, I'd need to choose word accepted
even though it was applied (I mean cherry-picked) by OpenWrt developers
to OpenWrt 19.07. So, we don't need to have it in backports anymore.

In OpenWrt 19.07 packages feed, there was updated zerotier from master
branch to version 1.6.2 and it includes same workaround

In OpenWrt 19.07 packages feed, there was updated zerotier from master
branch and it includes same workaround

dhparam package was removed so it is not longer available for install.
Instead turris-cagen dhparam generation should be used.

Karel Koci authored 4 years ago and Karel Koci committed 4 years ago

This triggers removal of uboot environment configuration from system so
we can set there correct one by installing new version of uboot-tools.

In case of Turris Mox we have to do also additional trickery as file we
want to generate is part of mox-support package originally. With this
update it is dropped but we have to let it be removed fist before we
recreate it by installing new version of uboot-tools otherwise updater
would remove that file as part of mox-support package cleanup phase.

Karel Koci authored 4 years ago and Karel Koci committed 4 years ago

Packge Mox-support was providing uboot environment configuration file
but more appropriate is to use OpenWrt approach with generated
configuration.

This should move Mox little bit closer to upstream.

By replacing Repository with provided one user can generate package
lists pointing to repo.turris.cz. That means it uses packages from there
but lists are local one. Of course user can replace hbs with any other
branch name and such.
The primary point there is to specify public keys.

Co-authored-by: Karel Kočí <cynerd@email.cz>
Co-authored-by: Vojtech Myslivec <vojtech.myslivec@nic.cz>

This is just stylistic change. It removes dirname and instead relies on
string operation. There should be no way of having in $0 anything else
other than some valid path. That is last part has to be at least
/name_of_file. This means that we do not have to rely on dirname and we
can safely just remove anything after last slash (including that slash).
We user dirname only for paths in sub-directories.

This fixes bug that caused force-make to go to fallback immediately
after first batch of retries.

As a benefit it also adds some reporting of what is going on. This
should help to locate fail point as well as to inform about execution
progress.