Marty R [Thu, 16 Oct 2014 04:30:25 +0000 (21:30 -0700)]
nut: New package for Network UPS Tools
An older version of nut is in oldpackages. This commit is based on that
package and bumps the version to the latest release along with unifying
the server and client packages. More build options are provided for any
custom configuration but the defaults provide a working client and
server install which only needs a UPS driver. Drivers still build as
individual packages to minimise bloat. SSL support has also been added.
sqm-scripts is a new QoS system from Cerowrt / Bufferbloat.
I have tested it in my own community build.
More info at:
http://www.bufferbloat.net/projects/cerowrt/wiki/Smart_Queue_Management
http://www.bufferbloat.net/projects/cerowrt/wiki/Setting_up_SQM_for_CeroWrt_310
Dave Täht submitted this first via openwrt-devel mailing list in March 2014,
but that was not picked up before the switch to Github:
http://patchwork.openwrt.org/patch/5078/
I have modified:
- set the default interface to be eth1, like the Openwrt default.
- modified config to hide some of the advanced options by default
- beautified Makefile a bit
- added Toke Høiland-Jørgensen <redacted> as the maintainer, as discussed
in https://github.com/dtaht/ceropackages-3.10/issues/8#issuecomment-59215905
According to PEP394 (http://legacy.python.org/dev/peps/pep-0394/)
the 'python' command should refer to 'python2'.
In our case, this means we should reboot the old python package.
We could rename the package name to python2, but that would
just complicate things a bit with other packages, and
since we're doing this reboot, such a complication would be
unnecessary.
John Crispin [Wed, 8 Oct 2014 01:45:30 +0000 (03:45 +0200)]
generic userid: use the new USERID api
recent changes in trunk allow us to specify the userid inside the openwrt makefile.
the info is stored int he meta data of the IPK contorl file and users are generated
by the new generic postinst trigger.
ruby-core is problematic as it is too big.
It is impossible to fix pkgs dependencies as
ruby-core would generate multiple cycled dependencies
between packages.
Also, "core" in ruby context means "classes that does not need a 'require'".
This is not the case of ruby-core classes. They are, actually, a subset of
Ruby Standard Library.
In every detected case where a portion of ruby-core could be isolated and
save another pkgs from requiring all ruby-core where spin-off into a new
subset. Also, big portions of ruby-core, not require by current ruby-* pkgs
where spin-off in new pkgs. The remaining of ruby-core was put into a new ruby-misc.
ruby-stdlib was created as a meta package that requires all ruby packages that are
part of Ruby Standard Library. For a full Ruby Standard Library, just install
ruby-stdlib and its deps.
Ruby encoding library was too big and bring unecessary encodings for a simple ruby usage.
All not directly required encodings from stdlib where moved to ruby-enc-extra.
Created pkg from ruby-enc
- ruby-enc-extra (from ruby-enc)
Signed-off-by: Luiz Angelo Daros de Luca <redacted>
ruby_find_pkgsdeps: look for file dependencies (checks
require and Encoding references) and extrapolate it to pkgs
deps. Also checks whether a dep is redundant or missing in pkgs.
Must run inside an OpenWRT with all ruby* pkgs installed.
ruby_missingfiles: list files in staging/target and from files
comparing side by side its contents. It helps to easly visualize
which file is not packaged in an ipk.
Signed-off-by: Luiz Angelo Daros de Luca <redacted>
Digest can use OpenSSL or ruby internal implementation of hash functions. The first
uses less disk space but requires openssl, that is relatively big. As internal hash
implementations are not too much bigger than openssl version, it is compiled by
default. A new config option can change it to use OpenSSL instead.
As digest is independent from openssl, ruby-digest was created as a new pkgs.
Adds pkgs:
- ruby-digest (from ruby-openssl)
Signed-off-by: Luiz Angelo Daros de Luca <redacted>
Some ruby gems where still in ruby-core pkg. These files where
moved outside ruby-core into ruby-gems or their own subpkg.
ruby-unit renamed to ruby-testunit as its gem is named test-unit.
ruby-rdoc left a file in ruby-core.
Psych is a gem and deserves its own subpkg. It replaces syck
(used by yaml) on recent ruby version (ref:
https://bugs.ruby-lang.org/projects/ruby-trunk/repository/revisions/36786)
Also, some psych files where packed incorrecly into ruby-json. The asterisk
in */json was intend to match <arch>/json/ and not psych/json.
Files where derived
from ruby-core and a lost file in ruby-json.
New subpkgs:
- ruby-bigdecimal
- ruby-io-console
- ruby-minitest
- ruby-psych
Signed-off-by: Luiz Angelo Daros de Luca <redacted>
rdoc seems to be written to run without gem. However,
some internal code still does not check for gems presence.
With a small patch, rdoc can run without gems.
Ref: https://bugs.ruby-lang.org/issues/10196
Signed-off-by: Luiz Angelo Daros de Luca <redacted>
ruby: wrap script to disable gems if not installed
ruby, by default, try to load gems on start. If not
present, it fails. As gems brings many deps, this make ruby
unsuitable for routers limited resources.
ruby can avoid to load gems with the option "--disable-gems".
So, a wrap script in the place of /usr/bin/ruby adds this option
if gems are not found.
Also add vendor/site directories
Signed-off-by: Luiz Angelo Daros de Luca <redacted>
Docs are not currently packaged. So, there is no need to install
them. Also, doxygen have some problem with the usaged of git in
build_dir as it takes ages running git commands on every file.
Disabling the docs reduces the compilation time.
Signed-off-by: Luiz Angelo Daros de Luca <redacted>
Some notes:
- Python 3 (at least version 3.4) is pretty cross-compiler-friendly,
so a lot of patches were thrown away.
- Arguments below were moved to ./files/config.site file, and disabled
ac_cv_have_chflags=no \
ac_cv_have_lchflags=no \
ac_cv_py_format_size_t=no \
ac_cv_have_long_long_format=yes \
ac_cv_buggy_getaddrinfo=no \
- --without-ensurepip added, because the build wants to ensure that
it works; that's a good idea, but for now, it requires special
setup, and we can do that later
- --without-pymalloc added, becase in Python 3, modules are suffixed
with m; e.g. so some paths are python3.4m instead of python3.4
all this is detailed here:
http://legacy.python.org/dev/peps/pep-3149/
Maybe it will be a good idea to re-add this back
Make python-mini the top-level and default Python package.
Make the other packages subpackages of the top-level python package.
Make old python package the python-full package.
Note: at this point, this package can be split/renamed to
python-legacy should anyone need Python 2. After this commit
it's going to be Python 3 all the way.