2017-03-04 13:15 GMT+01:00 Ilario Gelmetti <iochesonome(a)gmail.com>om>:
I copy and paste the pad sorted by assigned person, there are many
easy tasks if you want to go for them :)
(sorry for cross posting)
===================
Whoever can do this
-------------------
== Improve online documentation ==
Fill the gaps in the quick starting guide and add examples of common
configs (e.g. 1 port router, set the port as WAN)
Add link to Italian translation (NinuxBO)
Complete the big tested hardware table
http://libremesh.org/docs/hardware/index.html
== Test LibreNet6 into LibreMesh and produce documentation on the process ==
This can be useful for GSoC project on LibreNet6 implementation in LibreMesh
https://wiki.freifunk.net/Ideas#LibreMesh_LibreNet6_integration
There are users trying it:
https://lists.libremesh.org/pipermail/lime-users/2017-March/000400.html
Gui and Nico are needed for deciding criteria for delivering /64 from LibreNet6
== WPA access point for clients and for backbone ==
Thread:
https://lists.libremesh.org/pipermail/lime-users/2017-January/000325.html
GSoC idea:
https://wiki.freifunk.net/Ideas#LibreMesh_WPA2_and_802.11s_mesh_encryption_…
It is already supported for AP and client, just we need to document it.
== Advertisement of student positions in GSoC program ==
If we don't advertise it and we're not present in IRC channel for
answering them, we'll have just a few students
General mailing list for Freifunk's GSoC:
http://lists.freifunk.net/mailman/listinfo/wlanware-freifunk.net
hackmeeting-es (Ilario)
hackmeeting-it (???)
guifi-users (Al)
hacklab Dontpanic (Al, Gio)
eigenLab (Ilario)
retroshare forums (Gio)
student application period: March 20th - April 3rd
suggesting to contact via lime-users list (our presence in IRC is quite scarce)
== Use Gio's thesis for lime-web ==
adapt contents and insert them in lime-web
and fix link to the thesis s/~gioacchino/~gio/ in lime-web
=====================================================
Whoever can do this but I will do if no one else does
-----------------------------------------------------
== Improve offline documentation ==
Add a reminder for "lime-config" command in the header of the
configuration files
== lime-build: mt7621 not an available target ==
just add that target in the available targets in targets.mk
https://github.com/libremesh/lime-build/issues/15
This has already been fixed by Pau in f92c2d5 in develop but that
commit is not present in branch 17.02
Do a new PR adding just this, ask for merge in 17.02
== lime-build: make bin-clean ==
A make option for deleting binary images in build/src/bin directory
can be useful and should be part of the make process for avoiding
confusion:
https://lists.libremesh.org/pipermail/lime-users/2017-February/000373.html
make clean is not the same as deletes a lot of stuff that has to be
compiled again
== Add spaces and client_encryption 'psk' to lime-example ==
https://github.com/libremesh/lime-packages/blob/develop/packages/lime-syste…
== lime-build: specify profile in make menuconfig in readme ==
https://lists.libremesh.org/pipermail/lime-users/2017-February/000371.html
== lime-build: Avoid error from make if targets/*.kernel is absent for
the current target ==
https://lists.libremesh.org/pipermail/lime-users/2017-February/000389.html
Decision: put an IF before the cp statement
=============
I can do this
-------------
== Update and reopen lime-example update PR ==
https://github.com/libremesh/lime-packages/pull/67
================
Devs can do this
----------------
== Add an option for customizing DHCP range ==
https://github.com/libremesh/lime-packages/issues/33
https://lists.libremesh.org/pipermail/lime-users/2017-February/000377.html
https://lists.libremesh.org/pipermail/lime-dev/2017-February/000842.html
== Merge and close PR on IPv6 GW detection ==
https://github.com/libremesh/lime-packages/pull/71
== Check compatibility with new versions of LuCI ==
Verify the PR and check for other similar issues
https://github.com/libremesh/lime-packages/issues/73
== luci-app-batman-adv and Alfred: a 3 years old problem, still open ==
https://github.com/libremesh/lime-packages/issues/1
confirmed is still an issue, maybe not needed because the web
interface will be rewritten
https://github.com/libremesh/lime-packages/blob/develop/packages/luci-app-b…
https://github.com/libremesh/lime-packages/blob/develop/packages/luci-app-b…
== lime-webui-ng will be included in 17.02? ==
Pau?
Unitl when it does not work should be moved in a specific branch, not
in develop as now is.
https://github.com/libremesh/lime-packages/tree/develop/packages/lime-basic…
== lime-build: Document how to add new hardware targets ==
Up to now just Pau did it
Split specific taget lines from generic lime configs in targets/*
files and move these ones to an unique file.
== Redirect
libre-mesh.org to
libremesh.org ==
And check on the internet where we can fix it
https://www.google.es/search?q=%22libre-mesh%22+-site:libre-mesh.org
Gui should do this
== Update or close repository on GitLab ==
Having an out of date version there is not good
https://gitlab.com/libre-mesh/lime-packages
Decision: we will set up syncronization
Gio should do this
== Update or close repositories on GitHub ==
We have a lot, that looks like a ugly list
https://github.com/libremesh
openwrt-oldpackages could be useful for someone compiling 15.09, shall
we keep it? Yes, add comments in the repo description
https://github.com/libremesh/openwrt-oldpackages
lime-field-tools, nicoechaniz... are we using it or updating it or something?
https://github.com/libremesh/lime-field-tools
an old presentation on LibreMesh, we could make a repo for graphics
and move it there, adding logos etc. Or can be moved in lime-web but
should be updated.
https://github.com/libremesh/libre-mesh-tecnical-presentation-1h-reveal.js
libremap moved to its own user 3 years ago, so the repo in our
organization is not useful. Delete it.
https://github.com/libremesh/libremap
ruci is an old script from Gui, let's ask him
https://github.com/libremesh/ruci
FFT_eval should be needed
https://github.com/libremesh/FFT_eval
another old presentation, see the previous comment
https://github.com/libremesh/libremap-talk-2013-is4cwn
super old repo superseeded by libremap
https://github.com/libremesh/altermap
ask Pau if we can delete this
https://github.com/libremesh/dhcpdiscover
let's remove this
https://github.com/libremesh/demeshtify
Devs should confirm we can delete these
== Test and document how to use nodogsplash with LibreMesh ==
During last LiMeCat2016q3 we decided to not include nodogsplash in
LibreMesh, but some GSoC projects go in that direction.
https://wiki.freifunk.net/Ideas#LibreMesh_Captive_Portal_and_Access_Control
Also nodogsplash is something which users ask for:
https://lists.libremesh.org/pipermail/lime-users/2016-December/000287.html
We need Gui for this.
== Verify users reports ==
=== Hostnames propagation ===
https://lists.libremesh.org/pipermail/lime-users/2016-November/000223.html
=== Strange errors ===
This happened long time ago, likely it's solved:
https://lists.libremesh.org/pipermail/lime-users/2016-October/000197.html
=== Publishing on the website a guide for captive portal and coupon sharing? ===
https://lists.libremesh.org/pipermail/lime-users/2017-January/000317.html
=== Speed and latency issues ===
https://lists.libremesh.org/pipermail/lime-users/2017-February/000374.html
=== Mysterious /8 route in BMX6 ===
https://lists.libremesh.org/pipermail/lime-dev/2016-October/000787.html
info in p.s., not in the main email body
not an error, means accept routes in this range
=== trap in bmx6-auto-gw-mode logic ===
https://lists.libremesh.org/pipermail/lime-dev/2016-November/000797.html
we can fix adding uci commit at the end of each watchping check
=== not existing luci-i18n-english ===
https://lists.libremesh.org/pipermail/lime-dev/2016-December/000818.html
=== Libremesh & qMp don't ping with same BMX6 revision ===
https://lists.libremesh.org/pipermail/lime-dev/2017-February/000837.html
== No watchping, no pingcheck, better a wget ==
meh meh meh??
== lime-build: Need an alert for non existent profile ==
It affected an user here:
https://lists.libremesh.org/pipermail/lime-users/2017-February/000389.html
The issue is equivalent to this but for profile:
https://github.com/libremesh/lime-build/issues/1
confirmed to do
== Chef should ask for a root user password to be included in cooked firmware ==
For avoiding attacks while the password is still not set
https://lists.libremesh.org/pipermail/lime-dev/2017-January/000827.html
Chef could leak cleartext password, Chef should make the hash itself.
== Discuss how a testbed for Battlemesh can be obtained from LiMe ==
Directions to forward to the student who will work on this GSoC idea
https://wiki.freifunk.net/Ideas#LibreMesh_based_Battle_Mesh_testbed
lime-proto-babel
lime-proto-***
ask Nemesis for how they did the alternative WBM firmware
== Develop a test suite with qemu or similar ==
We need to test the new commits against a few sample configurations
and check that everything continues working.
As doing this with real devices would be hard, something virtual would
be accessible.
Decision: much more useful to make something with real hardware, but
needs to be able to reset to previous firmware. So we will never do it
this way.
So... Which virtual machines platform we can use for this?
==================================================
Devs can do this but I will do if no one else does
--------------------------------------------------
== Remove luci-mod-lime-basic ==
https://github.com/libremesh/lime-packages/issues/72
== Complete and test lime-proto-client ==
https://github.com/libremesh/lime-packages/issues/47
For now the network "wan" is hardcoded, it should be possible to use
it for lan also
When we have it for lan, we will have to test a non-mesh network
(backbone made by AP-sta links)
prio
Remove args["network"] = "wan" for avoiding lime-proto-wan dependency
of lime-system
== lime-build: close an old PR - build sdk for all profiles ==
either completing it or refusing it
https://github.com/libremesh/lime-build/pull/8
== lime-build: Re-add bmx7olsr1 profile to PROFILES_AVAILABLE ==
https://github.com/libremesh/lime-build/commit/f92c2d5354a62518435823153df2…
Pau deleted this by error
== Start using projects or labels and milestones for issues on Github ==
lables can indicate easy-to-address bugs for newbies
we can transfer on github the majority of the issues listed in this document
or include them as tasks in "projects" in Github
Decision: yes for labels
== Add topics to lime-packages repository for imrpoving visibility ==
As we're on Github, let's take advance for having a bit of visibility
https://github.com/blog/2309-introducing-topics
Decision: yes
== Change link to old PabloCastellano repo in alterchef repo ==
https://github.com/libremesh/alterchef/blob/master/README.md
And add description
=============================
Completed tasks and decisions
-----------------------------
== Discuss if "lime-config" could be inserted in LEDE init ==
This could broke custom configurations but avoids the common mistake
of forgetting lime-config
Decision: no.
== lime-build: Discuss if we need a profile with 802.11s configured as
default ==
This way it would be easier for the users to build a 802.11s based mesh.
Decision: we will have just 802.11s
https://github.com/libremesh/lime-web/commit/0d147690bb1a0a97f8c301dfcbe1ed…
== Change BSSID and VLAN ID of BMX6 for compatibility with QMP ==
As we will broke compatibility with 802.11s, we can rediscuss other
things like BSSID and VLAN ID of BMX6, we can change these for making
easier the interoperability with QMP.
For wireless link the same BSSID is not enough as QMP uses ad-hoc and
we will use 802.11s
Decision: yes
In qMp
default BSSID: 02:CA:FF:EE:BA:BE
default VLANID: 12
== General switch to 802.11s for link layer? ==
ad-hoc seems to have a poor support by widely used drivers, 802.11s
seems better (e.g.
https://lists.libremesh.org/pipermail/lime-users/2017-January/000318.html
)
The later the more painful will be the change.
In LiMeCat2016q3 the change was decided for the next release, which is the 17.02
(cfr the pad of LiMeCat2016q3
https://pad.eigenlab.org/p/LiMeCat2016q3 )
Decision: confirmed, we need testing. We won't have
back-compatibility. Network who will stick to ad-hoc will have to keep
16.07. Or configuring 17.02 with ad-hoc in Chef or post-flashing. Can
cause problems.
Decision: we need a super easy way in the webui for switching from
802.11s to ad-hoc
== Plan deployment of LibreNet6 configuration helper webpage ==
In "mid-2017", similar to vpn.qmp.cat
https://lists.libremesh.org/pipermail/lime-users/2016-November/000204.html
Decision: WTF? Centralizing TINC configuration on a server? No.