User:DanielDrake/11.3.0 goals

From OLPC

These are my personal goals/plans for 11.3.0, for discussion.

Budget: 10-12 hours per week

[edit] Network stability

• I'd like to make this my main goal - improve network stability by fixing all those pesty issues we have lying around
• XO-1 not connecting to encrypted networks on reboot, consistently need to click once
• Check for optimal programming speed of encryption keys, needed for reliable WPA1 (revisit the libertas work from 8.2.1 that was never carried forward)
• usb_reset() needlessly called on XO-1 wifi during resume - causes light to blink
• ticket #11062: XO-1 actrd can't connect to some access points
• ticket #9153: XO no longer tries to connect wlan after resume from suspend
• ticket #10971: Failure to automatically reconnect to Sugar Ad-hoc network
• ticket #10725: Sugar Ad-hoc network: reconnect after suspend
• ticket #10366: idle suspend causes network connection in progress to fail
• ticket #10543: Wireless disabled but Sugar Ad-hoc icons present when switching back from GNOME
• ticket #10549: Disabling the "Radio" in Sugar disables ability to use any networking adapter
• ticket #10694: After XO-1 sleep, there is a 3-4min delay before wireless access points appear
• ticket #10780: Sugar Network view looses track of available networks but iwlist is fine
• ticket #10232: WiFi dies on suspended XO-1, os300
• ticket #10375: Lid close/open turns on WiFi on suspended XO-1
• ticket #10627: os5, XO-1: Network fails to reconnect after suspend
• (many of the above tickets overlap greatly)

[edit] Image size reduction - particularly for XO-1

• Remove mesa-dri-drivers
• Remove cracklib dicts? linux-firmware?
• Move to ubifs again, but use zlib compression
• In boot partition on signed images, use signed files even in unsigned mode, no need to store unsigned copies

[edit] Misc low-level

• ticket #10967: Synaptics touchpad may result in spurious PS/2 mouse data upon resume
• ticket #10568: XO-1 graphical boot hangs if SISUSB VGA adapter attached

[edit] Release management, deployment and community engagement

• Continue weekly triage meetings, even if not so triage-orientated
• Invite Chris Leonard to weekly meetings for better coordination with localization efforts
• Document out how localization interacts with the release process (when do string freezes happen? who announces them? etc.)
• Try to get a couple of teams meeting on a weekly basis for testing and awareness (Nicaragua & La Rioja?)
• Look closely and prioritise pending requests from responsive deployments, encourage more of this