Platform/GFX/2011-September-26: Difference between revisions

no edit summary
No edit summary
No edit summary
 
Line 7: Line 7:


Agenda:
Agenda:
* GFX engagement steward? (joe)
* GFX engagement steward? (joe)  
** bjacob volunteers
** bjacob volunteers  
** More information: https://wiki.mozilla.org/Stewards and http://htmlpad.org/steward-program/
** BenWa talks about marking bugs as good first bugs and having a mentor. Mark your bugs as [mentor=you@yours.com] in the status whiteboard, and make sure there's a good description. Include a list of files that will likely be affected.
* Bugsmash day/week/month? (joe)
* Bugsmash day/week/month? (joe)
** We have lots of bugs, and no idea how many of them are real.
** We have lots of bugs, and no idea how many of them are real. 2964 bugs: http://bit.ly/qzW7ie
** We can fix this! And then we can keep it that way!
** We can fix this!
* GL Layers on Mobile (ajuma, benwa)
*** jpr's experience is that a developer can get 25-30 bugs dealt with a day. That's 100 person-days
** And then we can keep it that way!  
*** "Graphics Sheriff" idea - rotating duty to make sure new bugs are correctly marked, assigned as necessary, etc.
*** Potential for multiple components, with one person assigned to a component to ensure that all "300-ish" bugs are real.
* Graphics & Layout work week (joe)
** New Zealand - Auckland work week for mid-November
** Need a head count; Joe's going to get that ball rolling.
* GL Layers on Mobile (ajuma, benwa)  
** Fix to the last major visual glitch has landed, so now it's just bugfixing
** Reftests were designed to run in a resolution of 800x1000, but when we want to debug on a Nexus S (which has a resolution of 480x800) several reftests fail for lack of rendered output.
** BenWa has been working on creating a resolution attribute for reftests to let us skip certain reftests on smaller devices.
* Testing on Mali GPU (cjones)
** Galaxy S II's GPU
** Needs testing, fixes
** We need a device in Toronto - will get. Have to be careful because there's one SII with the Mali, and one with the Tegra, and we want to avoid the latter.
* Mobile-specific optimizations
** Screen on the Galaxy S II is natively 24-bit, not 16-bit, and the Mali's driver apparently doesn't support 16-bit surfaces very well.
** Need to investigate the optimizations we've put in, and make sure we've made the quality/performance tradeoff in the right place.
Confirmed users
856

edits