Hi Entropy512,
Thank you for your work on CM9 and CM10. I really appreciate the time and effort you have spent. For the last year I have been keeping a track of (and using) CyanogenMod on the Note and curiousity prompts me to ask a question of you if it's not too much trouble (or impertinent): what drives you (or any dev) to spend effort doing this? Is it for the challenge, the pleasure that you derive working on the kernel?
As a full-time software engineer myself (and also an Android app developer), sometimes the last thing I want to do when I get home is to check the bug reports and support emails. I have become wary of adding new features simply because of the time that gets eaten up. I am really curious about those who give of their free time to the community.
Thanks.
It started as personal hacking just to tweak my own device, to some degree as a learning experience. Keeping my mind sharp, so to say. It's also a great software development learning experience. In my day job I'm a hardware guy/systems engineer although I work closely with embedded software developers - for me, a lot of this is about learning new hardware.
If I did this stuff for a living I'd probably hate it with a passion. (There's the answer, again, for those saying Samsung or Google should hire me.) To some degree I get bound by a sense of moral duty to stay with some devices I should just throw into a wall, but sometimes when you make a breakthrough with those it's exhilarating (like Bluetooth on Infuse).
That said I'm rapidly approaching burnout... I'm likely to easymode it on a Nexus doing minimal work, and/or returning to what I really wanted to do when I was hacking kernels, more highly experimental work. I've had a project (reducing suspend/resume power consumption cost) I haven't been able to even think about in months.
yeah it'd probably require Entropy and other devs to be simultaneously online I guess, so that they can discuss these things out.
Yeah, and this is always a problem. Usually I'm heading to work when espenfjo and codeworkx are awake, and when I get home from work, they're sleeping.
Hwcomposer - may be, memory leak - I hardly beleive. Just a leaked I9305 (SGS3 LTE) GPU drivers. Anyway should be very good commit.
Memleak - almost definitely fixed. Mali r2p4 drivers + gralloc leak like crazy on AOSP for some reason, r3p0 works far better. My build (only on my device right now...) and espen's are still using ics hwcomposer though, and it's highly likely we'll be stuck there. Opensource hwc has never worked quite correctly. This is really the root of codeworkx's anti-Exynos rant.
I think that these aren't related with our main problem.
Correct. Those fix minor memleaks elsewhere, not the massive gralloc leak from hell.