[DEV] sfdroid: run android in a window
Update from the field:
https://www.youtube.com/watch?v=78N1C2-6t3I This will be released soon :). Instructions are on wiki.merproject.org https://wiki.merproject.org/wiki/Ada...for_hammerhead https://wiki.merproject.org/wiki/Ada...ishOS_for_mako Old instructions: Code:
Don't flash this if you're not able to recover from it. It breaks some things. I hope i can make it better. ONLY FLASH ON A ROM YOU DON'T NEED. |
Re: Idea for an Android Compatibility Layer
just out of curiousity, would the use of lxc/docker help?
|
Re: Idea for an Android Compatibility Layer
Quote:
|
Re: Idea for an Android Compatibility Layer
I wonder why a TCP connection. Shouldn't a local Unix datagram socket be good enough but without the additional overhead of network and TCP?
|
Re: Idea for an Android Compatibility Layer
Quote:
|
Re: Idea for an Android Compatibility Layer
update:
surfaceflinger does not crash anymore: https://bpaste.net/show/5520d6e5bc9b but rendering does not work yet... initializing renderer process listening on port 22468 renderer process started creating sub window created sub window FrameBuffer::post accepted client accepted client |
Re: Idea for an Android Compatibility Layer
Thanks! Does the chroot work(run!) on N9 or it should be compiled from sources?
|
Re: Idea for an Android Compatibility Layer
Quote:
|
Re: Idea for an Android Compatibility Layer
Quote:
i fear we wont get it to work on the n9 though, because the renderer requires some egl/gles features which we don't have on the n9 |
Re: Idea for an Android Compatibility Layer
current issue, please help, eglCreateContext from libhybris segfaults if share_context != EGL_NO_CONTEXT
it crashes inside the adreno blobs, can we do it without that pbuffer context? |
Re: Idea for an Android Compatibility Layer
Quote:
About the missing egl/gles features, isn't it possible to work around them somehow (e.g. providing a software implementation using available features?)? |
Re: Idea for an Android Compatibility Layer
Quote:
|
Re: Idea for an Android Compatibility Layer
i have updated the chroot, this one does not segfault anymore in surfaceflinger, and i have uploaded the renderer sources to github so you can contribute :)
|
Re: Idea for an Android Compatibility Layer
current issue (the eglCreateContext one is still there, but i think this one is more pressing):
rcFBPost is never called EDIT: fixed, current issues: 1) how to create subwindows in wayland 1.3 2) how to make the wayland window appear on screen |
Re: Idea for an Android Compatibility Layer
ok i updated the original post, there are two files which you can copy over the current chroot (if you have the most recent one) init.rc will fix usb problems and gralloc.default.so will make it use the correct gralloc version.
now we have to figure out why the sdl window does not appear |
Re: Idea for an Android Compatibility Layer
sorry but i did not understood. it is a way to using android app on sailfish for nexus 5? is viable? have u tryed to suggest it on #sailfish-porters?
|
Re: Idea for an Android Compatibility Layer
Quote:
|
Re: Idea for an Android Compatibility Layer
i have a n5, i m not skilled but if u tell me what to do i could try,maybe in next week
|
Re: Idea for an Android Compatibility Layer
Quote:
|
Re: Idea for an Android Compatibility Layer
maybe u dont know with who u are speaking...i m completely noob,so u have to followme step by step..ahahhaah
|
Re: Idea for an Android Compatibility Layer
Quote:
And another question: The EGL/GLES problem with N9 you mentioned is with N9 itself or with Harmattan? (will it work on SFOS port on N9?) Regards |
Re: Idea for an Android Compatibility Layer
Quote:
|
Re: Idea for an Android Compatibility Layer
You should say in the first post that you are working only for the N9 and Sailfish, but not the N900 ! :'(
Few months ago, I discovered DebianKit, which aims to install a full classical Debian directly in parallel with the existing Android install. Since Harmattan is more or less a Debian and Sailfish is still GNU/Linux, maybe something like that could be achieved for using Android apps (in addition, if it works with Harmattan, we will be able to do the same on the N(eo)900). |
Re: Idea for an Android Compatibility Layer
Quote:
|
Re: Idea for an Android Compatibility Layer
https://bpaste.net/show/5338d7a8aafe
this is a current log with some debug info, wl_surface.attach is never called, this is the current issue |
Re: Idea for an Android Compatibility Layer
current state: i talked to the #wayland guys and it seems egl pbuffers are not supported in conjunction with wayland...
EDIT: maybe it is a bug on my end... |
Re: Idea for an Android Compatibility Layer
so what it means for common mortals?
|
Re: Idea for an Android Compatibility Layer
Quote:
|
Re: Idea for an Android Compatibility Layer
no way to connect those systems? u have to look for a different solution?
|
Re: Idea for an Android Compatibility Layer
Quote:
EDIT: i added some contact information on the first post |
Re: Idea for an Android Compatibility Layer
Quote:
|
Re: Idea for an Android Compatibility Layer
hmm...sorry, what is the advantage of using your own module over using unix local sockets (or even FIFOs). If both ends of the communication are in userspace, I wonder if a custom kernel module will be faster than unix sockets?
|
Re: Idea for an Android Compatibility Layer
Quote:
|
Re: Idea for an Android Compatibility Layer
okay i am facing a pretty weird bug... no window is displayed... i tried to make an application which des exactly the same as the renderer and that works... then i tried to copy this application into the renderer and that still does not display a window! :O???
|
Re: Idea for an Android Compatibility Layer
once i link against utils the test application fails, but i don't know why yet...
|
Re: Idea for an Android Compatibility Layer
ok the window does appear now :) but it renders garbage
|
Re: Idea for an Android Compatibility Layer
|
Re: Idea for an Android Compatibility Layer
Quote:
|
Re: Idea for an Android Compatibility Layer
Quote:
|
Re: Idea for an Android Compatibility Layer
Quote:
EDIT: current issue: GL_INVALID_OPERATION, this error comes from the blob drivers when the chroot calls glCopyTexImage2D, what is weird though, i added printfs in the decoder methods and it seems glCopyTexImage2D is never called on the host |
All times are GMT. The time now is 02:54. |
vBulletin® Version 3.8.8