-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Somebody in the thread at some point said: | On Thursday 08 January 2009 16:58:32 Andy Green wrote: |> Somebody in the thread at some point said: |> | 1.) Get the code upstream into the linux kernel |> |> We feel showing it can remove the need for tslib gives us a story that |> gives a better chance upstream. | | I think one can imagine that this will work. But what about actually sending | the idea + patch set to the linux input device list? I think it makes sense to | check if the people maintaining the linux input framework would be willing to | have and maintain code that is doing that.
The implementation already exists. |> | that goal, we walk into the other direction. |> |> We went in the direction of resolving a longstanding problem for users, |> GTA02 touchscreen was unusable until this was done. |> |> There's no point to repeat the cycle of "shoulda done it in tslib" |> because that's out of scope for me and nobody else took action. | | I don't know what is wrong... I highlighted a migration path. If you don't | want to do this, fine. Just to repeat what I said: Yes I read and understood it. But it seems I need to repeat what I said back on 29th Dec: ''The two things about ~ - cleaning up the touchscreen data in kernel, and ~ - whether to maintain tslib linkage in userspace for basically API compatability reasons are separate issues... the first is the right thing to do IMO and the second is a useful option we're looking at because we think it might help get the kernel filter stuff upstream, but not proposing to get changed in OM userspace.'' | How to test the touchscreen: | Kernel: cat device... spitting out things but no numbers | tslib: ts_print and ts_print_raw with nice output | | How to calibrate the touchscreen: | Kernel: Aeehh... find the Determinate yourself solve the equations | tslib: ts_calibrate ... This just polarizes the nonexistent question, nobody is saying not to use tslib where it provides a service like creating calibration data and I and most recently Nelson point out tslib works fine with a clean input device. It's just fun to keep repeating it? | I just highlight what is missing in your story, if you want to push the | migration you should offer the same features as tslib or convince people why | these are not needed... I'm willing to help you migrating but it starts with See above. | 1st) convincing people maintaining the linux input device system that this is | a good idea (otherwise we will end up with GTA06 and have 100 "feature | patches" because we do clever things that upstream does not understand... and | this will be a nightmare for everyone involved). Wow nobody could have imagined. Thank goodness that wasn't the situation when I inherited the kernel. - -Andy -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.9 (GNU/Linux) Comment: Using GnuPG with Fedora - http://enigmail.mozdev.org iEYEARECAAYFAklmVi4ACgkQOjLpvpq7dMoxrwCcDSE00wHipKcjCAig4QI+tSgJ AYgAniyq1OuO6vMQshdolf1ogRajw8zY =soJw -----END PGP SIGNATURE----- _______________________________________________ devel mailing list [email protected] https://lists.openmoko.org/mailman/listinfo/devel
