search

OpenPDroid / platform_frameworks_opt_telephony

1 stars 4 forks source link

Phone FCs #1

Open mateor opened 11 years ago

mateor commented 11 years ago

With my recently remade Cyanogen10.1 patches (0209 I think) we are having a phone FC problem. I cannot yet tell if this is a patcher problem or what.

A user (with grouper, mind you, no Mms usually) reported that removing Mms.apk fixed the problem. So I thought that since the Mms.apk uses telephony-common libraries, that the prebuilt Mms.apk I sometimes copy over was at fault. But another user tried merely replacing it in his rom and still got FCs somehow. I don't see why deleting it in grouper would fix it but not in another rom...maybe grouper doesn't ever enable the telephony-common.jar unless the Mms.apk is available.

So now I am waiting for someone to try manually placing the older telephony-common.patch (from 20130201) into the 20130208 folder and see if the older telephony patch fixes the issue. It will produce an update.zip, i have verified that.

Here is the logcat from a phone FC...it is from grouper, but I associate these errors with the classpath not being able to find the telephony-common jar. I am just starting to look at it, but I figure i would post here as well in case another eyeball makes the difference. http://codepad.org/GVoYtNiy

mateor commented 11 years ago

Okay, this is real. My initial, unexamined, sense is that this is from the upstream commits, whatever it did that broke our original 1.0.1 smali patches is not playing nice with our build patches either. The build patches are applying fine but it is overwriting something it needs. I will look at it some tomorrow night, but it is Mardi Gras day, my family expects me...

mateor commented 11 years ago

I have fixes in place for this (basically just recycling the telephony-common patches from the 0201 patches). I am closing this until it rears its head again. Happened in PA 3.0 and CM10.1, but no evidence that it isn't just a local problem of some sort. I hope it is.

mateor commented 11 years ago

There is a problem here. I will be requesting logs from users. But even though the smali patches apply perfectly, all OS's report FCS with patches made from 0208 or so source. When I replace those patches (telephony with 4.2.1, no solution for CM10 yet...) with patches made from 0201, then the FCs stop. Somewhere we have a problem. and I am not sure where.

PD2.0 patches were applied at build time by karamelos, but both users report constant FCs as well. This may need to be our #1 priority, although I will not rule out user (read: me) error. Maybe there is a corruption or fault in my build tree, and maybe karamelos did something wrong.

But we have these reports in PA3.0, CM10.1 and CM10. They could be unrelated, but this has never happened in the year I have been doing this until now.

mateor commented 11 years ago

Well, here is one user error. It turns out that when the PDroid2.0 user, even though she said she was using PDroid2.0 1,54, they actually meant Autopatcherv2.6.67. So the Pdroid2.0 report in invalid. That does, however, mean that AOKP is added to the list of roms that have a telephony-common.patch problem.

I would think that maybe it was related to a problem with my telephony tree, yet it was also reported with the new Cm10 patches, which do not have a telephony patch.

wsot commented 11 years ago

Ok - so, I'm back. Finally.

AOKP have now integrated Android 4.2.2, which means everything is up for grabs (again) it seems. I'm in the process of fixing up my build environment for AOKP branch jb-mr1 (had some problems after the move) so that will be built very soon, then I can check it and provide you the files.

mateor commented 11 years ago

Most everything is current for 4.2.2 except for evervolv, which is going up this morning. We need rebuilt 4.1.2 for AOKP though, I believe. Checking the merge in then telephony is needed, though. Auto-patcher produces FCs with 4.1.2 up.except I have circumnavigated that by using an older telephony patch. I don't think of is a local issue, as it has persisted on builds from different build machines.

wsot commented 11 years ago

Cool, I'll get the build going for 4.1.2 AOKP. My builds of CM 10.1 since 4.2.2 was integrated don't boot - even without OpenPDroid, so I'll need to look into that.