[bugs now fixed!] LG kernel source code for Optimus V is broken

[picasticks]

You're right, I knew they were distinct filenames but I didn't realize that one got overwritten with the other. That explains a lot. It does compile just fine now.

I'm a little disappointed that they still stripped out all of their comments in this tree and not just the emails (except for the Makefile, surprised they didn't catch the emails there). It sure makes the diffs harder to compare with the Sprint code. Two files that are otherwise exactly the same have lots of diff entries simply because all the LG comments are gone. Oh well.

Yeah, I agree. I read through the first 10% or so of my "mega diff" the other day, and in a lot of those they were now just removing the email address from inside the comments, which is good. They may have other reasons for doing it that way -- since their comment-scrubbing script was borked, they probably did what you suggested and realized just stripping \S+@\S+\.\S+ or whatever regex for an email address was simpler than trying to delete entire multiline comments cleanly.

If you Google "diff ignore c-style comments" you will see there are some people with sed scripts that will strip comments, so you can make a better diff.

You can also run diff with -I option to ignore lines matching a regex. So, something like "diff -I '^\s*//' -I '^\s*/\*.*\*/\s*$'" would ignore single-line comments at least.

If you open the diffs in an editor that does color highlighting of diff files, that also makes it a lot easier to just scan for "real" differences (I use vim which does this ... it actually doesn't take that long to go through the diff and delete all the "comment only" sections).

The kernel I posted compiled from the new source apparently has the home and menu buttons swapped (a hardware difference between the S and V) which is the usual problem with running an S kernel on the V. You'd expect a kernel compiled with the V source not to have this problem ... so it's possible there's still something iffy with the code (i.e. it's not the actual V kernel shipped with the V but some other fork of the S code). But there are other possible (nonkernel) explanations for this behavior, so it's also quite likely that the kernel source is correct. Not having a V myself to dig through and test on, I can't really investigate that any deeper, so I'm happy at least to declare victory on the "getting buildable source released" issue in this thread.

 

[picasticks]

I am not working on that so I don't know, you'd have to ask those folks. I don't know if zefie and other people building kernels for Cyanogen-based ROMs like CM7 have already gone through the VM670 kernel code and merged it or not, you could definitely suggest it on one of those threads now that the code is fixed.

Worst case they will say "duh, we went through that when it first came out a month or 2 ago".