Brice Arnould

Hi, and thanks for your contribution! It prompted me to actually simplify the build system using cmake and published a documentation, it should be mush easier now. Please check it...

It would invaluable indeed, I'd expect dependancies to be kernel headers, cmake and build-essentials. Could you please have a look at the documentation and tell me if it's good enough?

`shake` tries really hard not to destroy files, at worst it may let you know that the file has been renamed. The error message should look like `failed to allocate...

I bumped the PPA, could you please let me know if it works now?

Which OS do you have in mind? On GNU/Linux, `posix_fallocate` is implemented using `fallocate` (see [posix_fallocate.c](https://github.com/bminor/glibc/blob/9bdf92c79d63b42f931101bb6df87129c408b0c4/sysdeps/unix/sysv/linux/posix_fallocate.c#L29)), so that wouldn't increase compatiblity.

> If you need help extracting the fiemap code into a separate function, I can help with that. Also, it should keep the part to skip extent marked as shared...

It's in the working directory actually, but I think you're right it should be `$TMPDIR`. Care to send a patch?

Oh, only seeing this now. This is a good idea. As you said, it seems that the "most popular" FS have a way of grafting extents from one file to...

Locks in Linux can be "advisory" (user space can ignore them) or "mandatory". Mandatory locks can expire, create responsiveness problems and will go away. Advisory locks are typically ignored by...

(As I go through the backlog, I see this is quite similar to what @kakra proposed in #13 )