Dan Burkert

Yep, @BurntSushi is correct. I wouldn't expect something like a system package manager updating a package to be an issue, since I would expect it to write a new file...

One thing that worries me is that we're changing the API to have a safe deref to `&[u8]` on `Mmap`, but make it unsafe to create a `Mmap` from a...

@iqualfragile what specifically do you mean by un-deleted?

@iqualfragile what specifically do you mean by re-add it to the file system? I'm not aware of an API to do that (even OS/FS dependent), so I'm intrigued. If there's...

Right, my take in the past has been that `madvise` should probably be provided by a separate crate. The non-cross-platform aspects of it make it a tough API to wrap....

I expected that the fix to this would be to use `mmap64` instead of `mmap` (on select platforms), or is that infeasible?

Hi @tinco, AFAIK this isn't something that can be done reliably across platform, and as such it isn't something I want to support in this crate. In particular the behavior...

OK this looks good, but please add some tests, with cases covering: |offset|len|outcome|notes| |---|---|---|---| |`0`|`isize::MAX`|fail|| |`0`|`isize::MAX - 1`|pass|| |`x`|`isize::MAX - x`|fail|`x > 0 && x < PAGE_SIZE`| |`x`|`isize::MAX - x...

The test can look like the one in https://github.com/danburkert/memmap-rs/pull/74, but with the parameters I mentioned here to test edge cases.

I'm open to this, but only if rustfmt can be configured to be fairly close to the existing style. Last time I checked out rustfmt it had a bunch of...