Process memory

[thijsc]

Measurements of current and max memory usage for processes. You can get current for all processes. Max seems to be only available for the current process, at least in Linux.

Fixes #12

[thijsc]

I haven't done this in the coding style of #14 by the way. Will try to convert it and see how I like that :-).

[thijsc]

I think this is the nicest we can get. It's not possible to combine raw_data and segments because segments is a Vector of references to raw_data. Therefore we do need to keep that variable around for the entire scope of the function.

[thijsc]

I tried a slightly different approach, something like this:

file_to_string(path).
    and_then(|raw_data| {
        let segments: Vec<&str> = raw_data.split_whitespace().collect();

        if segments.len() < 2 {
            return Err(ProbeError::UnexpectedContent("Incorrect number of segments".to_owned()))
        }

This isn't very ergonomic because the errors within the block are not of the same type as the IO error from file_to_string. We'd have to add try!s around them. I think I now understand why you added these manual error conversions in #14 @timonv.

Overall I feel that a more functional style in the implementation of these functions causes more hassle than the benefit it delivers.

[timonv]

It's a balancing act 🐵

[timonv]

Ok, but why is it unsafe? Or is just whole libc unsafe?

[thijsc]

Using libc is almost always unsafe since its a C library that doesn't offer the guarantees Rust needs. Therefore we have to check things manually and make sure they are correct.

[thijsc]

The compiler will emit a warning if you use an unnecessary unsafe block by the way, so we'd know about it if this was unnecessary.

[thijsc]

Yeah we'll plan that evening after its done, didn't mean to imply that I wanted to impose deadlines :-).

[thijsc]

Any objections to merging this?

[timonv]

None at all! +1