notify VS perf-tools

To support hot reloading, the asset manager definitely has to have some watcher thread, but I have that part dealt with already since I have decided to utilize notify. The other half is to find a way to reload all assets that have changes as determined by the watcher thread. What I am not sure about is how to implement this in a Rustic fashion and with clean architecture while also considering the fact that asset pointers may be held by other objects (such as an object representing an OpenGL program object).

It is a great question. We have based the development on a library called notify, kudos to https://github.com/notify-rs/notify. This library adds a layer of abstraction to each system. It implements kernel-specific hooks as you mentioned. In some cases like Audit extended data, we have developed an integration that detects changes on the Audit log file and processes the given information including a lot of information into Linux systems. We have plans to include it in Windows as well.

In order to detect files changes, we use hotwatch, a simple wrapper over notify that will allow us to save a few lines.

The notify crate uses the inotify API on Linux. However, it's probably not what you want, since one of its limitations is:

If you want to build this yourself, you'll want to build on something like notify - there are libraries like linemux built on top of it that will do a lot of this for you too.

Hey rust community, I'm currently writing a small app using Tokio, but running into a problem when trying to add a file watcher to it. I looked into available crates and found https://github.com/notify-rs/notify to be a good option, though it only exposes a sync interface.

I'm most well versed in Python so I started with that and eventually ended up with a POC using watchdog, but the program took more than a day and a half to traverse everything and register all of the watches. I've been trying to learn Rust for a while, and decided to perhaps use this as an excuse to try something 'real' and not just a learning exercise. I found the notify crate (https://github.com/notify-rs/notify) and basically copied their example listed on their GitHub, but even this takes about 16-18 hours to place all the watches before it starts processing events. I did not see any obvious ways to enable asynchronous or async/await code in notify, so I don't know of a way to parallelize the disk I/O with this approach.

The notify crate is probably what you're looking for. It uses inotify on linux by default.

For execsnoop, people running systems with DTrace can find the same:

* https://github.com/jorgev/dtrace-scripts/blob/master/execsno...

On macOS Monterey+ you'll probably have to install the Kernel Debug Kit per:

* https://developer.apple.com/forums/thread/692444

The Linux variant was written Brendan Gregg (who previous did a lot of work on Solaris, where DTrace was created):

* https://github.com/brendangregg/perf-tools/blob/master/execs...

by https://github.com/brendangregg/perf-tools

It's not a library you can use but you might be able to take inspiration from something like opensnoop.

There are several tools that might go by the name 'perf', are you talking about this one? https://github.com/brendangregg/perf-tools

I have an application and I'd like to measure its page cache hit/miss numbers. I found this tool called cachestat by Brendan Gregg, which uses the ftrace tool to count the number of 4 specific functions (mark_page_accessed(), mark_buffer_dirty(), add_to_page_cache_lru(), and account_page_dirtied()) to find out the total number of page cache accesses and page cache misses. Detailed information can be found here, but basically, total number of page cache accesses and page cache misses can be calculated as below:

Program(Process) Tracing (ptrace) Tools: Establishes tracing operation during the execution of the application. Contains the traces of the index of instructions executed and the data referenced during execution. These are greatly used by developers for debugging purposes. Some examples of ptrace tools are, Strace, Ltrace, Opensnoop, and Valgrind Lackey.