Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Restart doesn't work with cargo run that starts a server #16

Closed
robinst opened this issue Oct 20, 2016 · 5 comments
Closed

Restart doesn't work with cargo run that starts a server #16

robinst opened this issue Oct 20, 2016 · 5 comments

Comments

@robinst
Copy link
Contributor

robinst commented Oct 20, 2016

With a simple Iron server project that listens on a port, the restarting doesn't work (on MacOS):

$ watchexec --restart cargo run
// made changes here
   Compiling iron-example v0.1.0 (file:///Users/robinst/Projects/iron-example)
    Finished debug [unoptimized + debuginfo] target(s) in 8.49 secs
     Running `target/debug/iron-example`
// made changes here
   Compiling iron-example v0.1.0 (file:///Users/robinst/Projects/iron-example)
    Finished debug [unoptimized + debuginfo] target(s) in 6.46 secs
     Running `target/debug/iron-example`
thread 'main' panicked at 'Unable to start server: Io(Error { repr: Os { code: 48, message: "Address already in use" } })', ../src/libcore/result.rs:788
note: Run with `RUST_BACKTRACE=1` for a backtrace.
error: Process didn't exit successfully: `target/debug/iron-example` (exit code: 101)
^C
[1]    56690 interrupt  watchexec --restart cargo run

The reason for this is that the restart didn't successfully kill the subprocess that cargo run started, hence "Address already in use".

Note that cargo watch has exactly the same problem: watchexec/cargo-watch#25

The cargo developers recommend using a process group and then killing the process group to make sure all subprocesses are killed: rust-lang/cargo#2818 (comment)
@robinst robinst mentioned this issue Oct 20, 2016
Merged
@mattgreen
Copy link
Collaborator

As of f3c6df8, we use setpgid(), killpg(), and waitpid() to handle launching the child process in a new process group on POSIX systems.

We still need to handle when SIGINT is sent to watchexec by killing the process group.

@robinst
Copy link
Contributor Author

robinst commented Oct 21, 2016

That was quick :). Tried out watchexec 1.1.1, but restarting still runs into the same problem. Was it expected to be fixed with 1.1.1 (not sure from your description)?

@robinst robinst mentioned this issue Oct 21, 2016
Closed
@mattgreen
Copy link
Collaborator

mattgreen commented Oct 21, 2016 •
edited
Loading

Targeting this feature for 1.2. 1.1.1 was a bugfix for 1.1.

Trying to get signal handling in so we can shutdown cleanly and destroy the process group at exit now. Unfortunately, Rust's signal situation is pretty barebones at the moment.

@mattgreen
Copy link
Collaborator

This landed in 1.2.0. Give it a try!

@robinst
Copy link
Contributor Author

robinst commented Oct 26, 2016

Yes, works nicely now :)! 🎉

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
2 participants
@robinst @mattgreen