Rollup merge of #38518 - nagisa:exec-doc, r=alexcrichton

Expand documentation of process::exit and exec

Show a conventional way to use process::exit when destructors are considered important and also
mention that the same caveats wrt destructors apply to exec as well.

1 files changed, 10 insertions, 0 deletions

diff --git a/src/libstd/sys/unix/ext/process.rs b/src/libstd/sys/unix/ext/process.rs
index 585dcbb9a34..2961c4ec582 100644
--- a/src/libstd/sys/unix/ext/process.rs
+++ b/src/libstd/sys/unix/ext/process.rs
@@ -67,10 +67,20 @@ pub trait CommandExt {
     /// an error indicating why the exec (or another part of the setup of the
     /// `Command`) failed.
     ///
+    /// `exec` not returning has the same implications as calling
+    /// [`process::exit`] – no destructors on the current stack or any other
+    /// thread’s stack will be run. Therefore, it is recommended to only call
+    /// `exec` at a point where it is fine to not run any destructors. Note,
+    /// that the `execvp` syscall independently guarantees that all memory is
+    /// freed and all file descriptors with the `CLOEXEC` option (set by default
+    /// on all file descriptors opened by the standard library) are closed.
+    ///
     /// This function, unlike `spawn`, will **not** `fork` the process to create
     /// a new child. Like spawn, however, the default behavior for the stdio
     /// descriptors will be to inherited from the current process.
     ///
+    /// [`process::exit`]: ../../../process/fn.exit.html
+    ///
     /// # Notes
     ///
     /// The process may be in a "broken state" if this function returns in