Pipelines

Connect streams and redirect output

A pipeline sends the standard output of one stage to the standard input of the next. && runs a following pipeline only after success; || runs it after failure. > replaces a file and >> appends to it.

import shell
let check = fn () -> i64 {
    let code = run printf "alpha\nbeta\n" | grep beta
    if code != 0 { return code }
    return 0
}
assert check() == 0

Exit status belongs to the whole pipeline

The invocation records the first failing stage, rather than merely the last program. A successful consumer does not hide a producer failure. Test the returned status before treating output as a successful result.

import shell
let verify_pipeline = fn () -> i64 {
    let code = run false | cat
    if code == 0 { return 1 }
    return 0
}
assert verify_pipeline() == 0

Use Shell:run_pipeline when constructing stages dynamically. The caller still owns that pipeline after execution. The special result 126 indicates that the pipeline could not be started; a nonzero program exit is also failure and should be handled by the caller.