保存和恢复陷阱状态?管理多个陷阱处理程序的简单方法? [英] Save and restore trap state? Easy way to manage multiple handlers for traps?

问题描述

什么是覆盖 bash 陷阱处理程序的好方法，这些处理程序不会永久践踏可能已经设置或尚未设置的现有处理程序?动态管理任意的陷阱例程链怎么样?

What is a good way to override bash trap handlers that don't permanently trample existing ones that may or may not already be set? What about dynamically managing arbitrary chains of trap routines?

有没有办法保存陷阱处理程序的当前状态，以便以后恢复它们?

Is there a way to save the current state of the trap handlers so they can be restored later?

推荐答案

在 Bash 中保存和恢复 Trap Handler 状态

我将提交以下堆栈实现来跟踪和恢复陷阱状态.使用这种方法，我可以推送陷阱更改，然后在完成后将它们弹出.这也可以用于将许多陷阱例程链接在一起.

Save and Restore Your Trap Handler State in Bash

I would submit the following stack implementation to track and restore trap state. Using this method, I am able to push trap changes and then pop them away when I'm done with them. This could also be used to chain many trap routines together.

查看以下源文件(.trap_stack.sh)

See the following source file (.trap_stack.sh)

#!/bin/bash
trap_stack_name() {
  local sig=${1//[^a-zA-Z0-9]/_}
  echo "__trap_stack_$sig"
}

extract_trap() {
  echo ${@:3:$(($#-3))}
}

get_trap() {
  eval echo $(extract_trap `trap -p $1`)
}

trap_push() {
  local new_trap=$1
  shift
  local sigs=$*
  for sig in $sigs; do
    local stack_name=`trap_stack_name "$sig"`
    local old_trap=$(get_trap $sig)
    eval "${stack_name}"'[${#'"${stack_name}"'[@]}]=$old_trap'
    trap "${new_trap}" "$sig"
  done
}

trap_pop() {
  local sigs=$*
  for sig in $sigs; do
    local stack_name=`trap_stack_name "$sig"`
    local count; eval 'count=${#'"${stack_name}"'[@]}'
    [[ $count -lt 1 ]] && return 127
    local new_trap
    local ref="${stack_name}"'[${#'"${stack_name}"'[@]}-1]'
    local cmd='new_trap=${'"$ref}"; eval $cmd
    trap "${new_trap}" "$sig"
    eval "unset $ref"
  done
}

trap_prepend() {
  local new_trap=$1
  shift
  local sigs=$*
  for sig in $sigs; do
    if [[ -z $(get_trap $sig) ]]; then
      trap_push "$new_trap" "$sig"
    else
      trap_push "$new_trap ; $(get_trap $sig)" "$sig"
    fi
  done
}

trap_append() {
  local new_trap=$1
  shift
  local sigs=$*
  for sig in $sigs; do
    if [[ -z $(get_trap $sig) ]]; then
      trap_push "$new_trap" "$sig"
    else
      trap_push "$(get_trap $sig) ; $new_trap" "$sig"
    fi
  done
}

这可以管理定义为命名函数的处理程序以及像此示例定义的临时例程 trap "kill $!"SIGTERM SIGINT.

This can manage handlers that are defined as named functions and also ad-hoc routines defined like this example trap "kill $!" SIGTERM SIGINT.

这是我用来帮我写的测试脚本:

This is the test script I used to help me write it:

#!/bin/bash
source .trap_stack.sh

initial_trap='echo "messy" ;'" echo 'handler'"
non_f_trap='echo "non-function trap"'
f_trap() {
  echo "function trap"
}

print_status() {
  echo "    SIGINT  trap: `get_trap SIGINT`"  
  echo "    SIGTERM trap: `get_trap SIGTERM`"
  echo "-------------"
  echo
}

echo "--- TEST START ---"
echo "Initial trap state (should be empty):"
print_status

echo 'Setting messy non-function handler for SIGINT ("original state")'
trap "$initial_trap" SIGINT
print_status

echo 'Pop empty stacks (still in original state)'
trap_pop SIGINT SIGTERM
print_status

echo 'Push non-function handler for SIGINT'
trap_push "$non_f_trap" SIGINT
print_status

echo 'Append function handler for SIGINT and SIGTERM'
trap_append f_trap SIGINT SIGTERM
print_status

echo 'Prepend function handler for SIGINT and SIGTERM'
trap_prepend f_trap SIGINT SIGTERM
print_status

echo 'Push non-function handler for SIGINT and SIGTERM'
trap_push "$non_f_trap" SIGINT SIGTERM
print_status

echo 'Pop both stacks'
trap_pop SIGINT SIGTERM
print_status

echo 'Prepend function handler for SIGINT and SIGTERM'
trap_prepend f_trap SIGINT SIGTERM
print_status

echo 'Pop both stacks thrice'
trap_pop SIGINT SIGTERM
trap_pop SIGINT SIGTERM
trap_pop SIGINT SIGTERM
print_status

echo 'Push non-function handler for SIGTERM'
trap_push "$non_f_trap" SIGTERM
print_status

echo 'Pop handler state for SIGINT (SIGINT is now back to original state)'
trap_pop SIGINT
print_status

echo 'Pop handler state for SIGTERM (SIGTERM is now back to original state)'
trap_pop SIGTERM
print_status

这篇关于保存和恢复陷阱状态?管理多个陷阱处理程序的简单方法?的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！