ruby/test/test_timeout.rb

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

272 строки
6.3 KiB
Ruby
Исходник Обычный вид История

# frozen_string_literal: false
require 'test/unit'
require 'timeout'
class TestTimeout < Test::Unit::TestCase
def test_work_is_done_in_same_thread_as_caller
assert_equal Thread.current, Timeout.timeout(10){ Thread.current }
end
def test_work_is_done_in_same_fiber_as_caller
require 'fiber' # needed for ruby 3.0 and lower
assert_equal Fiber.current, Timeout.timeout(10){ Fiber.current }
end
def test_non_timing_out_code_is_successful
assert_nothing_raised do
assert_equal :ok, Timeout.timeout(1){ :ok }
end
end
def test_allows_zero_seconds
assert_nothing_raised do
assert_equal :ok, Timeout.timeout(0){:ok}
end
end
def test_allows_nil_seconds
assert_nothing_raised do
assert_equal :ok, Timeout.timeout(nil){:ok}
end
end
def test_included
c = Class.new do
include Timeout
def test
timeout(1) { :ok }
end
end
assert_nothing_raised do
assert_equal :ok, c.new.test
end
end
def test_yield_param
assert_equal [5, :ok], Timeout.timeout(5){|s| [s, :ok] }
end
def test_queue
q = Thread::Queue.new
assert_raise(Timeout::Error, "[ruby-dev:32935]") {
Timeout.timeout(0.01) { q.pop }
}
end
def test_timeout
assert_raise(Timeout::Error) do
Timeout.timeout(0.1) {
nil while true
}
end
end
[ruby/timeout] Raise exception instead of throw/catch for timeouts (https://github.com/ruby/timeout/pull/30) throw/catch is used for non-local control flow, not for exceptional situations. For exceptional situations, raise should be used instead. A timeout is an exceptional situation, so it should use raise, not throw/catch. Timeout's implementation that uses throw/catch internally causes serious problems. Consider the following code: ```ruby def handle_exceptions yield rescue Exception => exc handle_error # e.g. ROLLBACK for databases raise ensure handle_exit unless exc # e.g. COMMIT for databases end Timeout.timeout(1) do handle_exceptions do do_something end end ``` This kind of design ensures that all exceptions are handled as errors, and ensures that all exits (normal exit, early return, throw/catch) are not handled as errors. With Timeout's throw/catch implementation, this type of code does not work, since a timeout triggers the normal exit path. See https://github.com/rails/rails/pull/29333 for an example of the damage Timeout's design has caused the Rails ecosystem. This switches Timeout.timeout to use raise/rescue internally. It adds a Timeout::ExitException subclass of exception for the internal raise/rescue, which Timeout.timeout will convert to Timeout::Error for backwards compatibility. Timeout::Error remains a subclass of RuntimeError. This is how timeout used to work in Ruby 2.0. It was changed in Ruby 2.1, after discussion in [Bug #8730] (commit https://github.com/ruby/timeout/commit/238c003c921e in the timeout repository). I think the change from using raise/rescue to using throw/catch has caused significant harm to the Ruby ecosystem at large, and reverting it is the most sensible choice. From the translation of [Bug #8730], it appears the issue was that someone could rescue Exception and not reraise the exception, causing timeout errors to be swallowed. However, such code is broken anyway. Using throw/catch causes far worse problems, because then it becomes impossible to differentiate between normal control flow and exceptional control flow. Also related to this is [Bug #11344], which changed how Thread.handle_interrupt interacted with Timeout. https://github.com/ruby/timeout/commit/f16545abe6 Co-authored-by: Nobuyoshi Nakada <nobu@ruby-lang.org>
2023-06-22 21:24:40 +03:00
def test_nested_timeout
a = nil
assert_raise(Timeout::Error) do
Timeout.timeout(0.1) {
Timeout.timeout(1) {
nil while true
}
a = 1
}
end
assert_nil a
end
class MyNewErrorOuter < StandardError; end
class MyNewErrorInner < StandardError; end
# DOES NOT fail with
# - raise new(message) if exc.equal?(e)
# + raise new(message) if exc.class == e.class
def test_nested_timeout_error_identity
begin
Timeout.timeout(0.1, MyNewErrorOuter) {
Timeout.timeout(1, MyNewErrorInner) {
nil while true
}
}
rescue => e
assert e.class == MyNewErrorOuter
end
end
# DOES fail with
# - raise new(message) if exc.equal?(e)
# + raise new(message) if exc.class == e.class
def test_nested_timeout_which_error_bubbles_up
raised_exception = nil
begin
Timeout.timeout(0.1) {
Timeout.timeout(1) {
raise Timeout::ExitException.new("inner message")
}
}
rescue Exception => e
raised_exception = e
end
assert_equal 'inner message', raised_exception.message
end
def test_cannot_convert_into_time_interval
bug3168 = '[ruby-dev:41010]'
def (n = Object.new).zero?; false; end
assert_raise(TypeError, bug3168) {Timeout.timeout(n) { sleep 0.1 }}
end
[ruby/timeout] Raise exception instead of throw/catch for timeouts (https://github.com/ruby/timeout/pull/30) throw/catch is used for non-local control flow, not for exceptional situations. For exceptional situations, raise should be used instead. A timeout is an exceptional situation, so it should use raise, not throw/catch. Timeout's implementation that uses throw/catch internally causes serious problems. Consider the following code: ```ruby def handle_exceptions yield rescue Exception => exc handle_error # e.g. ROLLBACK for databases raise ensure handle_exit unless exc # e.g. COMMIT for databases end Timeout.timeout(1) do handle_exceptions do do_something end end ``` This kind of design ensures that all exceptions are handled as errors, and ensures that all exits (normal exit, early return, throw/catch) are not handled as errors. With Timeout's throw/catch implementation, this type of code does not work, since a timeout triggers the normal exit path. See https://github.com/rails/rails/pull/29333 for an example of the damage Timeout's design has caused the Rails ecosystem. This switches Timeout.timeout to use raise/rescue internally. It adds a Timeout::ExitException subclass of exception for the internal raise/rescue, which Timeout.timeout will convert to Timeout::Error for backwards compatibility. Timeout::Error remains a subclass of RuntimeError. This is how timeout used to work in Ruby 2.0. It was changed in Ruby 2.1, after discussion in [Bug #8730] (commit https://github.com/ruby/timeout/commit/238c003c921e in the timeout repository). I think the change from using raise/rescue to using throw/catch has caused significant harm to the Ruby ecosystem at large, and reverting it is the most sensible choice. From the translation of [Bug #8730], it appears the issue was that someone could rescue Exception and not reraise the exception, causing timeout errors to be swallowed. However, such code is broken anyway. Using throw/catch causes far worse problems, because then it becomes impossible to differentiate between normal control flow and exceptional control flow. Also related to this is [Bug #11344], which changed how Thread.handle_interrupt interacted with Timeout. https://github.com/ruby/timeout/commit/f16545abe6 Co-authored-by: Nobuyoshi Nakada <nobu@ruby-lang.org>
2023-06-22 21:24:40 +03:00
def test_skip_rescue_standarderror
e = nil
[ruby/timeout] Raise exception instead of throw/catch for timeouts (https://github.com/ruby/timeout/pull/30) throw/catch is used for non-local control flow, not for exceptional situations. For exceptional situations, raise should be used instead. A timeout is an exceptional situation, so it should use raise, not throw/catch. Timeout's implementation that uses throw/catch internally causes serious problems. Consider the following code: ```ruby def handle_exceptions yield rescue Exception => exc handle_error # e.g. ROLLBACK for databases raise ensure handle_exit unless exc # e.g. COMMIT for databases end Timeout.timeout(1) do handle_exceptions do do_something end end ``` This kind of design ensures that all exceptions are handled as errors, and ensures that all exits (normal exit, early return, throw/catch) are not handled as errors. With Timeout's throw/catch implementation, this type of code does not work, since a timeout triggers the normal exit path. See https://github.com/rails/rails/pull/29333 for an example of the damage Timeout's design has caused the Rails ecosystem. This switches Timeout.timeout to use raise/rescue internally. It adds a Timeout::ExitException subclass of exception for the internal raise/rescue, which Timeout.timeout will convert to Timeout::Error for backwards compatibility. Timeout::Error remains a subclass of RuntimeError. This is how timeout used to work in Ruby 2.0. It was changed in Ruby 2.1, after discussion in [Bug #8730] (commit https://github.com/ruby/timeout/commit/238c003c921e in the timeout repository). I think the change from using raise/rescue to using throw/catch has caused significant harm to the Ruby ecosystem at large, and reverting it is the most sensible choice. From the translation of [Bug #8730], it appears the issue was that someone could rescue Exception and not reraise the exception, causing timeout errors to be swallowed. However, such code is broken anyway. Using throw/catch causes far worse problems, because then it becomes impossible to differentiate between normal control flow and exceptional control flow. Also related to this is [Bug #11344], which changed how Thread.handle_interrupt interacted with Timeout. https://github.com/ruby/timeout/commit/f16545abe6 Co-authored-by: Nobuyoshi Nakada <nobu@ruby-lang.org>
2023-06-22 21:24:40 +03:00
assert_raise_with_message(Timeout::Error, /execution expired/) do
Timeout.timeout 0.01 do
begin
sleep 3
[ruby/timeout] Raise exception instead of throw/catch for timeouts (https://github.com/ruby/timeout/pull/30) throw/catch is used for non-local control flow, not for exceptional situations. For exceptional situations, raise should be used instead. A timeout is an exceptional situation, so it should use raise, not throw/catch. Timeout's implementation that uses throw/catch internally causes serious problems. Consider the following code: ```ruby def handle_exceptions yield rescue Exception => exc handle_error # e.g. ROLLBACK for databases raise ensure handle_exit unless exc # e.g. COMMIT for databases end Timeout.timeout(1) do handle_exceptions do do_something end end ``` This kind of design ensures that all exceptions are handled as errors, and ensures that all exits (normal exit, early return, throw/catch) are not handled as errors. With Timeout's throw/catch implementation, this type of code does not work, since a timeout triggers the normal exit path. See https://github.com/rails/rails/pull/29333 for an example of the damage Timeout's design has caused the Rails ecosystem. This switches Timeout.timeout to use raise/rescue internally. It adds a Timeout::ExitException subclass of exception for the internal raise/rescue, which Timeout.timeout will convert to Timeout::Error for backwards compatibility. Timeout::Error remains a subclass of RuntimeError. This is how timeout used to work in Ruby 2.0. It was changed in Ruby 2.1, after discussion in [Bug #8730] (commit https://github.com/ruby/timeout/commit/238c003c921e in the timeout repository). I think the change from using raise/rescue to using throw/catch has caused significant harm to the Ruby ecosystem at large, and reverting it is the most sensible choice. From the translation of [Bug #8730], it appears the issue was that someone could rescue Exception and not reraise the exception, causing timeout errors to be swallowed. However, such code is broken anyway. Using throw/catch causes far worse problems, because then it becomes impossible to differentiate between normal control flow and exceptional control flow. Also related to this is [Bug #11344], which changed how Thread.handle_interrupt interacted with Timeout. https://github.com/ruby/timeout/commit/f16545abe6 Co-authored-by: Nobuyoshi Nakada <nobu@ruby-lang.org>
2023-06-22 21:24:40 +03:00
rescue => e
flunk "should not see any exception but saw #{e.inspect}"
end
end
end
[ruby/timeout] Raise exception instead of throw/catch for timeouts (https://github.com/ruby/timeout/pull/30) throw/catch is used for non-local control flow, not for exceptional situations. For exceptional situations, raise should be used instead. A timeout is an exceptional situation, so it should use raise, not throw/catch. Timeout's implementation that uses throw/catch internally causes serious problems. Consider the following code: ```ruby def handle_exceptions yield rescue Exception => exc handle_error # e.g. ROLLBACK for databases raise ensure handle_exit unless exc # e.g. COMMIT for databases end Timeout.timeout(1) do handle_exceptions do do_something end end ``` This kind of design ensures that all exceptions are handled as errors, and ensures that all exits (normal exit, early return, throw/catch) are not handled as errors. With Timeout's throw/catch implementation, this type of code does not work, since a timeout triggers the normal exit path. See https://github.com/rails/rails/pull/29333 for an example of the damage Timeout's design has caused the Rails ecosystem. This switches Timeout.timeout to use raise/rescue internally. It adds a Timeout::ExitException subclass of exception for the internal raise/rescue, which Timeout.timeout will convert to Timeout::Error for backwards compatibility. Timeout::Error remains a subclass of RuntimeError. This is how timeout used to work in Ruby 2.0. It was changed in Ruby 2.1, after discussion in [Bug #8730] (commit https://github.com/ruby/timeout/commit/238c003c921e in the timeout repository). I think the change from using raise/rescue to using throw/catch has caused significant harm to the Ruby ecosystem at large, and reverting it is the most sensible choice. From the translation of [Bug #8730], it appears the issue was that someone could rescue Exception and not reraise the exception, causing timeout errors to be swallowed. However, such code is broken anyway. Using throw/catch causes far worse problems, because then it becomes impossible to differentiate between normal control flow and exceptional control flow. Also related to this is [Bug #11344], which changed how Thread.handle_interrupt interacted with Timeout. https://github.com/ruby/timeout/commit/f16545abe6 Co-authored-by: Nobuyoshi Nakada <nobu@ruby-lang.org>
2023-06-22 21:24:40 +03:00
end
def test_raises_exception_internally
e = nil
assert_raise_with_message(Timeout::Error, /execution expired/) do
Timeout.timeout 0.01 do
begin
sleep 3
rescue Exception => exc
e = exc
raise
end
end
end
assert_equal Timeout::ExitException, e.class
end
def test_rescue_exit
exc = Class.new(RuntimeError)
e = nil
assert_nothing_raised(exc) do
Timeout.timeout 0.01, exc do
begin
sleep 3
rescue exc => e
end
end
end
assert_raise_with_message(exc, 'execution expired') {raise e if e}
end
def test_custom_exception
bug9354 = '[ruby-core:59511] [Bug #9354]'
err = Class.new(StandardError) do
def initialize(msg) super end
end
assert_nothing_raised(ArgumentError, bug9354) do
assert_equal(:ok, Timeout.timeout(100, err) {:ok})
end
assert_raise_with_message(err, 'execution expired') do
Timeout.timeout 0.01, err do
sleep 3
end
end
assert_raise_with_message(err, /connection to ruby-lang.org expired/) do
Timeout.timeout 0.01, err, "connection to ruby-lang.org expired" do
sleep 3
end
end
end
def test_exit_exception
assert_raise_with_message(Timeout::Error, "boon") do
Timeout.timeout(10, Timeout::Error) do
raise Timeout::Error, "boon"
end
end
end
def test_raise_with_message
bug17812 = '[ruby-core:103502] [Bug #17812]: Timeout::Error doesn\'t let two-argument raise() set a new message'
exc = Timeout::Error.new('foo')
assert_raise_with_message(Timeout::Error, 'bar', bug17812) do
raise exc, 'bar'
end
end
def test_enumerator_next
bug9380 = '[ruby-dev:47872] [Bug #9380]: timeout in Enumerator#next'
e = (o=Object.new).to_enum
def o.each
sleep
end
assert_raise_with_message(Timeout::Error, 'execution expired', bug9380) do
Timeout.timeout(0.01) {e.next}
end
end
def test_handle_interrupt
bug11344 = '[ruby-dev:49179] [Bug #11344]'
ok = false
assert_raise(Timeout::Error) {
[ruby/timeout] Raise exception instead of throw/catch for timeouts (https://github.com/ruby/timeout/pull/30) throw/catch is used for non-local control flow, not for exceptional situations. For exceptional situations, raise should be used instead. A timeout is an exceptional situation, so it should use raise, not throw/catch. Timeout's implementation that uses throw/catch internally causes serious problems. Consider the following code: ```ruby def handle_exceptions yield rescue Exception => exc handle_error # e.g. ROLLBACK for databases raise ensure handle_exit unless exc # e.g. COMMIT for databases end Timeout.timeout(1) do handle_exceptions do do_something end end ``` This kind of design ensures that all exceptions are handled as errors, and ensures that all exits (normal exit, early return, throw/catch) are not handled as errors. With Timeout's throw/catch implementation, this type of code does not work, since a timeout triggers the normal exit path. See https://github.com/rails/rails/pull/29333 for an example of the damage Timeout's design has caused the Rails ecosystem. This switches Timeout.timeout to use raise/rescue internally. It adds a Timeout::ExitException subclass of exception for the internal raise/rescue, which Timeout.timeout will convert to Timeout::Error for backwards compatibility. Timeout::Error remains a subclass of RuntimeError. This is how timeout used to work in Ruby 2.0. It was changed in Ruby 2.1, after discussion in [Bug #8730] (commit https://github.com/ruby/timeout/commit/238c003c921e in the timeout repository). I think the change from using raise/rescue to using throw/catch has caused significant harm to the Ruby ecosystem at large, and reverting it is the most sensible choice. From the translation of [Bug #8730], it appears the issue was that someone could rescue Exception and not reraise the exception, causing timeout errors to be swallowed. However, such code is broken anyway. Using throw/catch causes far worse problems, because then it becomes impossible to differentiate between normal control flow and exceptional control flow. Also related to this is [Bug #11344], which changed how Thread.handle_interrupt interacted with Timeout. https://github.com/ruby/timeout/commit/f16545abe6 Co-authored-by: Nobuyoshi Nakada <nobu@ruby-lang.org>
2023-06-22 21:24:40 +03:00
Thread.handle_interrupt(Timeout::ExitException => :never) {
Timeout.timeout(0.01) {
sleep 0.2
ok = true
[ruby/timeout] Raise exception instead of throw/catch for timeouts (https://github.com/ruby/timeout/pull/30) throw/catch is used for non-local control flow, not for exceptional situations. For exceptional situations, raise should be used instead. A timeout is an exceptional situation, so it should use raise, not throw/catch. Timeout's implementation that uses throw/catch internally causes serious problems. Consider the following code: ```ruby def handle_exceptions yield rescue Exception => exc handle_error # e.g. ROLLBACK for databases raise ensure handle_exit unless exc # e.g. COMMIT for databases end Timeout.timeout(1) do handle_exceptions do do_something end end ``` This kind of design ensures that all exceptions are handled as errors, and ensures that all exits (normal exit, early return, throw/catch) are not handled as errors. With Timeout's throw/catch implementation, this type of code does not work, since a timeout triggers the normal exit path. See https://github.com/rails/rails/pull/29333 for an example of the damage Timeout's design has caused the Rails ecosystem. This switches Timeout.timeout to use raise/rescue internally. It adds a Timeout::ExitException subclass of exception for the internal raise/rescue, which Timeout.timeout will convert to Timeout::Error for backwards compatibility. Timeout::Error remains a subclass of RuntimeError. This is how timeout used to work in Ruby 2.0. It was changed in Ruby 2.1, after discussion in [Bug #8730] (commit https://github.com/ruby/timeout/commit/238c003c921e in the timeout repository). I think the change from using raise/rescue to using throw/catch has caused significant harm to the Ruby ecosystem at large, and reverting it is the most sensible choice. From the translation of [Bug #8730], it appears the issue was that someone could rescue Exception and not reraise the exception, causing timeout errors to be swallowed. However, such code is broken anyway. Using throw/catch causes far worse problems, because then it becomes impossible to differentiate between normal control flow and exceptional control flow. Also related to this is [Bug #11344], which changed how Thread.handle_interrupt interacted with Timeout. https://github.com/ruby/timeout/commit/f16545abe6 Co-authored-by: Nobuyoshi Nakada <nobu@ruby-lang.org>
2023-06-22 21:24:40 +03:00
Thread.handle_interrupt(Timeout::ExitException => :on_blocking) {
sleep 0.2
}
}
}
}
assert(ok, bug11344)
end
def test_fork
omit 'fork not supported' unless Process.respond_to?(:fork)
r, w = IO.pipe
pid = fork do
r.close
begin
r = Timeout.timeout(0.01) { sleep 5 }
w.write r.inspect
rescue Timeout::Error
w.write 'timeout'
ensure
w.close
end
end
w.close
Process.wait pid
assert_equal 'timeout', r.read
r.close
end
def test_threadgroup
assert_separately(%w[-rtimeout], <<-'end;')
tg = ThreadGroup.new
thr = Thread.new do
tg.add(Thread.current)
Timeout.timeout(10){}
end
thr.join
assert_equal [].to_s, tg.list.to_s
end;
end
# https://github.com/ruby/timeout/issues/24
def test_handling_enclosed_threadgroup
assert_separately(%w[-rtimeout], <<-'end;')
Thread.new {
t = Thread.current
group = ThreadGroup.new
group.add(t)
group.enclose
assert_equal 42, Timeout.timeout(1) { 42 }
}.join
end;
end
end