class Ameba::Rule::Lint::TrailingRescueException
- Ameba::Rule::Lint::TrailingRescueException
- Ameba::Rule::Base
- Reference
- Object
Overview
A rule that prohibits the misconception about how trailing rescue
statements work,
preventing Paths (exception class names or otherwise) from being used as the
trailing value. The value after the trailing rescue
statement is the value
to use if an exception occurs, not the exception class to rescue from.
For example, this is considered invalid - if an exception occurs,
response
will be assigned with the value of IO::Error
instead of nil
:
response = HTTP::Client.get("http://www.example.com") rescue IO::Error
And should instead be written as this in order to capture only IO::Error
exceptions:
response = begin
HTTP::Client.get("http://www.example.com")
rescue IO::Error
"default value"
end
Or to rescue all exceptions (instead of just IO::Error
):
response = HTTP::Client.get("http://www.example.com") rescue "default value"
YAML configuration example:
Lint/TrailingRescueException:
Enabled: true
Included Modules
- YAML::Serializable
- YAML::Serializable::Strict
Defined in:
ameba/rule/lint/trailing_rescue_exception.crConstant Summary
-
MSG =
"Use a block variant of `rescue` to filter by the exception type"
Constructors
- .new(ctx : YAML::ParseContext, node : YAML::Nodes::Node)
-
.new(config = nil)
A rule that prohibits the misconception about how trailing
rescue
statements work, preventing Paths (exception class names or otherwise) from being used as the trailing value.
Class Method Summary
-
.parsed_doc : String | Nil
Returns documentation for this rule, if there is any.
Instance Method Summary
- #description : String
- #description=(description : String)
- #enabled=(enabled : Bool)
- #enabled? : Bool
- #excluded : Array(String) | Nil
- #excluded=(excluded : Array(String) | Nil)
- #severity : Ameba::Severity
- #severity=(severity : Ameba::Severity)
- #since_version : SemanticVersion | Nil
- #since_version=(since_version : String)
- #test(source, node : Crystal::ExceptionHandler)
Instance methods inherited from class Ameba::Rule::Base
==(other)
==,
catch(source : Source)
catch,
excluded?(source)
excluded?,
group
group,
hash
hash,
name
name,
special?
special?,
test(source : Source, node : Crystal::ASTNode, *opts)test(source : Source) test
Class methods inherited from class Ameba::Rule::Base
default_severity : Ameba::Severity
default_severity
Macros inherited from class Ameba::Rule::Base
issue_for(*args, **kwargs, &block)
issue_for
Macros inherited from module Ameba::Config::RuleConfig
properties(&block)
properties
Constructor Detail
A rule that prohibits the misconception about how trailing rescue
statements work,
preventing Paths (exception class names or otherwise) from being used as the
trailing value. The value after the trailing rescue
statement is the value
to use if an exception occurs, not the exception class to rescue from.
For example, this is considered invalid - if an exception occurs,
response
will be assigned with the value of IO::Error
instead of nil
:
response = HTTP::Client.get("http://www.example.com") rescue IO::Error
And should instead be written as this in order to capture only IO::Error
exceptions:
response = begin
HTTP::Client.get("http://www.example.com")
rescue IO::Error
"default value"
end
Or to rescue all exceptions (instead of just IO::Error
):
response = HTTP::Client.get("http://www.example.com") rescue "default value"
YAML configuration example:
Lint/TrailingRescueException:
Enabled: true
Class Method Detail
Returns documentation for this rule, if there is any.
module Ameba
# This is a test rule.
# Does nothing.
class MyRule < Ameba::Rule::Base
def test(source)
end
end
end
MyRule.parsed_doc # => "This is a test rule.\nDoes nothing."