# frozen_string_literal: true # :markup: markdown # irb.rb - irb main module # by Keiju ISHITSUKA(keiju@ruby-lang.org) # require "ripper" require "reline" require_relative "irb/init" require_relative "irb/context" require_relative "irb/default_commands" require_relative "irb/ruby-lex" require_relative "irb/statement" require_relative "irb/history" require_relative "irb/input-method" require_relative "irb/locale" require_relative "irb/color" require_relative "irb/version" require_relative "irb/easter-egg" require_relative "irb/debug" require_relative "irb/pager" # ## IRB # # Module IRB ("Interactive Ruby") provides a shell-like interface that supports # user interaction with the Ruby interpreter. # # It operates as a *read-eval-print loop* # ([REPL](https://en.wikipedia.org/wiki/Read%E2%80%93eval%E2%80%93print_loop)) # that: # # * ***Reads*** each character as you type. You can modify the IRB context to # change the way input works. See [Input](rdoc-ref:IRB@Input). # * ***Evaluates*** the code each time it has read a syntactically complete # passage. # * ***Prints*** after evaluating. You can modify the IRB context to change # the way output works. See [Output](rdoc-ref:IRB@Output). # # # Example: # # $ irb # irb(main):001> File.basename(Dir.pwd) # => "irb" # irb(main):002> Dir.entries('.').size # => 25 # irb(main):003* Dir.entries('.').select do |entry| # irb(main):004* entry.start_with?('R') # irb(main):005> end # => ["README.md", "Rakefile"] # # The typed input may also include [\IRB-specific # commands](rdoc-ref:IRB@IRB-Specific+Commands). # # As seen above, you can start IRB by using the shell command `irb`. # # You can stop an IRB session by typing command `exit`: # # irb(main):006> exit # $ # # At that point, IRB calls any hooks found in array `IRB.conf[:AT_EXIT]`, then # exits. # # ## Startup # # At startup, IRB: # # 1. Interprets (as Ruby code) the content of the [configuration # file](rdoc-ref:IRB@Configuration+File) (if given). # 2. Constructs the initial session context from [hash # IRB.conf](rdoc-ref:IRB@Hash+IRB.conf) and from default values; the hash # content may have been affected by [command-line # options](rdoc-ref:IB@Command-Line+Options), and by direct assignments in # the configuration file. # 3. Assigns the context to variable `conf`. # 4. Assigns command-line arguments to variable `ARGV`. # 5. Prints the [prompt](rdoc-ref:IRB@Prompt+and+Return+Formats). # 6. Puts the content of the [initialization # script](rdoc-ref:IRB@Initialization+Script) onto the IRB shell, just as if # it were user-typed commands. # # # ### The Command Line # # On the command line, all options precede all arguments; the first item that is # not recognized as an option is treated as an argument, as are all items that # follow. # # #### Command-Line Options # # Many command-line options affect entries in hash `IRB.conf`, which in turn # affect the initial configuration of the IRB session. # # Details of the options are described in the relevant subsections below. # # A cursory list of the IRB command-line options may be seen in the [help # message](https://raw.githubusercontent.com/ruby/irb/master/lib/irb/lc/help-message), # which is also displayed if you use command-line option `--help`. # # If you are interested in a specific option, consult the # [index](rdoc-ref:doc/irb/indexes.md@Index+of+Command-Line+Options). # # #### Command-Line Arguments # # Command-line arguments are passed to IRB in array `ARGV`: # # $ irb --noscript Foo Bar Baz # irb(main):001> ARGV # => ["Foo", "Bar", "Baz"] # irb(main):002> exit # $ # # Command-line option `--` causes everything that follows to be treated as # arguments, even those that look like options: # # $ irb --noscript -- --noscript -- Foo Bar Baz # irb(main):001> ARGV # => ["--noscript", "--", "Foo", "Bar", "Baz"] # irb(main):002> exit # $ # # ### Configuration File # # You can initialize IRB via a *configuration file*. # # If command-line option `-f` is given, no configuration file is looked for. # # Otherwise, IRB reads and interprets a configuration file if one is available. # # The configuration file can contain any Ruby code, and can usefully include # user code that: # # * Can then be debugged in IRB. # * Configures IRB itself. # * Requires or loads files. # # # The path to the configuration file is the first found among: # # * The value of variable `$IRBRC`, if defined. # * The value of variable `$XDG_CONFIG_HOME/irb/irbrc`, if defined. # * File `$HOME/.irbrc`, if it exists. # * File `$HOME/.config/irb/irbrc`, if it exists. # * File `.irbrc` in the current directory, if it exists. # * File `irb.rc` in the current directory, if it exists. # * File `_irbrc` in the current directory, if it exists. # * File `$irbrc` in the current directory, if it exists. # # # If the search fails, there is no configuration file. # # If the search succeeds, the configuration file is read as Ruby code, and so # can contain any Ruby programming you like. # # Method `conf.rc?` returns `true` if a configuration file was read, `false` # otherwise. Hash entry `IRB.conf[:RC]` also contains that value. # # ### Hash `IRB.conf` # # The initial entries in hash `IRB.conf` are determined by: # # * Default values. # * Command-line options, which may override defaults. # * Direct assignments in the configuration file. # # # You can see the hash by typing `IRB.conf`. # # Details of the entries' meanings are described in the relevant subsections # below. # # If you are interested in a specific entry, consult the # [index](rdoc-ref:doc/irb/indexes.md@Index+of+IRB.conf+Entries). # # ### Notes on Initialization Precedence # # * Any conflict between an entry in hash `IRB.conf` and a command-line option # is resolved in favor of the hash entry. # * Hash `IRB.conf` affects the context only once, when the configuration file # is interpreted; any subsequent changes to it do not affect the context and # are therefore essentially meaningless. # # # ### Initialization Script # # By default, the first command-line argument (after any options) is the path to # a Ruby initialization script. # # IRB reads the initialization script and puts its content onto the IRB shell, # just as if it were user-typed commands. # # Command-line option `--noscript` causes the first command-line argument to be # treated as an ordinary argument (instead of an initialization script); # `--script` is the default. # # ## Input # # This section describes the features that allow you to change the way IRB input # works; see also [Input and Output](rdoc-ref:IRB@Input+and+Output). # # ### Input Command History # # By default, IRB stores a history of up to 1000 input commands in a file named # `.irb_history`. The history file will be in the same directory as the # [configuration file](rdoc-ref:IRB@Configuration+File) if one is found, or in # `~/` otherwise. # # A new IRB session creates the history file if it does not exist, and appends # to the file if it does exist. # # You can change the filepath by adding to your configuration file: # `IRB.conf[:HISTORY_FILE] = *filepath*`, where *filepath* is a string filepath. # # During the session, method `conf.history_file` returns the filepath, and # method `conf.history_file = *new_filepath*` copies the history to the file at # *new_filepath*, which becomes the history file for the session. # # You can change the number of commands saved by adding to your configuration # file: `IRB.conf[:SAVE_HISTORY] = *n*`, where *n* is one of: # # * Positive integer: the number of commands to be saved. # * Negative integer: all commands are to be saved. # * Zero or `nil`: no commands are to be saved. # # # During the session, you can use methods `conf.save_history` or # `conf.save_history=` to retrieve or change the count. # # ### Command Aliases # # By default, IRB defines several command aliases: # # irb(main):001> conf.command_aliases # => {:"$"=>:show_source, :"@"=>:whereami} # # You can change the initial aliases in the configuration file with: # # IRB.conf[:COMMAND_ALIASES] = {foo: :show_source, bar: :whereami} # # You can replace the current aliases at any time with configuration method # `conf.command_aliases=`; Because `conf.command_aliases` is a hash, you can # modify it. # # ### End-of-File # # By default, `IRB.conf[:IGNORE_EOF]` is `false`, which means that typing the # end-of-file character `Ctrl-D` causes the session to exit. # # You can reverse that behavior by adding `IRB.conf[:IGNORE_EOF] = true` to the # configuration file. # # During the session, method `conf.ignore_eof?` returns the setting, and method # `conf.ignore_eof = *boolean*` sets it. # # ### SIGINT # # By default, `IRB.conf[:IGNORE_SIGINT]` is `true`, which means that typing the # interrupt character `Ctrl-C` causes the session to exit. # # You can reverse that behavior by adding `IRB.conf[:IGNORE_SIGING] = false` to # the configuration file. # # During the session, method `conf.ignore_siging?` returns the setting, and # method `conf.ignore_sigint = *boolean*` sets it. # # ### Automatic Completion # # By default, IRB enables [automatic # completion](https://en.wikipedia.org/wiki/Autocomplete#In_command-line_interpr # eters): # # You can disable it by either of these: # # * Adding `IRB.conf[:USE_AUTOCOMPLETE] = false` to the configuration file. # * Giving command-line option `--noautocomplete` (`--autocomplete` is the # default). # # # Method `conf.use_autocomplete?` returns `true` if automatic completion is # enabled, `false` otherwise. # # The setting may not be changed during the session. # # ### Automatic Indentation # # By default, IRB automatically indents lines of code to show structure (e.g., # it indent the contents of a block). # # The current setting is returned by the configuration method # `conf.auto_indent_mode`. # # The default initial setting is `true`: # # irb(main):001> conf.auto_indent_mode # => true # irb(main):002* Dir.entries('.').select do |entry| # irb(main):003* entry.start_with?('R') # irb(main):004> end # => ["README.md", "Rakefile"] # # You can change the initial setting in the configuration file with: # # IRB.conf[:AUTO_INDENT] = false # # Note that the *current* setting *may not* be changed in the IRB session. # # ### Input Method # # The IRB input method determines how command input is to be read; by default, # the input method for a session is IRB::RelineInputMethod. Unless the # value of the TERM environment variable is 'dumb', in which case the # most simplistic input method is used. # # You can set the input method by: # # * Adding to the configuration file: # # * `IRB.conf[:USE_SINGLELINE] = true` or `IRB.conf[:USE_MULTILINE]= # false` sets the input method to IRB::ReadlineInputMethod. # * `IRB.conf[:USE_SINGLELINE] = false` or `IRB.conf[:USE_MULTILINE] = # true` sets the input method to IRB::RelineInputMethod. # # # * Giving command-line options: # # * `--singleline` or `--nomultiline` sets the input method to # IRB::ReadlineInputMethod. # * `--nosingleline` or `--multiline` sets the input method to # IRB::RelineInputMethod. # * `--nosingleline` together with `--nomultiline` sets the # input to IRB::StdioInputMethod. # # # Method `conf.use_multiline?` and its synonym `conf.use_reline` return: # # * `true` if option `--multiline` was given. # * `false` if option `--nomultiline` was given. # * `nil` if neither was given. # # # Method `conf.use_singleline?` and its synonym `conf.use_readline` return: # # * `true` if option `--singleline` was given. # * `false` if option `--nosingleline` was given. # * `nil` if neither was given. # # # ## Output # # This section describes the features that allow you to change the way IRB # output works; see also [Input and Output](rdoc-ref:IRB@Input+and+Output). # # ### Return-Value Printing (Echoing) # # By default, IRB prints (echoes) the values returned by all input commands. # # You can change the initial behavior and suppress all echoing by: # # * Adding to the configuration file: `IRB.conf[:ECHO] = false`. (The default # value for this entry is `nil`, which means the same as `true`.) # * Giving command-line option `--noecho`. (The default is `--echo`.) # # # During the session, you can change the current setting with configuration # method `conf.echo=` (set to `true` or `false`). # # As stated above, by default IRB prints the values returned by all input # commands; but IRB offers special treatment for values returned by assignment # statements, which may be: # # * Printed with truncation (to fit on a single line of output), which is the # default; an ellipsis (`...` is suffixed, to indicate the truncation): # # irb(main):001> x = 'abc' * 100 # # # > "abcabcabcabcabcabcabcabcabcabcabcabcabcabcabcabcabcabcabcabcabcabcabcabcabc... # # * Printed in full (regardless of the length). # * Suppressed (not printed at all) # # # You can change the initial behavior by: # # * Adding to the configuration file: `IRB.conf[:ECHO_ON_ASSIGNMENT] = false`. # (The default value for this entry is `niL`, which means the same as # `:truncate`.) # * Giving command-line option `--noecho-on-assignment` or # `--echo-on-assignment`. (The default is `--truncate-echo-on-assignment`.) # # # During the session, you can change the current setting with configuration # method `conf.echo_on_assignment=` (set to `true`, `false`, or `:truncate`). # # By default, IRB formats returned values by calling method `inspect`. # # You can change the initial behavior by: # # * Adding to the configuration file: `IRB.conf[:INSPECT_MODE] = false`. (The # default value for this entry is `true`.) # * Giving command-line option `--noinspect`. (The default is `--inspect`.) # # # During the session, you can change the setting using method # `conf.inspect_mode=`. # # ### Multiline Output # # By default, IRB prefixes a newline to a multiline response. # # You can change the initial default value by adding to the configuration file: # # IRB.conf[:NEWLINE_BEFORE_MULTILINE_OUTPUT] = false # # During a session, you can retrieve or set the value using methods # `conf.newline_before_multiline_output?` and # `conf.newline_before_multiline_output=`. # # Examples: # # irb(main):001> conf.inspect_mode = false # => false # irb(main):002> "foo\nbar" # => # foo # bar # irb(main):003> conf.newline_before_multiline_output = false # => false # irb(main):004> "foo\nbar" # => foo # bar # # ### Evaluation History # # By default, IRB saves no history of evaluations (returned values), and the # related methods `conf.eval_history`, `_`, and `__` are undefined. # # You can turn on that history, and set the maximum number of evaluations to be # stored: # # * In the configuration file: add `IRB.conf[:EVAL_HISTORY] = *n*`. (Examples # below assume that we've added `IRB.conf[:EVAL_HISTORY] = 5`.) # * In the session (at any time): `conf.eval_history = *n*`. # # # If `n` is zero, all evaluation history is stored. # # Doing either of the above: # # * Sets the maximum size of the evaluation history; defines method # `conf.eval_history`, which returns the maximum size `n` of the evaluation # history: # # irb(main):001> conf.eval_history = 5 # => 5 # irb(main):002> conf.eval_history # => 5 # # * Defines variable `_`, which contains the most recent evaluation, or `nil` # if none; same as method `conf.last_value`: # # irb(main):003> _ # => 5 # irb(main):004> :foo # => :foo # irb(main):005> :bar # => :bar # irb(main):006> _ # => :bar # irb(main):007> _ # => :bar # # * Defines variable `__`: # # * `__` unadorned: contains all evaluation history: # # irb(main):008> :foo # => :foo # irb(main):009> :bar # => :bar # irb(main):010> :baz # => :baz # irb(main):011> :bat # => :bat # irb(main):012> :bam # => :bam # irb(main):013> __ # => # 9 :bar # 10 :baz # 11 :bat # 12 :bam # irb(main):014> __ # => # 10 :baz # 11 :bat # 12 :bam # 13 ...self-history... # # Note that when the evaluation is multiline, it is displayed # differently. # # * `__[`*m*`]`: # # * Positive *m*: contains the evaluation for the given line number, # or `nil` if that line number is not in the evaluation history: # # irb(main):015> __[12] # => :bam # irb(main):016> __[1] # => nil # # * Negative *m*: contains the `mth`-from-end evaluation, or `nil` if # that evaluation is not in the evaluation history: # # irb(main):017> __[-3] # => :bam # irb(main):018> __[-13] # => nil # # * Zero *m*: contains `nil`: # # irb(main):019> __[0] # => nil # # # # # ### Prompt and Return Formats # # By default, IRB uses the prompt and return value formats defined in its # `:DEFAULT` prompt mode. # # #### The Default Prompt and Return Format # # The default prompt and return values look like this: # # irb(main):001> 1 + 1 # => 2 # irb(main):002> 2 + 2 # => 4 # # The prompt includes: # # * The name of the running program (`irb`); see [IRB # Name](rdoc-ref:IRB@IRB+Name). # * The name of the current session (`main`); See [IRB # Sessions](rdoc-ref:IRB@IRB+Sessions). # * A 3-digit line number (1-based). # # # The default prompt actually defines three formats: # # * One for most situations (as above): # # irb(main):003> Dir # => Dir # # * One for when the typed command is a statement continuation (adds trailing # asterisk): # # irb(main):004* Dir. # # * One for when the typed command is a string continuation (adds trailing # single-quote): # # irb(main):005' Dir.entries('. # # # You can see the prompt change as you type the characters in the following: # # irb(main):001* Dir.entries('.').select do |entry| # irb(main):002* entry.start_with?('R') # irb(main):003> end # => ["README.md", "Rakefile"] # # #### Pre-Defined Prompts # # IRB has several pre-defined prompts, stored in hash `IRB.conf[:PROMPT]`: # # irb(main):001> IRB.conf[:PROMPT].keys # => [:NULL, :DEFAULT, :CLASSIC, :SIMPLE, :INF_RUBY, :XMP] # # To see the full data for these, type `IRB.conf[:PROMPT]`. # # Most of these prompt definitions include specifiers that represent values like # the IRB name, session name, and line number; see [Prompt # Specifiers](rdoc-ref:IRB@Prompt+Specifiers). # # You can change the initial prompt and return format by: # # * Adding to the configuration file: `IRB.conf[:PROMPT] = *mode*` where # *mode* is the symbol name of a prompt mode. # * Giving a command-line option: # # * `--prompt *mode*`: sets the prompt mode to *mode*. where *mode* is the # symbol name of a prompt mode. # * `--simple-prompt` or `--sample-book-mode`: sets the prompt mode to # `:SIMPLE`. # * `--inf-ruby-mode`: sets the prompt mode to `:INF_RUBY` and suppresses # both `--multiline` and `--singleline`. # * `--noprompt`: suppresses prompting; does not affect echoing. # # # # You can retrieve or set the current prompt mode with methods # # `conf.prompt_mode` and `conf.prompt_mode=`. # # If you're interested in prompts and return formats other than the defaults, # you might experiment by trying some of the others. # # #### Custom Prompts # # You can also define custom prompts and return formats, which may be done # either in an IRB session or in the configuration file. # # A prompt in IRB actually defines three prompts, as seen above. For simple # custom data, we'll make all three the same: # # irb(main):001* IRB.conf[:PROMPT][:MY_PROMPT] = { # irb(main):002* PROMPT_I: ': ', # irb(main):003* PROMPT_C: ': ', # irb(main):004* PROMPT_S: ': ', # irb(main):005* RETURN: '=> ' # irb(main):006> } # => {:PROMPT_I=>": ", :PROMPT_C=>": ", :PROMPT_S=>": ", :RETURN=>"=> "} # # If you define the custom prompt in the configuration file, you can also make # it the current prompt by adding: # # IRB.conf[:PROMPT_MODE] = :MY_PROMPT # # Regardless of where it's defined, you can make it the current prompt in a # session: # # conf.prompt_mode = :MY_PROMPT # # You can view or modify the current prompt data with various configuration # methods: # # * `conf.prompt_mode`, `conf.prompt_mode=`. # * `conf.prompt_c`, `conf.c=`. # * `conf.prompt_i`, `conf.i=`. # * `conf.prompt_s`, `conf.s=`. # * `conf.return_format`, `return_format=`. # # # #### Prompt Specifiers # # A prompt's definition can include specifiers for which certain values are # substituted: # # * `%N`: the name of the running program. # * `%m`: the value of `self.to_s`. # * `%M`: the value of `self.inspect`. # * `%l`: an indication of the type of string; one of `"`, `'`, `/`, `]`. # * `%NNi`: Indentation level. NN is a 2-digit number that specifies the number # of digits of the indentation level (03 will result in 001). # * `%NNn`: Line number. NN is a 2-digit number that specifies the number # of digits of the line number (03 will result in 001). # * `%%`: Literal `%`. # # # ### Verbosity # # By default, IRB verbosity is disabled, which means that output is smaller # rather than larger. # # You can enable verbosity by: # # * Adding to the configuration file: `IRB.conf[:VERBOSE] = true` (the default # is `nil`). # * Giving command-line options `--verbose` (the default is `--noverbose`). # # # During a session, you can retrieve or set verbosity with methods # `conf.verbose` and `conf.verbose=`. # # ### Help # # Command-line option `--version` causes IRB to print its help text and exit. # # ### Version # # Command-line option `--version` causes IRB to print its version text and exit. # # ## Input and Output # # ### Color Highlighting # # By default, IRB color highlighting is enabled, and is used for both: # # * Input: As you type, IRB reads the typed characters and highlights elements # that it recognizes; it also highlights errors such as mismatched # parentheses. # * Output: IRB highlights syntactical elements. # # # You can disable color highlighting by: # # * Adding to the configuration file: `IRB.conf[:USE_COLORIZE] = false` (the # default value is `true`). # * Giving command-line option `--nocolorize` # # # ## Debugging # # Command-line option `-d` sets variables `$VERBOSE` and `$DEBUG` to `true`; # these have no effect on IRB output. # # ### Warnings # # Command-line option `-w` suppresses warnings. # # Command-line option `-W[*level*]` sets warning level; # # * 0=silence # * 1=medium # * 2=verbose # # ## Other Features # # ### Load Modules # # You can specify the names of modules that are to be required at startup. # # Array `conf.load_modules` determines the modules (if any) that are to be # required during session startup. The array is used only during session # startup, so the initial value is the only one that counts. # # The default initial value is `[]` (load no modules): # # irb(main):001> conf.load_modules # => [] # # You can set the default initial value via: # # * Command-line option `-r` # # $ irb -r csv -r json # irb(main):001> conf.load_modules # => ["csv", "json"] # # * Hash entry `IRB.conf[:LOAD_MODULES] = *array*`: # # IRB.conf[:LOAD_MODULES] = %w[csv, json] # # # Note that the configuration file entry overrides the command-line options. # # ### RI Documentation Directories # # You can specify the paths to RI documentation directories that are to be # loaded (in addition to the default directories) at startup; see details about # RI by typing `ri --help`. # # Array `conf.extra_doc_dirs` determines the directories (if any) that are to be # loaded during session startup. The array is used only during session startup, # so the initial value is the only one that counts. # # The default initial value is `[]` (load no extra documentation): # # irb(main):001> conf.extra_doc_dirs # => [] # # You can set the default initial value via: # # * Command-line option `--extra_doc_dir` # # $ irb --extra-doc-dir your_doc_dir --extra-doc-dir my_doc_dir # irb(main):001> conf.extra_doc_dirs # => ["your_doc_dir", "my_doc_dir"] # # * Hash entry `IRB.conf[:EXTRA_DOC_DIRS] = *array*`: # # IRB.conf[:EXTRA_DOC_DIRS] = %w[your_doc_dir my_doc_dir] # # # Note that the configuration file entry overrides the command-line options. # # ### IRB Name # # You can specify a name for IRB. # # The default initial value is `'irb'`: # # irb(main):001> conf.irb_name # => "irb" # # You can set the default initial value via hash entry `IRB.conf[:IRB_NAME] = # *string*`: # # IRB.conf[:IRB_NAME] = 'foo' # # ### Application Name # # You can specify an application name for the IRB session. # # The default initial value is `'irb'`: # # irb(main):001> conf.ap_name # => "irb" # # You can set the default initial value via hash entry `IRB.conf[:AP_NAME] = # *string*`: # # IRB.conf[:AP_NAME] = 'my_ap_name' # # ### Configuration Monitor # # You can monitor changes to the configuration by assigning a proc to # `IRB.conf[:IRB_RC]` in the configuration file: # # IRB.conf[:IRB_RC] = proc {|conf| puts conf.class } # # Each time the configuration is changed, that proc is called with argument # `conf`: # # ### Encodings # # Command-line option `-E *ex*[:*in*]` sets initial external (ex) and internal # (in) encodings. # # Command-line option `-U` sets both to UTF-8. # # ### Commands # # Please use the `help` command to see the list of available commands. # # ### IRB Sessions # # IRB has a special feature, that allows you to manage many sessions at once. # # You can create new sessions with Irb.irb, and get a list of current sessions # with the `jobs` command in the prompt. # # #### Configuration # # The command line options, or IRB.conf, specify the default behavior of # Irb.irb. # # On the other hand, each conf in IRB@Command-Line+Options is used to # individually configure IRB.irb. # # If a proc is set for `IRB.conf[:IRB_RC]`, its will be invoked after execution # of that proc with the context of the current session as its argument. Each # session can be configured using this mechanism. # # #### Session variables # # There are a few variables in every Irb session that can come in handy: # # `_` # : The value command executed, as a local variable # `__` # : The history of evaluated commands. Available only if # `IRB.conf[:EVAL_HISTORY]` is not `nil` (which is the default). See also # IRB::Context#eval_history= and IRB::History. # `__[line_no]` # : Returns the evaluation value at the given line number, `line_no`. If # `line_no` is a negative, the return value `line_no` many lines before the # most recent return value. # # # ## Restrictions # # Ruby code typed into IRB behaves the same as Ruby code in a file, except that: # # * Because IRB evaluates input immediately after it is syntactically # complete, some results may be slightly different. # * Forking may not be well behaved. # module IRB # An exception raised by IRB.irb_abort class Abort < Exception;end class << self # The current IRB::Context of the session, see IRB.conf # # irb # irb(main):001:0> IRB.CurrentContext.irb_name = "foo" # foo(main):002:0> IRB.conf[:MAIN_CONTEXT].irb_name #=> "foo" def CurrentContext # :nodoc: conf[:MAIN_CONTEXT] end # Initializes IRB and creates a new Irb.irb object at the `TOPLEVEL_BINDING` def start(ap_path = nil) STDOUT.sync = true $0 = File::basename(ap_path, ".rb") if ap_path setup(ap_path) if @CONF[:SCRIPT] irb = Irb.new(nil, @CONF[:SCRIPT]) else irb = Irb.new end irb.run(@CONF) end # Quits irb def irb_exit(*) # :nodoc: throw :IRB_EXIT, false end # Aborts then interrupts irb. # # Will raise an Abort exception, or the given `exception`. def irb_abort(irb, exception = Abort) # :nodoc: irb.context.thread.raise exception, "abort then interrupt!" end end class Irb # Note: instance and index assignment expressions could also be written like: # "foo.bar=(1)" and "foo.[]=(1, bar)", when expressed that way, the former be # parsed as :assign and echo will be suppressed, but the latter is parsed as a # :method_add_arg and the output won't be suppressed PROMPT_MAIN_TRUNCATE_LENGTH = 32 PROMPT_MAIN_TRUNCATE_OMISSION = '...' CONTROL_CHARACTERS_PATTERN = "\x00-\x1F" # Returns the current context of this irb session attr_reader :context # The lexer used by this irb session attr_accessor :scanner attr_reader :from_binding # Creates a new irb session def initialize(workspace = nil, input_method = nil, from_binding: false) @from_binding = from_binding @context = Context.new(self, workspace, input_method) @context.workspace.load_helper_methods_to_main @signal_status = :IN_IRB @scanner = RubyLex.new @line_no = 1 end # A hook point for `debug` command's breakpoint after :IRB_EXIT as well as its # clean-up def debug_break # it means the debug integration has been activated if defined?(DEBUGGER__) && DEBUGGER__.respond_to?(:capture_frames_without_irb) # after leaving this initial breakpoint, revert the capture_frames patch DEBUGGER__.singleton_class.send(:alias_method, :capture_frames, :capture_frames_without_irb) # and remove the redundant method DEBUGGER__.singleton_class.send(:undef_method, :capture_frames_without_irb) end end def debug_readline(binding) workspace = IRB::WorkSpace.new(binding) context.replace_workspace(workspace) context.workspace.load_helper_methods_to_main @line_no += 1 # When users run: # 1. Debugging commands, like `step 2` # 2. Any input that's not irb-command, like `foo = 123` # # # Irb#eval_input will simply return the input, and we need to pass it to the # debugger. input = nil forced_exit = catch(:IRB_EXIT) do if History.save_history? && context.io.support_history_saving? # Previous IRB session's history has been saved when `Irb#run` is exited We need # to make sure the saved history is not saved again by resetting the counter context.io.reset_history_counter begin input = eval_input ensure context.io.save_history end else input = eval_input end false end Kernel.exit if forced_exit if input&.include?("\n") @line_no += input.count("\n") - 1 end input end def run(conf = IRB.conf) in_nested_session = !!conf[:MAIN_CONTEXT] conf[:IRB_RC].call(context) if conf[:IRB_RC] prev_context = conf[:MAIN_CONTEXT] conf[:MAIN_CONTEXT] = context load_history = !in_nested_session && context.io.support_history_saving? save_history = load_history && History.save_history? if load_history context.io.load_history end prev_trap = trap("SIGINT") do signal_handle end begin if defined?(RubyVM.keep_script_lines) keep_script_lines_backup = RubyVM.keep_script_lines RubyVM.keep_script_lines = true end forced_exit = catch(:IRB_EXIT) do eval_input end ensure # Do not restore to nil. It will cause IRB crash when used with threads. IRB.conf[:MAIN_CONTEXT] = prev_context if prev_context RubyVM.keep_script_lines = keep_script_lines_backup if defined?(RubyVM.keep_script_lines) trap("SIGINT", prev_trap) conf[:AT_EXIT].each{|hook| hook.call} context.io.save_history if save_history Kernel.exit if forced_exit end end # Evaluates input for this session. def eval_input configure_io each_top_level_statement do |statement, line_no| signal_status(:IN_EVAL) do begin # If the integration with debugger is activated, we return certain input if it # should be dealt with by debugger if @context.with_debugger && statement.should_be_handled_by_debugger? return statement.code end @context.evaluate(statement, line_no) if @context.echo? && !statement.suppresses_echo? if statement.is_assignment? if @context.echo_on_assignment? output_value(@context.echo_on_assignment? == :truncate) end else output_value end end rescue SystemExit, SignalException raise rescue Interrupt, Exception => exc handle_exception(exc) @context.workspace.local_variable_set(:_, exc) end end end end def read_input(prompt) signal_status(:IN_INPUT) do @context.io.prompt = prompt if l = @context.io.gets print l if @context.verbose? else if @context.ignore_eof? and @context.io.readable_after_eof? l = "\n" if @context.verbose? printf "Use \"exit\" to leave %s\n", @context.ap_name end else print "\n" if @context.prompting? end end l end end def readmultiline prompt = generate_prompt([], false, 0) # multiline return read_input(prompt) if @context.io.respond_to?(:check_termination) # nomultiline code = +'' line_offset = 0 loop do line = read_input(prompt) unless line return code.empty? ? nil : code end code << line return code if command?(code) tokens, opens, terminated = @scanner.check_code_state(code, local_variables: @context.local_variables) return code if terminated line_offset += 1 continue = @scanner.should_continue?(tokens) prompt = generate_prompt(opens, continue, line_offset) end end def each_top_level_statement loop do code = readmultiline break unless code yield build_statement(code), @line_no @line_no += code.count("\n") rescue RubyLex::TerminateLineInput end end def build_statement(code) if code.match?(/\A\n*\z/) return Statement::EmptyInput.new end code = code.dup.force_encoding(@context.io.encoding) if (command, arg = @context.parse_command(code)) command_class = Command.load_command(command) Statement::Command.new(code, command_class, arg) else is_assignment_expression = @scanner.assignment_expression?(code, local_variables: @context.local_variables) Statement::Expression.new(code, is_assignment_expression) end end def command?(code) !!@context.parse_command(code) end def configure_io if @context.io.respond_to?(:check_termination) @context.io.check_termination do |code| if Reline::IOGate.in_pasting? rest = @scanner.check_termination_in_prev_line(code, local_variables: @context.local_variables) if rest Reline.delete_text rest.bytes.reverse_each do |c| Reline.ungetc(c) end true else false end else next true if command?(code) _tokens, _opens, terminated = @scanner.check_code_state(code, local_variables: @context.local_variables) terminated end end end if @context.io.respond_to?(:dynamic_prompt) @context.io.dynamic_prompt do |lines| tokens = RubyLex.ripper_lex_without_warning(lines.map{ |l| l + "\n" }.join, local_variables: @context.local_variables) line_results = IRB::NestingParser.parse_by_line(tokens) tokens_until_line = [] line_results.map.with_index do |(line_tokens, _prev_opens, next_opens, _min_depth), line_num_offset| line_tokens.each do |token, _s| # Avoid appending duplicated token. Tokens that include "n" like multiline # tstring_content can exist in multiple lines. tokens_until_line << token if token != tokens_until_line.last end continue = @scanner.should_continue?(tokens_until_line) generate_prompt(next_opens, continue, line_num_offset) end end end if @context.io.respond_to?(:auto_indent) and @context.auto_indent_mode @context.io.auto_indent do |lines, line_index, byte_pointer, is_newline| next nil if lines == [nil] # Workaround for exit IRB with CTRL+d next nil if !is_newline && lines[line_index]&.byteslice(0, byte_pointer)&.match?(/\A\s*\z/) code = lines[0..line_index].map { |l| "#{l}\n" }.join tokens = RubyLex.ripper_lex_without_warning(code, local_variables: @context.local_variables) @scanner.process_indent_level(tokens, lines, line_index, is_newline) end end end def convert_invalid_byte_sequence(str, enc) str.force_encoding(enc) str.scrub { |c| c.bytes.map{ |b| "\\x#{b.to_s(16).upcase}" }.join } end def encode_with_invalid_byte_sequence(str, enc) conv = Encoding::Converter.new(str.encoding, enc) dst = String.new begin ret = conv.primitive_convert(str, dst) case ret when :invalid_byte_sequence conv.insert_output(conv.primitive_errinfo[3].dump[1..-2]) redo when :undefined_conversion c = conv.primitive_errinfo[3].dup.force_encoding(conv.primitive_errinfo[1]) conv.insert_output(c.dump[1..-2]) redo when :incomplete_input conv.insert_output(conv.primitive_errinfo[3].dump[1..-2]) when :finished end break end while nil dst end def handle_exception(exc) if exc.backtrace[0] =~ /\/irb(2)?(\/.*|-.*|\.rb)?:/ && exc.class.to_s !~ /^IRB/ && !(SyntaxError === exc) && !(EncodingError === exc) # The backtrace of invalid encoding hash (ex. {"\xAE": 1}) raises EncodingError without lineno. irb_bug = true else irb_bug = false # To support backtrace filtering while utilizing Exception#full_message, we need to clone # the exception to avoid modifying the original exception's backtrace. exc = exc.clone filtered_backtrace = exc.backtrace.map { |l| @context.workspace.filter_backtrace(l) }.compact backtrace_filter = IRB.conf[:BACKTRACE_FILTER] if backtrace_filter if backtrace_filter.respond_to?(:call) filtered_backtrace = backtrace_filter.call(filtered_backtrace) else warn "IRB.conf[:BACKTRACE_FILTER] #{backtrace_filter} should respond to `call` method" end end exc.set_backtrace(filtered_backtrace) end highlight = Color.colorable? order = if RUBY_VERSION < '3.0.0' STDOUT.tty? ? :bottom : :top else # '3.0.0' <= RUBY_VERSION :top end message = exc.full_message(order: order, highlight: highlight) message = convert_invalid_byte_sequence(message, exc.message.encoding) message = encode_with_invalid_byte_sequence(message, IRB.conf[:LC_MESSAGES].encoding) unless message.encoding.to_s.casecmp?(IRB.conf[:LC_MESSAGES].encoding.to_s) message = message.gsub(/((?:^\t.+$\n)+)/) { |m| case order when :top lines = m.split("\n") when :bottom lines = m.split("\n").reverse end unless irb_bug if lines.size > @context.back_trace_limit omit = lines.size - @context.back_trace_limit lines = lines[0..(@context.back_trace_limit - 1)] lines << "\t... %d levels..." % omit end end lines = lines.reverse if order == :bottom lines.map{ |l| l + "\n" }.join } # The "" in "(irb)" may be the top level of IRB so imitate the main object. message = message.gsub(/\(irb\):(?\d+):in (?[`'])<(?top \(required\))>'/) { "(irb):#{$~[:num]}:in #{$~[:open_quote]}
'" } puts message puts 'Maybe IRB bug!' if irb_bug rescue Exception => handler_exc begin puts exc.inspect puts "backtraces are hidden because #{handler_exc} was raised when processing them" rescue Exception puts 'Uninspectable exception occurred' end end # Evaluates the given block using the given `path` as the Context#irb_path and # `name` as the Context#irb_name. # # Used by the irb command `source`, see IRB@IRB+Sessions for more information. def suspend_name(path = nil, name = nil) @context.irb_path, back_path = path, @context.irb_path if path @context.irb_name, back_name = name, @context.irb_name if name begin yield back_path, back_name ensure @context.irb_path = back_path if path @context.irb_name = back_name if name end end # Evaluates the given block using the given `workspace` as the # Context#workspace. # # Used by the irb command `irb_load`, see IRB@IRB+Sessions for more information. def suspend_workspace(workspace) current_workspace = @context.workspace @context.replace_workspace(workspace) yield ensure @context.replace_workspace current_workspace end # Evaluates the given block using the given `input_method` as the Context#io. # # Used by the irb commands `source` and `irb_load`, see IRB@IRB+Sessions for # more information. def suspend_input_method(input_method) back_io = @context.io @context.instance_eval{@io = input_method} begin yield back_io ensure @context.instance_eval{@io = back_io} end end # Handler for the signal SIGINT, see Kernel#trap for more information. def signal_handle unless @context.ignore_sigint? print "\nabort!\n" if @context.verbose? exit end case @signal_status when :IN_INPUT print "^C\n" raise RubyLex::TerminateLineInput when :IN_EVAL IRB.irb_abort(self) when :IN_LOAD IRB.irb_abort(self, LoadAbort) when :IN_IRB # ignore else # ignore other cases as well end end # Evaluates the given block using the given `status`. def signal_status(status) return yield if @signal_status == :IN_LOAD signal_status_back = @signal_status @signal_status = status begin yield ensure @signal_status = signal_status_back end end def output_value(omit = false) # :nodoc: str = @context.inspect_last_value multiline_p = str.include?("\n") if omit winwidth = @context.io.winsize.last if multiline_p first_line = str.split("\n").first result = @context.newline_before_multiline_output? ? (@context.return_format % first_line) : first_line output_width = Reline::Unicode.calculate_width(result, true) diff_size = output_width - Reline::Unicode.calculate_width(first_line, true) if diff_size.positive? and output_width > winwidth lines, _ = Reline::Unicode.split_by_width(first_line, winwidth - diff_size - 3) str = "%s..." % lines.first str += "\e[0m" if Color.colorable? multiline_p = false else str = str.gsub(/(\A.*?\n).*/m, "\\1...") str += "\e[0m" if Color.colorable? end else output_width = Reline::Unicode.calculate_width(@context.return_format % str, true) diff_size = output_width - Reline::Unicode.calculate_width(str, true) if diff_size.positive? and output_width > winwidth lines, _ = Reline::Unicode.split_by_width(str, winwidth - diff_size - 3) str = "%s..." % lines.first str += "\e[0m" if Color.colorable? end end end if multiline_p && @context.newline_before_multiline_output? str = "\n" + str end Pager.page_content(format(@context.return_format, str), retain_content: true) end # Outputs the local variables to this current session, including #signal_status # and #context, using IRB::Locale. def inspect ary = [] for iv in instance_variables case (iv = iv.to_s) when "@signal_status" ary.push format("%s=:%s", iv, @signal_status.id2name) when "@context" ary.push format("%s=%s", iv, eval(iv).__to_s__) else ary.push format("%s=%s", iv, eval(iv)) end end format("#<%s: %s>", self.class, ary.join(", ")) end private def generate_prompt(opens, continue, line_offset) ltype = @scanner.ltype_from_open_tokens(opens) indent = @scanner.calc_indent_level(opens) continue = opens.any? || continue line_no = @line_no + line_offset if ltype f = @context.prompt_s elsif continue f = @context.prompt_c else f = @context.prompt_i end f = "" unless f if @context.prompting? p = format_prompt(f, ltype, indent, line_no) else p = "" end if @context.auto_indent_mode and !@context.io.respond_to?(:auto_indent) unless ltype prompt_i = @context.prompt_i.nil? ? "" : @context.prompt_i ind = format_prompt(prompt_i, ltype, indent, line_no)[/.*\z/].size + indent * 2 - p.size p += " " * ind if ind > 0 end end p end def truncate_prompt_main(str) # :nodoc: str = str.tr(CONTROL_CHARACTERS_PATTERN, ' ') if str.size <= PROMPT_MAIN_TRUNCATE_LENGTH str else str[0, PROMPT_MAIN_TRUNCATE_LENGTH - PROMPT_MAIN_TRUNCATE_OMISSION.size] + PROMPT_MAIN_TRUNCATE_OMISSION end end def format_prompt(format, ltype, indent, line_no) # :nodoc: format.gsub(/%([0-9]+)?([a-zA-Z%])/) do case $2 when "N" @context.irb_name when "m" main_str = @context.main.to_s rescue "!#{$!.class}" truncate_prompt_main(main_str) when "M" main_str = @context.main.inspect rescue "!#{$!.class}" truncate_prompt_main(main_str) when "l" ltype when "i" if indent < 0 if $1 "-".rjust($1.to_i) else "-" end else if $1 format("%" + $1 + "d", indent) else indent.to_s end end when "n" if $1 format("%" + $1 + "d", line_no) else line_no.to_s end when "%" "%" unless $1 end end end end end class Binding # Opens an IRB session where `binding.irb` is called which allows for # interactive debugging. You can call any methods or variables available in the # current scope, and mutate state if you need to. # # Given a Ruby file called `potato.rb` containing the following code: # # class Potato # def initialize # @cooked = false # binding.irb # puts "Cooked potato: #{@cooked}" # end # end # # Potato.new # # Running `ruby potato.rb` will open an IRB session where `binding.irb` is # called, and you will see the following: # # $ ruby potato.rb # # From: potato.rb @ line 4 : # # 1: class Potato # 2: def initialize # 3: @cooked = false # => 4: binding.irb # 5: puts "Cooked potato: #{@cooked}" # 6: end # 7: end # 8: # 9: Potato.new # # irb(#):001:0> # # You can type any valid Ruby code and it will be evaluated in the current # context. This allows you to debug without having to run your code repeatedly: # # irb(#):001:0> @cooked # => false # irb(#):002:0> self.class # => Potato # irb(#):003:0> caller.first # => ".../2.5.1/lib/ruby/2.5.0/irb/workspace.rb:85:in `eval'" # irb(#):004:0> @cooked = true # => true # # You can exit the IRB session with the `exit` command. Note that exiting will # resume execution where `binding.irb` had paused it, as you can see from the # output printed to standard output in this example: # # irb(#):005:0> exit # Cooked potato: true # # See IRB for more information. def irb(show_code: true) # Setup IRB with the current file's path and no command line arguments IRB.setup(source_location[0], argv: []) unless IRB.initialized? # Create a new workspace using the current binding workspace = IRB::WorkSpace.new(self) # Print the code around the binding if show_code is true STDOUT.print(workspace.code_around_binding) if show_code # Get the original IRB instance debugger_irb = IRB.instance_variable_get(:@debugger_irb) irb_path = File.expand_path(source_location[0]) if debugger_irb # If we're already in a debugger session, set the workspace and irb_path for the original IRB instance debugger_irb.context.replace_workspace(workspace) debugger_irb.context.irb_path = irb_path # If we've started a debugger session and hit another binding.irb, we don't want # to start an IRB session instead, we want to resume the irb:rdbg session. IRB::Debug.setup(debugger_irb) IRB::Debug.insert_debug_break debugger_irb.debug_break else # If we're not in a debugger session, create a new IRB instance with the current # workspace binding_irb = IRB::Irb.new(workspace, from_binding: true) binding_irb.context.irb_path = irb_path binding_irb.run(IRB.conf) binding_irb.debug_break end end end