shell

The Erlang Shell

The module shell implements an Erlang shell.

The shell is a user interface program for entering expression sequences. The expressions are evaluated and a value is returned. A history mechanism saves previous commands and their values, which can then be incorporated in later commands. How many commands and results to save can be determined by the user, either interactively, by calling shell:history/1 and shell:results/1, or by setting the application configuration parameters shell_history_length and shell_saved_results for the application STDLIB.

The shell uses a helper process for evaluating commands in order to protect the history mechanism from exceptions. By default the evaluator process is killed when an exception occurs, but by calling shell:catch_exception/1 or by setting the application configuration parameter shell_catch_exception for the application STDLIB this behavior can be changed. See also the example below.

Variable bindings, and local process dictionary changes which are generated in user expressions are preserved, and the variables can be used in later commands to access their values. The bindings can also be forgotten so the variables can be re-used.

The special shell commands all have the syntax of (local) function calls. They are evaluated as normal function calls and many commands can be used in one expression sequence.

If a command (local function call) is not recognized by the shell, an attempt is first made to find the function in the module user_default, where customized local commands can be placed. If found, then the function is evaluated. Otherwise, an attempt is made to evaluate the function in the module shell_default. The module user_default must be explicitly loaded.

The shell also permits the user to start multiple concurrent jobs. A job can be regarded as a set of processes which can communicate with the shell.

There is some support for reading and printing records in the shell. During compilation record expressions are translated to tuple expressions. In runtime it is not known whether a tuple actually represents a record. Nor are the record definitions used by compiler available at runtime. So in order to read the record syntax and print tuples as records when possible, record definitions have to be maintained by the shell itself. The shell commands for reading, defining, forgetting, listing, and printing records are described below. Note that each job has its own set of record definitions. To facilitate matters record definitions in the modules shell_default and user_default (if loaded) are read each time a new job is started. For instance, adding the line

          -include_lib("kernel/include/file.hrl").

to user_default makes the definition of file_info readily available in the shell.

The shell runs in two modes:

Only the currently connected job can 'talk' to the shell.

Shell Commands

b()

Prints the current variable bindings.

f()

Removes all variable bindings.

f(X)

Removes the binding of variable X.

h()

Prints the history list.

history(N)

Sets the number of previous commands to keep in the history list to N. The previous number is returned. The default number is 20.

results(N)

Sets the number of results from previous commands to keep in the history list to N. The previous number is returned. The default number is 20.

e(N)

Repeats the command N, if N is positive. If it is negative, the Nth previous command is repeated (i.e., e(-1) repeats the previous command).

v(N)

Uses the return value of the command N in the current command, if N is positive. If it is negative, the return value of the Nth previous command is used (i.e., v(-1) uses the value of the previous command).

help()

Evaluates shell_default:help().

c(File)

Evaluates shell_default:c(File). This compiles and loads code in File and purges old versions of code, if necessary. Assumes that the file and module names are the same.

catch_exception(Bool)

Sets the exception handling of the evaluator process. The previous exception handling is returned. The default (false) is to kill the evaluator process when an exception occurs, which causes the shell to create a new evaluator process. When the exception handling is set to true the evaluator process lives on which means that for instance ports and ETS tables as well as processes linked to the evaluator process survive the exception.

rd(RecordName, RecordDefinition)

Defines a record in the shell. RecordName is an atom and RecordDefinition lists the field names and the default values. Usually record definitions are made known to the shell by use of the rr commands described below, but sometimes it is handy to define records on the fly.

rf()

Removes all record definitions, then reads record definitions from the modules shell_default and user_default (if loaded). Returns the names of the records defined.

rf(RecordNames)

Removes selected record definitions. RecordNames is a record name or a list of record names. Use '_' to remove all record definitions.

rl()

Prints all record definitions.

rl(RecordNames)

Prints selected record definitions. RecordNames is a record name or a list of record names.

rp(Term)

Prints a term using the record definitions known to the shell. All of Term is printed; the depth is not limited as is the case when a return value is printed.

rr(Module)

Reads record definitions from a module's BEAM file. If there are no record definitions in the BEAM file, the source file is located and read instead. Returns the names of the record definitions read. Module is an atom.

rr(Wildcard)

Reads record definitions from files. Existing definitions of any of the record names read are replaced. Wildcard is a wildcard string as defined in filelib(3) but not an atom.

rr(WildcardOrModule, RecordNames)

Reads record definitions from files but discards record names not mentioned in RecordNames (a record name or a list of record names).

rr(WildcardOrModule, RecordNames, Options)

Reads record definitions from files. The compiler options {i, Dir}, {d, Macro}, and {d, Macro, Value} are recognized and used for setting up the include path and macro definitions. Use '_' as value of RecordNames to read all record definitions.

Example

The following example is a long dialogue with the shell. Commands starting with > are inputs to the shell. All other lines are output from the shell. All commands in this example are explained at the end of the dialogue. .

strider 1> erl
Erlang (BEAM) emulator version 5.3 [hipe] [threads:0]

Eshell V5.3  (abort with ^G)
1>Str = "abcd".
"abcd"
2> L = length(Str).
4
3> Descriptor = {L, list_to_atom(Str)}.
{4,abcd}
4> L. 
4
5> b().
Descriptor = {4,abcd}
L = 4
Str = "abcd"
ok
6> f(L). 
ok
7> b().
Descriptor = {4,abcd}
Str = "abcd"
ok
8> f(L).
ok
9> {L, _} = Descriptor.
{4,abcd}
10> L.
4
11> {P, Q, R} = Descriptor.
** exception error: no match of right hand side value {4,abcd}
12> P.
* 1: variable 'P' is unbound **
13> Descriptor.
{4,abcd}
14>{P, Q} = Descriptor.
{4,abcd}
15> P.
4
16> f().
ok
17> put(aa, hello).
undefined
18> get(aa).
hello
19> Y = test1:demo(1).
11
20> get().
[{aa,worked}]
21> put(aa, hello).
worked
22> Z = test1:demo(2).
** exception error: no match of right hand side value 1
     in function  test1:demo/1
23> Z.
* 1: variable 'Z' is unbound **
24> get(aa).
hello
25> erase(), put(aa, hello).
undefined
26> spawn(test1, demo, [1]).
<0.57.0>
27> get(aa).
hello
28> io:format("hello hello\n").
hello hello
ok
29> e(28).
hello hello
ok
30> v(28).
ok
31> c(ex).
{ok,ex}
32> rr(ex).
[rec]
33> rl(rec).
-record(rec,{a,b = val()}).
ok
34> #rec{}.
** exception error: undefined shell command val/0
35> #rec{b = 3}.
#rec{a = undefined,b = 3}
36> rp(v(-1)).
#rec{a = undefined,b = 3}
ok
37> rd(rec, {f = orddict:new()}).
rec
38> #rec{}.
#rec{f = []}
ok
39> rd(rec, {c}), A.
* 1: variable 'A' is unbound **
40> #rec{}.
#rec{c = undefined}
ok
41> test1:loop(0).
Hello Number: 0
Hello Number: 1
Hello Number: 2
Hello Number: 3

User switch command
 --> i
 --> c
.
.
.
Hello Number: 3374
Hello Number: 3375
Hello Number: 3376
Hello Number: 3377
Hello Number: 3378
** exception exit: killed
42> E = ets:new(t, []).
17
43> ets:insert({d,1,2}).
** exception error: undefined function ets:insert/1
44> ets:insert(E, {d,1,2}).
** exception error: argument is of wrong type
     in function  ets:insert/2
        called as ets:insert(16,{d,1,2})
45> f(E).
ok
46> catch_exception(true).
false
47> E = ets:new(t, []).
18
48> ets:insert({d,1,2}).
* exception error: undefined function ets:insert/1
49> ets:insert(E, {d,1,2}).
true
50> halt().
strider 2>

Comments

Command 1 sets the variable Str to the string "abcd".

Command 2 sets L to the length of the string evaluating the BIF atom_to_list.

Command 3 builds the tuple Descriptor.

Command 4 prints the value of the variable L.

Command 5 evaluates the internal shell command b(), which is an abbreviation of "bindings". This prints the current shell variables and their bindings. The ok at the end is the return value of the b() function.

Command 6 f(L) evaluates the internal shell command f(L) (abbreviation of "forget"). The value of the variable L is removed.

Command 7 prints the new bindings.

Command 8 has no effect since L has no value.

Command 9 performs a pattern matching operation on Descriptor, binding a new value to L.

Command 10 prints the current value of L.

Command 11 tries to match {P, Q, R} against Descriptor which is {4, abc}. The match fails and none of the new variables become bound. The printout starting with "** exception error:" is not the value of the expression (the expression had no value because its evaluation failed), but rather a warning printed by the system to inform the user that an error has occurred. The values of the other variables (L, Str, etc.) are unchanged.

Commands 12 and 13 show that P is unbound because the previous command failed, and that Descriptor has not changed.

Commands 14 and 15 show a correct match where P and Q are bound.

Command 16 clears all bindings.

The next few commands assume that test1:demo(X) is defined in the following way:

demo(X) ->
    put(aa, worked),
    X = 1,
    X + 10.    

Commands 17 and 18 set and inspect the value of the item aa in the process dictionary.

Command 19 evaluates test1:demo(1). The evaluation succeeds and the changes made in the process dictionary become visible to the shell. The new value of the dictionary item aa can be seen in command 20.

Commands 21 and 22 change the value of the dictionary item aa to hello and call test1:demo(2). Evaluation fails and the changes made to the dictionary in test1:demo(2), before the error occurred, are discarded.

Commands 23 and 24 show that Z was not bound and that the dictionary item aa has retained its original value.

Commands 25, 26 and 27 show the effect of evaluating test1:demo(1) in the background. In this case, the expression is evaluated in a newly spawned process. Any changes made in the process dictionary are local to the newly spawned process and therefore not visible to the shell.

Commands 28, 29 and 30 use the history facilities of the shell.

Command 29 is e(28). This re-evaluates command 28. Command 30 is v(28). This uses the value (result) of command 28. In the cases of a pure function (a function with no side effects), the result is the same. For a function with side effects, the result can be different.

The next few commands show some record manipulation. It is assumed that ex.erl defines a record like this:

-record(rec, {a, b = val()}).

val() ->
    3.    

Commands 31 and 32 compiles the file ex.erl and reads the record definitions in ex.beam. If the compiler did not output any record definitions on the BEAM file, rr(ex) tries to read record definitions from the source file instead.

Command 33 prints the definition of the record named rec.

Command 34 tries to create a rec record, but fails since the function val/0 is undefined. Command 35 shows the workaround: explicitly assign values to record fields that cannot otherwise be initialized.

Command 36 prints the newly created record using record definitions maintained by the shell.

Command 37 defines a record directly in the shell. The definition replaces the one read from the file ex.beam.

Command 38 creates a record using the new definition, and prints the result.

Command 39 and 40 show that record definitions are updated as side effects. The evaluation of the command fails but the definition of rec has been carried out.

For the next command, it is assumed that test1:loop(N) is defined in the following way:

loop(N) ->
    io:format("Hello Number: ~w~n", [N]), 
    loop(N+1).

Command 41 evaluates test1:loop(0), which puts the system into an infinite loop. At this point the user types Control G, which suspends output from the current process, which is stuck in a loop, and activates JCL mode. In JCL mode the user can start and stop jobs.

In this particular case, the i command ("interrupt") is used to terminate the looping program, and the c command is used to connect to the shell again. Since the process was running in the background before we killed it, there will be more printouts before the "** exception exit: killed" message is shown.

Command 42 creates an ETS table.

Command 43 tries to insert a tuple into the ETS table but the first argument (the table) is missing. The exception kills the evaluator process.

Command 44 corrects the mistake, but the ETS table has been destroyed since it was owned by the killed evaluator process.

Command 46 sets the exception handling of the evaluator process to true. The exception handling can also be set when starting Erlang, like this: erl -stdlib shell_catch_exception true.

Command 48 makes the same mistake as in command 43, but this time the evaluator process lives on. The single star at the beginning of the printout signals that the exception has been caught.

Command 49 successfully inserts the tuple into the ETS table.

The halt() command exits the Erlang runtime system.

JCL Mode

When the shell starts, it starts a single evaluator process. This process, together with any local processes which it spawns, is referred to as a job. Only the current job, which is said to be connected, can perform operations with standard IO. All other jobs, which are said to be detached, are blocked if they attempt to use standard IO.

All jobs which do not use standard IO run in the normal way.

The shell escape key ^G (Control G) detaches the current job and activates JCL mode. The JCL mode prompt is "-->". If "?" is entered at the prompt, the following help message is displayed:

          --> ?
          c [nn]            - connect to job
          i [nn]            - interrupt job
          k [nn]            - kill job
          j                 - list all jobs
          s [shell]         - start local shell
          r [node [shell]]  - start remote shell
          q        - quit erlang
          ? | h             - this message    

The JCL commands have the following meaning:

c [nn]

Connects to job number <nn> or the current job. The standard shell is resumed. Operations which use standard IO by the current job will be interleaved with user inputs to the shell.

i [nn]

Stops the current evaluator process for job number nn or the current job, but does not kill the shell process. Accordingly, any variable bindings and the process dictionary will be preserved and the job can be connected again. This command can be used to interrupt an endless loop.

k [nn]

Kills job number nn or the current job. All spawned processes in the job are killed, provided they have not evaluated the group_leader/1 BIF and are located on the local machine. Processes spawned on remote nodes will not be killed.

j

Lists all jobs. A list of all known jobs is printed. The current job name is prefixed with '*'.

s

Starts a new job. This will be assigned the new index [nn] which can be used in references.

s [shell]

Starts a new job. This will be assigned the new index [nn] which can be used in references. If the optional argument shell is given, it is assumed to be a module that implements an alternative shell.

r [node]

Starts a remote job on node. This is used in distributed Erlang to allow a shell running on one node to control a number of applications running on a network of nodes. If the optional argument shell is given, it is assumed to be a module that implements an alternative shell.

q

Quits Erlang. Note that this option is disabled if Erlang is started with the ignore break, +Bi, system flag (which may be useful e.g. when running a restricted shell, see below).

?

Displays this message.

It is possible to alter the behavior of shell escape by means of the STDLIB application variable shell_esc. The value of the variable can be either jcl (erl -stdlib shell_esc jcl) or abort (erl -stdlib shell_esc abort). The first option sets ^G to activate JCL mode (which is also default behavior). The latter sets ^G to terminate the current shell and start a new one. JCL mode cannot be invoked when shell_esc is set to abort.

If you want an Erlang node to have a remote job active from the start (rather than the default local job), you start Erlang with the -remsh flag. Example: erl -sname this_node -remsh other_node@other_host

Restricted Shell

The shell may be started in a restricted mode. In this mode, the shell evaluates a function call only if allowed. This feature makes it possible to, for example, prevent a user from accidentally calling a function from the prompt that could harm a running system (useful in combination with the the system flag +Bi).

When the restricted shell evaluates an expression and encounters a function call or an operator application, it calls a callback function (with information about the function call in question). This callback function returns true to let the shell go ahead with the evaluation, or false to abort it. There are two possible callback functions for the user to implement:

local_allowed(Func, ArgList, State) -> {true,NewState} | {false,NewState}

to determine if the call to the local function Func with arguments ArgList should be allowed.

non_local_allowed(FuncSpec, ArgList, State) -> {true,NewState} | {false,NewState} | {{redirect,NewFuncSpec,NewArgList},NewState}

to determine if the call to non-local function FuncSpec ({Module,Func} or a fun) with arguments ArgList should be allowed. The return value {redirect,NewFuncSpec,NewArgList} can be used to let the shell evaluate some other function than the one specified by FuncSpec and ArgList.

These callback functions are in fact called from local and non-local evaluation function handlers, described in the erl_eval manual page. (Arguments in ArgList are evaluated before the callback functions are called.)

The State argument is a tuple {ShellState,ExprState}. The return value NewState has the same form. This may be used to carry a state between calls to the callback functions. Data saved in ShellState lives through an entire shell session. Data saved in ExprState lives only through the evaluation of the current expression.

There are two ways to start a restricted shell session:

Notes:

Errors when loading the callback module is handled in different ways depending on how the restricted shell is activated:

Prompting

The default shell prompt function displays the name of the node (if the node can be part of a distributed system) and the current command number. The user can customize the prompt function by calling shell:prompt_func/1 or by setting the application configuration parameter shell_prompt_func for the application STDLIB.

A customized prompt function is stated as a tuple {Mod, Func}. The function is called as Mod:Func(L), where L is a list of key-value pairs created by the shell. Currently there is only one pair: {history, N}, where N is the current command number. The function should return a list of characters or an atom. This constraint is due to the Erlang I/O-protocol. Unicode characters beyond codepoint 255 are allowed in the list. Note that in restricted mode the call Mod:Func(L) must be allowed or the default shell prompt function will be called.

Functions


history(N) -> integer()

  • N = integer()

Sets the number of previous commands to keep in the history list to N. The previous number is returned. The default number is 20.

results(N) -> integer()

  • N = integer()

Sets the number of results from previous commands to keep in the history list to N. The previous number is returned. The default number is 20.

catch_exception(Bool) -> Bool

  • Bool = bool()

Sets the exception handling of the evaluator process. The previous exception handling is returned. The default (false) is to kill the evaluator process when an exception occurs, which causes the shell to create a new evaluator process. When the exception handling is set to true the evaluator process lives on which means that for instance ports and ETS tables as well as processes linked to the evaluator process survive the exception.

prompt_func(PromptFunc) -> prompt_func()

  • PromptFunc = prompt_func()
  • prompt_func() = default | {Mod, Func}
  • Mod = Func = atom()

Sets the shell prompt function to PromptFunc. The previous prompt function is returned.

start_restricted(Module) -> ok | {error, Reason}

  • Module = atom()
  • Reason = atom()

Exits a normal shell and starts a restricted shell. Module specifies the callback module for the functions local_allowed/3 and non_local_allowed/3. The function is meant to be called from the shell.

If the callback module cannot be loaded, an error tuple is returned. The Reason in the error tuple is the one returned by the code loader when trying to load the code of the callback module.

stop_restricted() -> ok

Exits a restricted shell and starts a normal shell. The function is meant to be called from the shell.