Why Sponsor Oils? | source | all docs for version 0.22.0 | all versions | oilshell.org

Warning: Work in progress! Leave feedback on Zulip or Github if you'd like this doc to be updated.

Interpreter State

The Oils project has a single interpreter that supports both the OSH and YSH languages.

In other words, It's useful to think of Unix shell in historical layers:

• OSH: A compatible but cleaned-up shell language.
  1. Thompson Shell (pipelines, exit status)
  2. Bourne Shell (variables, functions)
  3. Korn Shell (indexed arrays)
  4. Bash (shopt, associative arrays)
• YSH: A new shell language that manipulates the same interpreter state in a cleaner way.

Example

Shell has many syntaxes for the same semantics, which can be confusing. For example, in bash, these four statements do similar things:

$ foo='bar'
$ declare -g foo=bar
$ x='foo=bar'; typeset $x
$ printf -v foo bar

$ echo $foo
bar

In addition, YSH adds JavaScript-like syntax:

var foo = 'bar'

YSH syntax can express more data types, but it may also confuse new users.

So the sections below describe the shell from a semantic perspective, which should help users reason about their programs.

Quick tip: Use the pp builtin to inspect shell variables.

Design Goals

Simplify and Rationalize bash

POSIX shell has a fairly simple model: everything is a string, and "$@" is a special case.

Bash adds many features on top of POSIX, including arrays and associative arrays. Oils implements those features, and a few more.

However, it also significantly simplifies the model.

A primary difference is mentioned in Known Differences:

• In bash, the locations of values are tagged with types, e.g. declare -A unset_assoc_array.
• In Oils, values are tagged with types. This is how common dynamic languages like Python and JavaScript behave.

In other words, Oils "salvages" the confusing semantics of bash and produces something simpler, while still being very compatible.

Add New Features and Types

TODO

• eggex type
• later: floating point type

High Level Description

Memory Is a Stack

• Shell has a stack but no heap. The stack stores:
  • Variables that are local to a function.
  • The arguments array which is spelled "$@" in shell, and @ARGV in YSH.
• Shell's memory has values and locations, but no references/pointers.

Environment Variables Become Global Variables

On initialization, environment variables like PYTHONPATH=. are copied into the shell's memory as global variables, with the export flag set.

Global variables are stored in the first stack frame, i.e. the one at index 0.

Functions and Variables Are Separate

There are two distinct namespaces. For example:

foo() {
  echo 'function named foo'
}
foo=bar   # a variable; doesn't affect the function

Variable Name Lookup with "Dynamic Scope"

OSH has it, but YSH limits it.

Limitations of Arrays And Compound Data Structures

Shell is a value-oriented language.

• Can't Be Nested
• Can't Be Passed to Functions or Returned From Functions
• Can't Take References; Must be Copied

Example:

declare -a myarray=("${other_array[@]}")   # shell

var myarray = :| @other_array |            # Oils

Reason: There's no Garbage collection.

Integers and Coercion

• Strings are coerced to integers to do math.
• What about -i in bash?

Unix fork() Has Copy-On-Write Semantics

See the Process Model document.

Key Data Types

TODO: core/runtime.asdl

cell

TODO

• export only applies to strings

value

Undef, Str, Sequential/Indexed Arrays, Associative Array

• OSH has value.BashArray, and YSH has value.List.
• no integers, but there is (( ))
• "$@" is an array, and "${a[@]}" too
  • not true in bash -- it's fuzzy there
  • but $@ and ${a[@]} are NOT arrays
• flags: readonly and exported (but arrays/assoc arrays shouldn't be exported)
  • TODO: find that

cmd_value for shell builtins

Another important type:

  assign_arg = (lvalue lval, value? rval, int spid)

  cmd_value =
    Argv(string* argv, int* arg_spids, command__BraceGroup? block)
  | Assign(builtin builtin_id,
           string* argv, int* arg_spids,
           assign_arg* pairs)

Printing State

Shell Builtins

Oils supports various shell and bash operations to view the interpreter state.

• set prints variables and their values
• set -o prints options
• declare/typeset/readonly/export -p prints a subset of variables
• test -v tests if a variable is defined.

pp in Oils

Pretty prints a cell.

This is cleaner!

TODO: What about functions

Modifying State

YSH Keywords

TODO: See YSH Keywords doc.

Shell Assignment Builtins: declare/typeset, readonly, export

...

unset

You can't unset an array in OSH? But you can in bash.

Other Builtins

• read. Sometimes sets the magic $REPLY variable.
• getopts

Links

• Process Model
• https://opensource.com/article/18/5/you-dont-know-bash-intro-bash-arrays
• https://www.thegeekstuff.com/2010/06/bash-array-tutorial

Appendix: Bash Issues

Strings and Arrays Are Confused

Horrible

a=('1 2' 3)
b=(1 '2 3')  # two different elements

[[ $a == $b ]]
[[ ${a[0]} == ${b[0]} ]]

[[ ${a[@]} == ${b[@]} ]]

Associative arrays and being undefined

• half an array type
  • strict_array removes this
  • case $x in "$@"
• half an associative array type

Indexed Arrays and Associative Arrays Are Confused

Empty and Unset Are Confused

• empty array conflicts with set -o nounset (in bash 4.3). I can't recommend in good faith.