Why Sponsor Oils? | source | all docs for version 0.22.0 | all versions | oilshell.org
Oils Reference — Chapter Types and Methods
This chapter describes YSH types and methods. There are also two OSH types for bash compatibility.
(in progress)
These two types are for OSH code only.
A bash array holds a sequence of strings. Some entries may be unset, i.e. not an empty string.
See sh-array for details. In YSH, prefer to use List instances.
A bash associative array is a mapping from strings to strings.
See sh-assoc for details. In YSH, prefer to use Dict instances.
The Null type has a single value spelled null. (Related:
atom-literal).
The Bool type has 2 values: true and false. (Related: atom-literal).
Integers are currently 64-bit signed integers (on all platforms). TODO: they should be arbitrary precision.
There are many way of writing integers; see int-literal.
In shell, ASCII strings like '42' are often used for calculations on
integers. But you can use a "real" integer type in YSH.
Floats are at least 32 bits wide.
See float-literal for how to denote them.
In Oils, strings may contains any sequence of bytes, which may be UTF-8 encoded.
Internal NUL bytes (0x00) are allowed.
When passing such strings to say the cd builtin, the string will be
truncated before the NUL. This is because most C functions like chdir() take
NUL-terminated strings.
Replace substrings with a given string.
= mystr => replace("OSH", "YSH")
Or match with an Eggex.
= mystr => replace(/ d+ /, "<redacted>") # => "code is <redacted>"
Refer to Eggex captures with replacement expressions. Captured values can be
referenced with $1, $2, etc.
var mystr = "1989-06-08"
var pat = / <capture d{4}> '-' <capture d{2}> '-' <capture d{2}> /
= mystr => replace(pat, ^"Year: $1, Month: $2, Day: $3")
Captures can also be named.
= mystr2 => replace(/ <capture digit{4} as year : int> /, ^"$[year + 1]")
$0 refers to the entire capture itself in a substitution string.
var mystr = "replace with mystr => replace()"
= mystr => replace(/ alpha+ '=>' alpha+ '()' /, ^"<code>$0</code>")
# => "replace with <code>mystr => replace()</code>"
In addition to captures, other variables can be referenced within a replacement expression.
= mystr => replace(/ <capture alpha+> /, ^"$1 and $anotherVar")
To limit the number of replacements, pass in a named count argument. By default
the count is -1. For any count in [0, MAX_INT], there will be at most count
replacements. Any negative count means "replace all" (ie. count=-2 behaves
exactly like count=-1).
var mystr = "bob has a friend named bob"
= mystr => replace("bob", "Bob", count=1) # => "Bob has a friend named bob"
= mystr => replace("bob", "Bob", count=-1) # => "Bob has a friend named Bob"
The following matrix of signatures are supported by replace():
s => replace(string_val, subst_str)
s => replace(string_val, subst_expr)
s => replace(eggex_val, subst_str)
s => replace(eggex_val, subst_expr)
Checks if a string starts with a pattern, returning true if it does or false if it does not.
= b'YSH123' => startsWith(b'YSH') # => true
= b'123YSH' => startsWith(b'YSH') # => false
= b'123YSH' => startsWith(/ d+ /) # => true
= b'YSH123' => startsWith(/ d+ /) # => false
Matching is done based on bytes, not runes.
= b'\yce\ya3' # => (Str) "Σ"
= 'Σ' => startsWith(b'\yce') # => true
= 'Σ' => endsWith(b'\ya3') # => true
Like 'startsWith()` but returns true if the end of the string matches.
= b'123YSH' => endsWith("YSH") # => true
= b'YSH123' => endsWith(/ d+ /) # => true
Removes characters matching a pattern from the start and end of a string. With no arguments, whitespace is removed. When given a string or eggex pattern, that pattern is removed if it matches the start or end.
= b' YSH\n' => trim() # => "YSH"
= b'xxxYSHxxx' => trim('xxx') # => "YSH"
= b'xxxYSH ' => trim('xxx') # => "YSH "
= b' YSHxxx' => trim('xxx') # => " YSH"
= b' YSH ' => trim('xxx') # => " YSH "
= b'123YSH456' => trim(/ d+ /) # => "YSH"
When stripping whitespace, Oils decodes the bytes in string as utf-8 characters. Only the following Unicode codepoints are considered to be whitespace.
\t)\n)\v)\f)\r)<NBSP><ZWNBSP>While the Unicode standard defines other codepoints as being spaces, Oils limits itself to just these codepoints so that the specification is stable, and doesn't depend on an external standard that has reclassify characters.
Like trim() but only removes characters from the start of the string.
= b' YSH\n' => trimStart() # => "YSH\n"
= b'xxxYSHxxx' => trimStart(b'xxx') # => "YSHxxx"
= b'123YSH456' => trimStart(/ d+ /) # => "YSH456"
Like trim() but only removes characters from the end of the string.
= b' YSH\n' => trimEnd() # => " YSH"
= b'xxxYSHxxx' => trimEnd(b'xxx') # => "YxxxSH"
= b'123YSH456' => trimEnd(/ d+ /) # => "123YSH"
Respects unicode.
Respects unicode.
Search for the first occurrence of a regex in the string.
var m = 'hi world' => search(/[aeiou]/) # search for vowels
# matches at position 1 for 'i'
Returns a value.Match() if it matches, otherwise null.
You can start searching in the middle of the string:
var m = 'hi world' => search(/dot 'orld'/, pos=3)
# also matches at position 4 for 'o'
The %start or ^ metacharacter will only match when pos is zero.
(Similar to Python's re.search().)
leftMatch() is like search(), but it checks
var m = 'hi world' => leftMatch(/[aeiou]/) # search for vowels
# doesn't match because h is not a vowel
var m = 'aye' => leftMatch(/[aeiou]/)
# matches 'a'
leftMatch() Can be used to implement lexers that consume every byte of input.
var lexer = / <capture digit+> | <capture space+> /
(Similar to Python's re.match().)
A List contains an ordered sequence of values.
Add an element to a list.
var fruits = :|apple banana pear|
call fruits->append("orange")
echo @fruits # => apple banana pear orange
Similar names: append
remove an element from a list and return it.
var fruits = :|apple banana pear orange|
var last = fruits->pop() # "orange" is removed AND returned
echo $last # => orange
echo @fruits # => apple banana pear
Extend an existing list with the elements of another list.
var foods = :|cheese chocolate|
var fruits = :|apple banana|
call foods->extend(fruits)
echo @foods # => cheese chocolate apple banana
Returns the first index of the element in the list, or -1 if it's not present.
var names = :| Jane Peter Joana Sam |
echo $[names => indexOf("Sam")] # => 3
echo $[names => indexOf("Simon")] # => -1
Reverses a list in place.
var fruits = :|apple banana pear|
call fruits->reverse()
echo @fruits # => pear banana apple
A Dict contains an ordered sequence of key-value pairs. Given the key, the value can be retrieved efficiently.
Returns all existing keys from a dict as a list of strings.
var en2fr = {
hello: "bonjour",
friend: "ami",
cat: "chat"
}
= en2fr => keys()
# => (List 0x4689) ["hello","friend","cat"]
Similar to keys(), but returns the values of the dictionary.
var person = {
name: "Foo",
age: 25,
hobbies: :|walking reading|
}
= en2fr => values()]
# => (List 0x4689) ["Foo",25,["walking","reading"]]
A Range is a pair of two numbers, like 42 .. 45.
Ranges are used for iteration; see ysh-for.
An Eggex is a composable regular expression. It can be spliced into other
regular expressions.
A Match is the result searching for an Eggex within a Str.
Returns the string that matched a regex capture group. Group 0 is the entire match.
var m = '10:59' => search(/ ':' <capture d+> /)
echo $[m => group(0)] # => ':59'
echo $[m => group(1)] # => '59'
Matches can be named with as NAME:
var m = '10:59' => search(/ ':' <capture d+ as minute> /)
And then accessed by the same name:
echo $[m => group('minute')] # => '59'
Like group(), but returns the start position of a regex capture group,
rather than its value.
var m = '10:59' => search(/ ':' <capture d+ as minute> /)
echo $[m => start(0)] # => position 2 for ':59'
echo $[m => start(1)] # => position 3 for '59'
echo $[m => start('minute')] # => position 3 for '59'
Like group(), but returns the end position of a regex capture group,
rather than its value.
var m = '10:59' => search(/ ':' <capture d+ as minute> /)
echo $[m => end(0)] # => position 5 for ':59'
echo $[m => end(1)] # => position 5 for '59'
echo $[m => end('minute')] # => 5 for '59'
A Place is used as an "out param" by calling setValue():
proc p (out) {
call out->setValue('hi')
}
var x
p (&x)
echo x=$x # => x=hi
An unevaluated expression. You can create an Expr with an expression literal
(expr-literal):
var expr = ^[42 + a[i]]
An unevaluated command. You can create a Command with a "block expression"
(block-expr):
var block = ^(echo $PWD; ls *.txt)
A func that's part of Oils, like len().
The thin-arrow and fat-arrow create bound funcs:
var bound = '' => upper
var bound2 = [] -> append
User-defined functions.
User-defined procs.
TODO:
A module is a file with YSH code.
Like the eval builtin, but useful in pure functions.
Like $(), but useful in pure functions.
An API the wraps the $PS1 language. For example, to simulate PS1='\w\$ ':
func renderPrompt(io) {
var parts = []
call parts->append(io->promptval('w')) # pass 'w' for \w
call parts->append(io->promptval('$')) # pass '$' for \$
call parts->append(' ')
return (join(parts))
}
TODO: Depends on system clock.
TODO: Like the awk function, this takes an timestamp directly.
In other words, it calls C localtime() (which depends on the time zone database), and then C strftime().
TODO: The free function glob() actually does I/O. Although maybe it doesn't fail?