MadLisp is a [Lisp](https://en.wikipedia.org/wiki/Lisp_%28programming_language%29) interpreter written in PHP. It is inspired by the [Make a Lisp](https://github.com/kanaka/mal) project, but does not follow that convention or syntax strictly. It provides a fun platform for learning [functional programming](https://en.wikipedia.org/wiki/Functional_programming).
* REPL environment where the user can interactively experiment with the language. Suitable for executing pieces of code one by one and examining the internal state of the system.
* Minimal safeguards or restrictions as to what can be done. Breaking things or using the language in unexpected ways should be part of the fun.
* Performance does not need to match commercial-grade languages, but needs to be fast enough for real-world programs and uses cases.
* Suitable to be used as a scripting language in Linux shell scripts and similar environments.
* Suitable to be used as an embedded scripting language inside another PHP application.
* Clear and intuitive error messages. This is important for pleasant user experience.
* Provide a library with commonly used features such as HTTP requests, JSON processing and SQL database support.
* Provide a clean [interface](src/Lib/ILib.php) for extending the language with your own functions defined in PHP.
* Provide a safe-mode where access to the file system and other external I/O is restricted.
* Provide a debug mode which shows what is happening inside the code evaluation.
* Loosely respect the Lisp legacy with things like naming conventions but do not be constrained by it.
## Non-goals
* Ability to call arbitrary PHP functions directly. The language should have control over which PHP functions can be called and how.
* Namespaces or similar mechanisms. The global namespace is a feature, not a bug! Use a prefix for your function names if this becomes a problem.
The core project does not have any dependencies to external [Composer](https://getcomposer.org/) libraries, but it does currently use Composer for the autoloader so you need to run **composer install** for that.
You can create an init file in your home directory with the name `.madlisp_init`. This file is automatically executed when the interpreter is started. It is useful for registering commonly used functions and performing other initialization.
## Using from PHP
You can use the [LispFactory](src/LispFactory.php) class to create an instance of the interpreter if you wish to embed the MadLisp language in your PHP application and call it directly from your code.
The language includes a safe-mode that disables functions which allow external I/O. This allows a "sandbox" to be created where the evaluated scripts do not have access to the file system or other resources.
Internally lists and vectors are just PHP arrays wrapped in a class, and the only difference between the two is how they are evaluated. Another reason for adding vectors is the familiarity of the square bracket syntax for PHP developers. They can be thought of as PHP arrays for most intents and purposes.
Hash maps are collections of key-value pairs. Keys are normal strings, not "keywords" starting with colon characters as in many Lisp languages.
Hash maps are defined using curly brackets or using the built-in `hash` function. Odd arguments are treated as keys and even arguments are treated as values. The key-value pair can optionally include colon as a separator to make it more readable, but it is ignored internally.
Symbols are words which do not match any other type and are separated by whitespace. They can contain special characters. Examples of symbols are `a`, `name` or `+`.
A function is applied or "called" when a list is evaluated. The function is the first item of the list and the remaining items are arguments to the function. When a function is applied, a new environment is created where the bindings are bound to the given arguments, and the function body is then evaluated in this new environment.
Environments are hash-maps which store key-value pairs and use symbols as keys. If the key is not defined in current environment the lookup proceeds to the parent environment and so forth. The initial environment is called `root` and contains all the built-in functions listed here. Another environment called `user` is created for anything the user wants to define.
The first argument to let is a list of bindings defined in the new environment. In this example the value of `a` is set to 1, and the value of `b` to 2. Then the body expression, `(+ a b)` in the example, is evaluated in the new environment.
Conditional evaluation is accomplished with the `if` expression which is of the form `(if test consequent alternate)`. If *test* evaluates to truthy value, *consequent* is evaluated and returned. If *test* evaluates to falsy value, *alternate* is evaluated and returned:
> (cond ((= n 2) "two") ((= n 4) "four") ((= n 6) "six"))
"four"
```
For `case`, the first argument is evaluated, and then it is matched against the first item of the remaining arguments. If there is a match, the following expression is evaluated and returned:
Looping is accomplished with the `while` expression which is of the form `(while test expr1 expr2 ...)`. The *test* is evaluated at the beginning of each iteration and if it returns truthy value, the remaining expressions are evaluated. The value of the whole expression is the value of the last evaluated sub-expression.
```text
> (let (i 5) (while (> i 0) (print i) (def i (dec i))))
543210
```
Although the above example illustrates how to use `while`, this type of code is discouraged. Generally it is recommended to use recursion instead of iteration in these type of scenarios. Usually it results in cleaner code as well. The `while` expression is better suited for something like the main loop of a program.
Use the `quasiquote` special form when you need to turn on evaluation temporarily inside the quoted element. The special forms `unquote` and `unquote-splice` are available for that purpose:
Internally `quasiquote` expands to `cons` and `concat` functions. We can use the `quasiquote-expand` special form to test this expansion without evaluation:
The language has support for Lisp-style macros. Macros are like preprocessor directives and allow the manipulation of the language syntax before evaluation.
There are two built-in macros: `defn` which is a shortcut for the form `(def ... (fn ...))` and `defmacro` which is a shortcut for the form `(def ... (macro ...))`. To illustrate how macros work, lets look at the definition of `defn`:
The language has support for `try-catch` style exception handlers. The syntax is `(try A (catch B C))` where A is evaluated first and if exception is thrown, then C is evaluated with the symbol B bound to the value of the exception. Exceptions are thrown using the `throw` core function. You can give any data structure as argument to `throw` and it is passed along to `catch`. This way exceptions can contain more data than just a string that represents an error message.
Simple example of throwing and catching an exception:
```
> (try (throw {"msg":"message"}) (catch ex (str "error: " (get ex "msg"))))
"error: message"
```
Exceptions generated by PHP are catched as well. Their value will be a hash-map with keys `type`, `file`, `line` and `message`:
print | no | `(print "hello world")` | `hello world` | Print expression on the screen. Give optional second argument as `true` to show strings in readable format. Print returns null (shown in REPL).
throw | yes | `(throw "invalid value")` | `error: "invalid value"` | Throw an exception. The given value is passed to catch. See the section Exceptions.
hash | `(hash "a" 1 "b" 2)` | `{"a":1 "b":2}` | Create a new hash-map.
list | `(list 1 2 3)` | `(1 2 3)` | Create a new list.
vector | `(vector 1 2 3)` | `[1 2 3]` | Create a new vector.
range | `(range 2 5)` | `[2 3 4]` | Create a vector with integer values from first to argument (inclusive) to second argument (exclusive).
range | `(range 5)` | `[0 1 2 3 4]` | Range can also be used with one argument in which case it is used as length for a vector of integers starting from 0.
map | `(map (fn (a) (* a 2)) [1 2 3])` | `[2 4 6]` | Create new sequence by calling a function for each item. Uses [array_map](https://www.php.net/manual/en/function.array-map.php) internally.
map2 | `(map2 + [1 2 3] [4 5 6])` | `[5 7 9]` | Create new sequence by calling a function for each item from both sequences.
reduce | `(reduce + [2 3 4] 1)` | `10` | Reduce a sequence to a single value by calling a function sequentially of all arguments. Optional third argument is used to give the initial value for first iteration. Uses [array_reduce](https://www.php.net/manual/en/function.array-reduce.php) internally.
filter | `(filter odd? [1 2 3 4 5])` | `[1 3 5]` | Create a new sequence by using the given function as a filter. Uses [array_filter](https://www.php.net/manual/en/function.array-filter.php) internally.
filterh | `(filterh (fn (v k) (prefix? k "a")) {"aa":1 "ab":2 "bb":3})` | `{"aa":1 "ab":2}` | Same as filter but for hash-maps. First argument passed to the callback is the value and second is the key.
reverse | `(reverse [1 2 3])` | `[3 2 1]` | Reverse the order of a sequence. Uses [array_reverse](https://www.php.net/manual/en/function.array-reverse.php) internally.
key? | `(key? {"a" "b"} "a")` | `true` | Return true if the hash-map contains the given key.
set | `(set {"a" 1} "b" 2)` | `{"a":1 "b":2}` | Create new hash-map which contains the given key-value pair.
set! | `(set! {"a" 1} "b" 2)` | `2` | Modify the given hash-map by setting the given key-value pair and return the set value. **This function is mutable!**
unset | `(unset {"a":1 "b":2 "c":3} "b")` | `{"a":1 "c":3}` | Create a new hash-map with the given key removed.
unset! | `(unset! {"a":1 "b":2 "c":3} "b")` | `2` | Modify the given hash-map by removing the given key and return the corresponding value. **This function is mutable!**
keys | `(keys {"a" 1 "b" 2})` | `("a" "b")` | Return a list of the keys for a hash-map.
values | `(values {"a" 1 "b" 2})` | `(1 2)` | Return a list of the values for a hash-map.
zip | `(zip ["a" "b"] [1 2])` | `{"a":1 "b":2}` | Create a hash-map using the first sequence as keys and the second as values. Uses [array_combine](https://www.php.net/manual/en/function.array-combine.php) internally.
sort | `(sort [6 4 8 1])` | `[1 4 6 8]` | Sort the sequence using [sort](https://www.php.net/manual/en/function.sort.php).
usort | `(usort (fn (a b) (if (< a b) 0 1)) [3 1 5 4 2])` | `[1 2 3 4 5]` | Sort the sequence using custom comparison function using [usort](https://www.php.net/manual/en/function.usort.php).
db-open | `(def d (db-open "mysql:host=localhost;dbname=test" "testuser" "testpw"))` | `<object<PDO>>` | Open a database connection.
db-execute | `(db-execute d "INSERT INTO test_table (col1, col2) values (?, ?)" [1, 2])` | `1` | Execute a SQL statement and return the number of affected rows.
db-query | `(db-query d "SELECT * FROM test_table WHERE col1 = ?" [1])` | | Execute a SELECT statement.
db-last-id | `(db-last-id d)` | `"1"` | Return the last id of auto-increment column.
db-trans | `(db-trans d)` | `true` | Start a transaction.
db-commit | `(db-commit d)` | `true` | Commit a transaction.
db-rollback | `(db-rollback d)` | `true` | Roll back a transaction.
http | `(http "POST" "http://example.com/" (to-json {"key":"value"}) {"Content-Type":"application/json"})` | `{"status":200 "body":"" "headers":{}}` | Perform a HTTP request. First argument is the HTTP method, second is URL, third is request body and fourth is headers as a hash-map. The function returns a hash-map which contains keys `status`, `body` and `headers`.
fget | `(fget "test.txt")` | `"content"` | Read the contents of a file using [file_get_contents](https://www.php.net/manual/en/function.file-get-contents.php).
fput | `(fput "test.txt" "content")` | `true` | Write string to file using [file_put_contents](https://www.php.net/manual/en/function.file-put-contents.php). Give optional third parameter as `true` to append.
fopen | `(def f (fopen "test.txt" "w"))` | `<resource>` | Open a file for reading or writing. Give the mode as second argument.
fclose | `(fclose f)` | `true` | Close a file resource.
fwrite | `(fwrite f "abc")` | `3` | Write to a file resource.
fflush | `(fflush f)` | `true` | Persist buffered writes to disk for a file resource.
fread | `(fread f 16)` | `"abc"` | Read from a file resource.
feof? | `(feof? f)` | `true` | Return true if end of file has been reached for a file resource.
readline | `(readline "What is your name? ")` | `What is your name? ` | Read line of user input using [readline](https://www.php.net/manual/en/function.readline.php).
readline-add | `(readline-add "What is your name? ")` | `true` | Add line of user input to readline history using [readline_add_history](https://www.php.net/manual/en/function.readline-add-history.php).
readline-load | `(readline-load "historyfile")` | `true` | Read readline history from file using [readline_read_history](https://www.php.net/manual/en/function.readline-read-history.php).
readline-save | `(readline-save "historyfile")` | `true` | Write readline history into file using [readline_write_history](https://www.php.net/manual/en/function.readline-write-history.php).
re-match | `(re-match "/^[a-z]{4}[0-9]{4}$/" "test1234")` | `true` | Match subject to regular expression using [preg_match](https://www.php.net/manual/en/function.preg-match.php).
re-match-all | `(re-match-all "/[A-Z][a-z]{2}[0-9]/" "One1 Two2 Three3")` | `["One1" "Two2"]` | Find multiple matches to regular expression using [preg_match_all](https://www.php.net/manual/en/function.preg-match-all.php).
re-replace | `(re-replace "/year ([0-9]{4}) month ([0-9]{2})/" "$1-$2-01" "year 2020 month 10")` | `"2020-10-01"` | Perform search and replace with regular expression using [preg_replace](https://www.php.net/manual/en/function.preg-replace.php).
re-split | `(re-split "/\\s+/" "aa bb cc ")` | `["aa" "bb" "cc"]` | Split the subject by regular expression using [preg_split](https://www.php.net/manual/en/function.preg-split.php). The flag `PREG_SPLIT_NO_EMPTY` is enabled.
Note that support for multibyte characters in strings is limited because the provided functions do not use the [mbstring](https://www.php.net/manual/en/book.mbstring.php) extension.
### Time functions
Name | Example | Example result | Description
------- | ------- | -------------- | -----------
time | `(time)` | `1592011969` | Return the current unix timestamp using [time](https://www.php.net/manual/en/function.time).
mtime | `(mtime)` | `1607696761.132` | Return the current unix timestamp as float that includes microseconds. Uses [microtime](https://www.php.net/manual/en/function.microtime).
date | `(date "Y-m-d H:i:s")` | `"2020-06-13 08:33:29"` | Format the current time and date using [date](https://www.php.net/manual/en/function.date.php).
strtotime | `(strtotime "2020-06-13 08:34:47")` | `1592012087` | Parse datetime string into unix timestamp using [strtotime](https://www.php.net/manual/en/function.strtotime.php).
### Type functions
Skipped examples here as these are pretty self-explanatory.
Name | Description
------- | -----------
bool | Convert the argument to boolean.
float | Convert the argument to floating-point value.
int | Convert the argument to integer.
str | Convert the argument to string. Also concatenate multiple strings together.
symbol | Convert the argument to symbol.
not | Turns true to false and vice versa.
type | Return the type of the argument as a string.