[normand.git] / README.adoc

// Show ToC at a specific location for a GitHub rendering
ifdef::env-github[]
:toc: macro
endif::env-github[]

ifndef::env-github[]
:toc: left
endif::env-github[]

// This is to mimic what GitHub does so that anchors work in an offline
// rendering too.
:idprefix:
:idseparator: -

// Other attributes
:py3: Python{nbsp}3

= Normand
Philippe Proulx

[.normal]
image:https://img.shields.io/pypi/v/normand.svg?label=Latest%20version[link="https://pypi.python.org/pypi/normand"]

[.lead]
_**Normand**_ is a text-to-binary processor with its own language.

This package offers both a portable {py3} module and a command-line
tool.

WARNING: This version of Normand is 0.1, meaning both the Normand
language and the module/CLI interface aren't stable.

ifdef::env-github[]
// ToC location for a GitHub rendering
toc::[]
endif::env-github[]

== Introduction

The purpose of Normand is to consume human-readable text representing
bytes and to produce the corresponding binary data.

.Simple bytes input.
====
Consider the following Normand input:

----
4f 55 32 bb $167 fe %10100111 a9 $-32
----

The generated nine bytes are:

----
4f 55 32 bb a7 fe a7 a9  e0
----
====

As you can see in the last example, the fundamental unit of the Normand
language is the _byte_. The order in which you list bytes will be the
order of the generated data.

The Normand language is more than simple lists of bytes, though. Its
main features are:

Comments, including a bunch of insignificant symbols which may improve readability::
+
Input:
+
----
ff bb %1101:0010 # This is a comment
78 29 af $192 # This too # 99 $-80
fe80::6257:18ff:fea3:4229
60:57:18:a3:42:29
10839636-5d65-4a68-8e6a-21608ddf7258
----
+
Output:
+
----
ff bb d2 78 29 af c0 99  b0 fe 80 62 57 18 ff fe
a3 42 29 60 57 18 a3 42  29 10 83 96 36 5d 65 4a
68 8e 6a 21 60 8d df 72  58
----

Hexadecimal, decimal, and binary byte constants::
+
Input:
+
----
aa bb $247 $-89 %0011_0010 %11.01= 10/10
----
+
Output:
+
----
aa bb f7 a7 32 da
----

UTF-8, UTF-16, and UTF-32 literal strings::
+
Input:
+
----
"hello world!" 00
u16le"stress\nverdict 🤣"
----
+
Output:
+
----
68 65 6c 6c 6f 20 77 6f  72 6c 64 21 00 73 00 74  ┆ hello world!•s•t
00 72 00 65 00 73 00 73  00 0a 00 76 00 65 00 72  ┆ •r•e•s•s•••v•e•r
00 64 00 69 00 63 00 74  00 20 00 3e d8 23 dd     ┆ •d•i•c•t• •>•#•
----

Labels: special variables holding the offset where they're defined::
+
----
<beg> b2 52 e3 bc 91 05
$100 $50 <chair> 33 9f fe
25 e9 89 8a <end>
----

Variables::
+
----
5e 65 {tower = 47} c6 7f f2 c4
44 {hurl = tower - 14} b5 {tower = hurl} 26 2d
----
+
The value of a variable assignment is the evaluation of a valid {py3}
expression which may include label and variable names.

Value encoding with a specific length (8{nbsp}bits to 64{nbsp}bits) and byte order::
+
Input:
+
----
{strength = 4}
{be} 67 <lbl> 44 $178 {(end - lbl) * 8 + strength : 16} $99 <end>
{le} {-1993 : 32}
----
+
Output:
+
----
67 44 b2 00 2c 63 37 f8  ff ff
----
+
The encoded value is the evaluation of a valid {py3} expression which
may include label and variable names.

Repetition::
+
Input:
+
----
aa bb * 5 cc "yeah\0" * 8
----
+
Output:
+
----
aa bb bb bb bb bb cc 79  65 61 68 00 79 65 61 68  ┆ •••••••yeah.yeah
00 79 65 61 68 00 79 65  61 68 00 79 65 61 68 00  ┆ •yeah•yeah•yeah•
79 65 61 68 00 79 65 61  68 00 79 65 61 68 00     ┆ yeah•yeah•yeah•
----


Multilevel grouping::
+
Input:
+
----
ff ((aa bb "zoom" cc) * 5) * 3 $-34 * 4
----
+
Output:
+
----
ff aa bb 7a 6f 6f 6d cc  aa bb 7a 6f 6f 6d cc aa  ┆ •••zoom•••zoom••
bb 7a 6f 6f 6d cc aa bb  7a 6f 6f 6d cc aa bb 7a  ┆ •zoom•••zoom•••z
6f 6f 6d cc aa bb 7a 6f  6f 6d cc aa bb 7a 6f 6f  ┆ oom•••zoom•••zoo
6d cc aa bb 7a 6f 6f 6d  cc aa bb 7a 6f 6f 6d cc  ┆ m•••zoom•••zoom•
aa bb 7a 6f 6f 6d cc aa  bb 7a 6f 6f 6d cc aa bb  ┆ ••zoom•••zoom•••
7a 6f 6f 6d cc aa bb 7a  6f 6f 6d cc aa bb 7a 6f  ┆ zoom•••zoom•••zo
6f 6d cc aa bb 7a 6f 6f  6d cc de de de de        ┆ om•••zoom•••••
----

Precise error reporting::
+
----
/tmp/meow.normand:10:24 - Expecting a bit (`0` or `1`).
----
+
----
/tmp/meow.normand:32:6 - Unexpected character `k`.
----
+
----
/tmp/meow.normand:24:19 - Unknown variable/label name `meow` in expression `(meow - 45) // 8`.
----
+
----
/tmp/meow.normand:18:9 - Value 315 is outside the 8-bit range when evaluating expression `end - ICITTE` at byte offset 45.
----

You can use Normand to track data source files in your favorite VCS
instead of raw binary files. The binary files that Normand generates can
be used to test file format decoding, including malformatted data, for
example, as well as for education.

See <<learn-normand>> to explore all the Normand features.

== Install Normand

Normand requires Python ≥ 3.4.

To install Normand:

----
$ python3 -m pip install --user normand
----

See
https://packaging.python.org/en/latest/tutorials/installing-packages/#installing-to-the-user-site[Installing to the User Site]
to learn more about a user site installation.

[NOTE]
====
Normand has a single module file, `normand.py`, which you can copy as is
to your project to use it (both the <<python-3-api,`normand.parse()`>>
function and the <<command-line-tool,command-line tool>>).

`normand.py` has _no external dependencies_, but if you're using
Python{nbsp}3.4, you'll need a local copy of the standard `typing`
module.
====

== Learn Normand

A Normand text input is a sequence of items which represent a sequence
of raw bytes.

[[state]] During the processing of items to data, Normand relies on a
current state:

[%header%autowidth]
|===
|State variable |Description |Initial value: <<python-3-api,{py3} API>> |Initial value: <<command-line-tool,CLI>>

|[[cur-offset]] Current offset
|
The current offset has an effect on the value of
<<label,labels>> and of the special `ICITTE` name in <<value,value>> and
<<variable-assignment,variable assignment>> expression evaluation.

Each generated byte increments the current offset.

A <<current-offset-setting,current offset setting>> may change the
current offset.
|`init_offset` parameter of the `parse()` function.
|`--offset` option.

|[[cur-bo]] Current byte order
|
The current byte order has an effect on the encoding of <<value,values>>.

A <<current-byte-order-setting,current byte order setting>> may change
the current byte order.
|`init_byte_order` parameter of the `parse()` function.
|`--byte-order` option.

|<<label,Labels>>
|Mapping of label names to integral values.
|`init_labels` parameter of the `parse()` function.
|One or more `--label` options.

|<<variable-assignment,Variables>>
|Mapping of variable names to integral values.
|`init_variables` parameter of the `parse()` function.
|One or more `--var` options.
|===

The available items are:

* A <<byte-constant,constant integer>> representing a single byte.

* A <<literal-string,literal string>> representing a sequence of bytes
  encoding UTF-8, UTF-16, or UTF-32 data.

* A <<current-byte-order-setting,current byte order setting>> (big or
  little endian).

* A <<value,{py3} expression to be evaluated>> as an unsigned or signed
  integer to be encoded on one or more bytes using the current byte
  order.

* A <<current-offset-setting,current offset setting>>.

* A <<label,label>>, that is, a named constant holding the current
  offset.
+
This is similar to an assembly label.

* A <<variable-assignment,variable assignment>> associating a name to
  the integral result of an evaluated {py3} expression.

* A <<group,group>>, that is, a scoped sequence of items.

Moreover, you can <<repetition,repeat>> any item above, except an offset
or a label, a given number of times. This is called a repetition.

A Normand comment may exist:

* Between items, possibly within a group.
* Between the nibbles of a constant hexadecimal byte.
* Between the bits of a constant binary byte.
* Between the last item and the ``pass:[*]`` character of a repetition,
  and between that ``pass:[*]`` character and the following number.

A comment is anything between two ``pass:[#]`` characters on the same
line, or from ``pass:[#]`` until the end of the line. Whitespaces and
the following symbol characters are also considered comments where a
comment may exist:

----
! @ / \ ? & : ; . , + [ ] _ = | -
----

The latter serve to improve readability so that you may write, for
example, a MAC address or a UUID as is.

You can test the examples of this section with the `normand`
<<command-line-tool,command-line tool>> as such:

----
$ normand file | hexdump -C
----

where `file` is the name of a file containing the Normand input.

=== Byte constant

A _byte constant_ represents a single byte.

A byte constant is:

Hexadecimal form::
    Two consecutive hexits.

Decimal form::
    A decimal number after the `$` prefix.

Binary form::
    Eight bits after the `%` prefix.

====
Input:

----
ab cd [3d 8F] CC
----

Output:

----
ab cd 3d 8f cc
----
====

====
Input:

----
$192 %1100/0011 $ -77
----

Output:

----
c0 c3 b3
----
====

====
Input:

----
58f64689-6316-4d55-8a1a-04cada366172
fe80::6257:18ff:fea3:4229
----

Output:

----
58 f6 46 89 63 16 4d 55  8a 1a 04 ca da 36 61 72  ┆ X•F•c•MU•••••6ar
fe 80 62 57 18 ff fe a3  42 29                    ┆ ••bW••••B)
----
====

====
Input:

----
%01110011 %01100001 %01101100 %01110101 %01110100
----

Output:

----
73 61 6c 75 74  ┆ salut
----
====

=== Literal string

A _literal string_ represents the UTF-8-, UTF-16-, or UTF-32-encoded
bytes of a string.

The string to encode isn't implicitly null-terminated: use `\0` at the
end of the string to add a null character.

A literal string is:

. **Optional**: one of the following encodings instead of UTF-8:
+
--
[horizontal]
`u16be`:: UTF-16BE.
`u16le`:: UTF-16LE.
`u32be`:: UTF-32BE.
`u32le`:: UTF-32LE.
--

. The ``pass:["]`` prefix.

. A sequence of zero or more characters, possibly containing escape
  sequences.
+
An escape sequence is the ``\`` character followed by one of:
+
--
[horizontal]
`0`:: Null (U+0000)
`a`:: Alert (U+0007)
`b`:: Backspace (U+0008)
`e`:: Escape (U+001B)
`f`:: Form feed (U+000C)
`n`:: End of line (U+000A)
`r`:: Carriage return (U+000D)
`t`:: Character tabulation (U+0009)
`v`:: Line tabulation (U+000B)
``\``:: Reverse solidus (U+005C)
``pass:["]``:: Quotation mark (U+0022)
--

. The ``pass:["]`` suffix.

====
Input:

----
"coucou tout le monde!"
----

Output:

----
63 6f 75 63 6f 75 20 74  6f 75 74 20 6c 65 20 6d  ┆ coucou tout le m
6f 6e 64 65 21                                    ┆ onde!
----
====

====
Input:

----
u16le"I am not young enough to know everything."
----

Output:

----
49 00 20 00 61 00 6d 00  20 00 6e 00 6f 00 74 00  ┆ I• •a•m• •n•o•t•
20 00 79 00 6f 00 75 00  6e 00 67 00 20 00 65 00  ┆  •y•o•u•n•g• •e•
6e 00 6f 00 75 00 67 00  68 00 20 00 74 00 6f 00  ┆ n•o•u•g•h• •t•o•
20 00 6b 00 6e 00 6f 00  77 00 20 00 65 00 76 00  ┆  •k•n•o•w• •e•v•
65 00 72 00 79 00 74 00  68 00 69 00 6e 00 67 00  ┆ e•r•y•t•h•i•n•g•
2e 00                                             ┆ .•
----
====

====
Input:

----
u32be "\"illusion is the first\nof all pleasures\" 🦉"
----

Output:

----
00 00 00 22 00 00 00 69  00 00 00 6c 00 00 00 6c  ┆ •••"•••i•••l•••l
00 00 00 75 00 00 00 73  00 00 00 69 00 00 00 6f  ┆ •••u•••s•••i•••o
00 00 00 6e 00 00 00 20  00 00 00 69 00 00 00 73  ┆ •••n••• •••i•••s
00 00 00 20 00 00 00 74  00 00 00 68 00 00 00 65  ┆ ••• •••t•••h•••e
00 00 00 20 00 00 00 66  00 00 00 69 00 00 00 72  ┆ ••• •••f•••i•••r
00 00 00 73 00 00 00 74  00 00 00 0a 00 00 00 6f  ┆ •••s•••t•••••••o
00 00 00 66 00 00 00 20  00 00 00 61 00 00 00 6c  ┆ •••f••• •••a•••l
00 00 00 6c 00 00 00 20  00 00 00 70 00 00 00 6c  ┆ •••l••• •••p•••l
00 00 00 65 00 00 00 61  00 00 00 73 00 00 00 75  ┆ •••e•••a•••s•••u
00 00 00 72 00 00 00 65  00 00 00 73 00 00 00 22  ┆ •••r•••e•••s•••"
00 00 00 20 00 01 f9 89                           ┆ ••• ••••
----
====

=== Current byte order setting

This special item sets the <<cur-bo,_current byte order_>>.

The two accepted forms are:

[horizontal]
``pass:[{be}]``:: Set the current byte order to big endian.
``pass:[{le}]``:: Set the current byte order to little endian.

=== Value

A _value_ represents a fixed number of bytes encoding an unsigned or
signed integer which is the result of evaluating a {py3} expression
using the <<cur-bo,current byte order>>.

For a value at some source location{nbsp}__**L**__, its {py3} expression
may contain the name of any accessible <<label,label>>, including the
name of a label defined after{nbsp}__**L**__, as well as the name of any
<<variable-assignment,variable>> known at{nbsp}__**L**__.

An accessible label is either:

* Outside of the current <<group,group>>.
* Within the same immediate group (not within a nested group).

In the {py3} expression of a value, the value of the special name
`ICITTE` is the <<cur-offset,current offset>> (before encoding the
value).

A value is:

. The ``pass:[{]`` prefix.

. A valid {py3} expression.

. The `:` character.

. An encoding length in bits amongst `8`, `16`, `24`, `32`, `40`,
  `48`, `56`, and `64`.

. The `}` suffix.

====
Input:

----
{le} {345:16}
{be} {-0xabcd:32}
----

Output:

----
59 01 ff ff 54 33
----
====

====
Input:

----
{be}

# String length in bits
{8 * (str_end - str_beg) : 16}

# String
<str_beg>
  "hello world!"
<str_end>
----

Output:

----
00 60 68 65 6c 6c 6f 20  77 6f 72 6c 64 21  ┆ •`hello world!
----
====

====
Input:

----
{20 - ICITTE : 8} * 10
----

Output:

----
14 13 12 11 10 0f 0e 0d  0c 0b
----
====

=== Current offset setting

This special item sets the <<cur-offset,_current offset_>>.

A current offset setting is:

. The `<` prefix.

. A positive integer (hexadecimal starting with `0x` or `0X` accepted)
  which is the new current offset.

. The `>` suffix.

====
Input:

----
       {ICITTE : 8} * 8
<0x61> {ICITTE : 8} * 8
----

Output:

----
00 01 02 03 04 05 06 07  61 62 63 64 65 66 67 68  ┆ ••••••••abcdefgh
----
====

====
Input:

----
aa bb cc dd <meow> ee ff
<12> 11 22 33 <mix> 44 55
{meow : 8} {mix : 8}
----

Output:

----
aa bb cc dd ee ff 11 22  33 44 55 04 0f  ┆ •••••••"3DU••
----
====

=== Label

A _label_ associates a name to the <<cur-offset,current offset>>.

All the labels of a whole Normand input must have unique names.

A label may not share the name of a <<variable-assignment,variable>>
name.

A label name may not be `ICITTE` (see <<value>> and
<<variable-assignment>> to learn more).

A label is:

. The `<` prefix.

. A valid {py3} name which is not `ICITTE`.

. The `>` suffix.

=== Variable assignment

A _variable assignment_ associates a name to the integral result of an
evaluated {py3} expression.

For a variable assignment at some source location{nbsp}__**L**__, its
{py3} expression may contain the name of any accessible <<label,label>>,
including the name of a label defined after{nbsp}__**L**__, as well as
the name of any variable known at{nbsp}__**L**__.

An accessible label is either:

* Outside of the current <<group,group>>.
* Within the same immediate group (not within a nested group).

A variable name may not be `ICITTE`.

In the {py3} expression of a variable assignment, the special name
`ICITTE` is the <<cur-offset,current offset>>.

A variable is:

. The ``pass:[{]`` prefix.

. A valid {py3} name which is not `ICITTE`.

. The `=` character.

. A valid {py3} expression.

. The `}` suffix.

====
Input:

----
{mix = 101} {le}
{meow = 42} 11 22 {meow:8} 33 {meow = ICITTE + 17}
"yooo" {meow + mix : 16}
----

Output:

----
11 22 2a 33 79 6f 6f 6f  7a 00  ┆ •"*3yoooz•
----
====

=== Group

A _group_ is a scoped sequence of items.

The <<label,labels>> within a group aren't visible outside of it.

The main purpose of a group is to <<repetition,repeat>> more than a
single item.

A group is:

. The `(` prefix.

. Zero or more items.

. The `)` suffix.

====
Input:

----
((aa bb cc) dd () ee) "leclerc"
----

Output:

----
aa bb cc dd ee 6c 65 63  6c 65 72 63  ┆ •••••leclerc
----
====

====
Input:

----
((aa bb cc) * 3 dd ee) * 5
----

Output:

----
aa bb cc aa bb cc aa bb  cc dd ee aa bb cc aa bb
cc aa bb cc dd ee aa bb  cc aa bb cc aa bb cc dd
ee aa bb cc aa bb cc aa  bb cc dd ee aa bb cc aa
bb cc aa bb cc dd ee
----
====

====
Input:

----
{be}
(
  <str_beg> u16le"sébastien diaz" <str_end>
  {ICITTE - str_beg : 8}
  {(end - str_beg) * 5 : 24}
) * 3
<end>
----

Output:

----
73 00 e9 00 62 00 61 00  73 00 74 00 69 00 65 00  ┆ s•••b•a•s•t•i•e•
6e 00 20 00 64 00 69 00  61 00 7a 00 1c 00 01 e0  ┆ n• •d•i•a•z•••••
73 00 e9 00 62 00 61 00  73 00 74 00 69 00 65 00  ┆ s•••b•a•s•t•i•e•
6e 00 20 00 64 00 69 00  61 00 7a 00 1c 00 01 40  ┆ n• •d•i•a•z••••@
73 00 e9 00 62 00 61 00  73 00 74 00 69 00 65 00  ┆ s•••b•a•s•t•i•e•
6e 00 20 00 64 00 69 00  61 00 7a 00 1c 00 00 a0  ┆ n• •d•i•a•z•••••
----
====

=== Repetition

A _repetition_ represents the bytes of an item repeated a given number
of times.

A repetition is:

. Any item.

. The ``pass:[*]`` character.

. A positive integer (hexadecimal starting with `0x` or `0X` accepted)
  which is the number of times to repeat the previous item.

====
Input:

----
{end - ICITTE - 1 : 8} * 0x100 <end>
----

Output:

----
ff fe fd fc fb fa f9 f8  f7 f6 f5 f4 f3 f2 f1 f0  ┆ ••••••••••••••••
ef ee ed ec eb ea e9 e8  e7 e6 e5 e4 e3 e2 e1 e0  ┆ ••••••••••••••••
df de dd dc db da d9 d8  d7 d6 d5 d4 d3 d2 d1 d0  ┆ ••••••••••••••••
cf ce cd cc cb ca c9 c8  c7 c6 c5 c4 c3 c2 c1 c0  ┆ ••••••••••••••••
bf be bd bc bb ba b9 b8  b7 b6 b5 b4 b3 b2 b1 b0  ┆ ••••••••••••••••
af ae ad ac ab aa a9 a8  a7 a6 a5 a4 a3 a2 a1 a0  ┆ ••••••••••••••••
9f 9e 9d 9c 9b 9a 99 98  97 96 95 94 93 92 91 90  ┆ ••••••••••••••••
8f 8e 8d 8c 8b 8a 89 88  87 86 85 84 83 82 81 80  ┆ ••••••••••••••••
7f 7e 7d 7c 7b 7a 79 78  77 76 75 74 73 72 71 70  ┆ •~}|{zyxwvutsrqp
6f 6e 6d 6c 6b 6a 69 68  67 66 65 64 63 62 61 60  ┆ onmlkjihgfedcba`
5f 5e 5d 5c 5b 5a 59 58  57 56 55 54 53 52 51 50  ┆ _^]\[ZYXWVUTSRQP
4f 4e 4d 4c 4b 4a 49 48  47 46 45 44 43 42 41 40  ┆ ONMLKJIHGFEDCBA@
3f 3e 3d 3c 3b 3a 39 38  37 36 35 34 33 32 31 30  ┆ ?>=<;:9876543210
2f 2e 2d 2c 2b 2a 29 28  27 26 25 24 23 22 21 20  ┆ /.-,+*)('&%$#"!
1f 1e 1d 1c 1b 1a 19 18  17 16 15 14 13 12 11 10  ┆ ••••••••••••••••
0f 0e 0d 0c 0b 0a 09 08  07 06 05 04 03 02 01 00  ┆ ••••••••••••••••
----
====

== Command-line tool

If you <<install-normand,installed>> the `normand` package, then you
can use the `normand` command-line tool:

----
$ normand <<< '"ma gang de malades"' | hexdump -C
----

----
00000000  6d 61 20 67 61 6e 67 20  64 65 20 6d 61 6c 61 64  |ma gang de malad|
00000010  65 73                                             |es|
----

If you copy the `normand.py` module to your own project, then you can
run the module itself:

----
$ python3 -m normand <<< '"ma gang de malades"' | hexdump -C
----

----
00000000  6d 61 20 67 61 6e 67 20  64 65 20 6d 61 6c 61 64  |ma gang de malad|
00000010  65 73                                             |es|
----

Without a path argument, the `normand` tool reads from the standard
input.

The `normand` tool prints the generated binary data to the standard
output.

Various options control the initial <<state,state>> of the processor:
use the `--help` option to learn more.

== {py3} API

The whole `normand` package/module API is:

[source,python]
----
class ByteOrder(enum.Enum):
    # Big endian.
    BE = ...

    # Little endian.
    LE = ...


VarsT = typing.Dict[str, int]


class TextLoc:
    # Line number.
    @property
    def line_no(self) -> int:
        ...

    # Column number.
    @property
    def col_no(self) -> int:
        ...


class ParseError(RuntimeError):
    # Source text location.
    @property
    def text_loc(self) -> TextLoc:
        ...


class ParseResult:
    # Generated data.
    @property
    def data(self) -> bytearray:
        ...

    # Updated variable values.
    @property
    def variables(self) -> VarsT:
        ...

    # Updated main group label values.
    @property
    def labels(self) -> VarsT:
        ...

    # Final offset.
    @property
    def offset(self) -> int:
        ...

    # Final byte order.
    @property
    def byte_order(self) -> typing.Optional[int]:
        ...

def parse(normand: str,
          init_variables: typing.Optional[VarsT] = None,
          init_labels: typing.Optional[VarsT] = None,
          init_offset: int = 0,
          init_byte_order: typing.Optional[ByteOrder] = None) -> ParseResult:
    ...
----

The `normand` parameter is the actual <<learn-normand,Normand input>>
while the other parameters control the initial <<state,state>>.

The `parse()` function raises a `ParseError` instance should it fail to
parse the `normand` string for any reason.