Acceleo Text Production Rules

Authors: Laurent Goubet
Laurent Delaigue
Contact: laurent.goubet@obeo.fr, laurent.delaigue@obeo.fr

Copyright © 2008, 2010 Obeo™.

Contents

Overview

This is still a draft and is only provided as information. Though this document is still missing some formatting and examples, it can be used to get a precise idea as to how indentation and carriage returns are handled in Acceleo.

Definitions

Text production rules will apply to all body elements. Each body element can be either stand alone or embedded within other body elements. It will be easier to understand text production rules by splitting these body elements in five main categories : expressions, static text, comments, template invocation and blocks.

"Blocks" include template, for, if, let, protected area, file, trace and macro blocks. A block will be considered « stand-alone » if it is either one of:

Any block that does not fall into these categories will be considered an "embedded" block.

Examples

Stand-alone single-line blocks

Nothing relevant for generation between the line start and the block head, nothing relevant between the block tail and the line end; both head and tail on the same line. All of the following are in this category.

[for (Sequence{false, false, false})][self/][/for]

[if (true)]output[/if]

[for (Sequence{false, false, false})][self/][/for]

[/if]         [if (true)]output[/if]

[/if] [comment .../]  [if (true)]output[/if]

Embedded single-line blocks

Head and tail of the block are on the same line, but there is something relevant to the generation other than the block on that same line.

Some text[for (Sequence{false, false, false})]output[/for]

[if (true)]output[/if] and some text

[for (Sequence{false, false, false})]   [self/][/for] and some text.

Stand-alone multi-line blocks

Nothing relevant for generation between the line start and the block head, nothing relevant between the block tail and the line end; head and tail on different lines.

[for (Sequence{false, false, false})]
    [self/]
[/for]

[if (true)]output
[/if]

[if (true)]
    output[/if]

[comment]Generate booleans [/comment][for (Sequence{false, false, false})]
    [self/][/for]

[for (Sequence{false, false, false})]
[self/]
[/for]

Embedded multi-line blocks

Head and tail of the block are on different lines, but there is something relevant to the generation other than the block on that same line.

text[for (Sequence{false, false, false})]
    [self/]
[/for]

[if (true)]
    output[/if] text

Identifying Body Element Boundaries

Block Body

Single-line block
Body starts after the closing bracket of the block head and ends before the starting bracket of the block tail.
Multi-line block
If the closing bracket of the block head is directly followed by a new line, the block's body starts at the beginning of the next line after the block head. Otherwise the block's body starts after the closing bracket of its head. The body ends before the starting bracket of the block tail, be it directly preceded by a new line or not.
Special handling of "Template" blocks
If the last characters preceding the starting bracket of the block tail is a new line followed by optional white spaces, the template body ends before the last new line character preceding its tail.

Static Text

We need to define the concept of line relevance to properly identify these boundaries. For this purpose, we will describe as "white spaces" the white space characters contained in a static text, whatever their position in the text.

A line is considered "relevant" if it contains anything else than white spaces, block head, block tail and comment. Note, however, that lines consisting of white spaces only are also considered relevant.

  • If a static text is entirely located on a non relevant line, it does not produce any text.
  • If a static text starts on a non relevant line, all characters located on this line are ignored and the static text is considered to start with the character after its very first new line character.
  • If a static text ends on a non relevant line, all characters located on this line are ignored and the static text is considered to end with the character before its very last new line character.
  • Special handling of static text following a « protected area » block: all white spaces following a protected area tail are retained, including the new line character. The static text then starts right after the closing bracket of the protected area tail.

Rules

Considering the boundaries outlined in section Identifying Body Element Boundaries above, the text production rules stand as follows:

Examples

In the following examples, invisible characters have been materialized:

· indicates a space character
indicates a horizontal tab character
indicates a carriage return

Blocks

Embedded Blocks

Example 1

The Acceleo code:

Some · text · [for · (Sequence{false, · false, · false}) · separator · (' · - · ')]output[/for]

produces the following result:

Some · text · output · - · output · - · output

Example 2

The Acceleo code:

[if · (true)]output[/if] · and · some · text

produces:

output · and · some · text

Example 3

The Acceleo code:

[for · (Sequence{false, · false})] · [self/][/for] · and · some · text

produces:

↪ · false · false · and · some · text

Example 4

The Acceleo code:

text[for · (Sequence{false, · false})]

[self/]

[/for]text

produces:

textfalse

false

text


Example 5

The Acceleo code:

text[for · (Sequence{false, · false})]

output[/for]text

produces:

textoutputoutputtext

Stand-alone single-line blocks

Example 1

The Acceleo code:

[for · (Sequence{false, · false, · false})][self/][/for]

produces:

falsefalsefalse

Example 2

The Acceleo code:

[if · (true)]output[/if]

produces:

output

Example 3

The Acceleo code:

[for · (Sequence{false, · false, · false})][self/][/for]

produces:

falsefalsefalse

Example 4

The Acceleo code:

[/if] ↪ ↪ [if · (true)]output[/if]

produces:

↪ ↪ output

Example 5

The Acceleo code:

[/if] · [comment .../][if · (true)]output[/if]

produces:

· ↪ output

Stand-alone multi-line blocks

Example 1

The Acceleo code:

[for · (Sequence{false, · false, · false})]

[self/]

[/for]

produces:

false

false

false


Example 2

The Acceleo code:

[if · (false)]output

[/if]

produces nothing, neither whitespace nor empty line.


Example 3

The Acceleo code:

[if · (true)]

output[/if]

produces:

output

Example 4

The Acceleo code:

[comment]for · loop[/comment][for · (Sequence{false, · false})]

[if · (self)]

↪ ↪ [self/]

[/if]

[/for]

produces nothing, neither whitespace nor empty line.


Example 5

The Acceleo code:

[for · (Sequence{false, · false, · false})]

[self/]

[/for]

produces:

false

false

false