Switch branches/tags
Nothing to show
Find file History
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Failed to load latest commit information.


Haskell Code and Text Blogging (HaCoTeB)

A. About

Haskell Code and Text Blogging (HaCoTeB) started as a tool which allowed easy posting of Wordpress.com blog posts containing text and source code without relying to online tools or other plugins.

It was intended to transform one single input file into a html fragment to be inserted into the blog editor. However, in time, I've decided to allow patches which will allow converting the same input file into another format, for example into a RST format or a PDF file.

B. Usage

B.1 The input file

B.1.1. Structure of the text file

The text file is divided into sections. The section separator is the first line in the file. That line will be repeated in the file for each end of section.

The only place where the section separator (also called terminator or marker) is optional is at the end of the file (as it is implied that there will be one section in every input file).

There is no problem if two consecutive lines are the section separator. Thus, you can delete a section and leave the marker there without any problem.

For example, this is a valid input file structure:

content of section
content of second section
not how a section was deleted above but it's marker was left in place
this section is the last in the file, we can safely forget about
marking the end of the file.

B.1.2. Structure of a section

There are at least two types of sections from the user point of view: text and code. However, there are 3, in fact, types of them: an empty section (to solve the problem caused by multiple end markers), an anonymous section and a complete one.

A complete section contains a header and a content while the anonymous one lacks the header (and will be parsed as a text section). The empty section is ignored.

The header is the first line in the section, only if it contains text in brackets. The type of the section can only be deduced by looking at the header. See next section for header description.

Depending on header's content, the parses determines if the section is either text or code. Then, it calls the appropriate next level parser to generate the required representation.

Taking this into account, a file would look like this:

Implicit text section. Following is an empty section.
This section has a header, thus it has a type.
[header for contentless section]

B.1.3. Header description

A section header contains text enclosed in brackets. That is, the first character of the line should be an opening bracket for that line to be interpreted as a section.

Each type of section header has two parts: a common part which must be the same across different types of headers and a part giving options to that section's specific parser. The common part is the one contained between the brackets.

That part contains the type of the section and an optional name. If there is a name for the section it must be separated from the type by a colon character (:) and must contain only [a-zA-Z0-9]*. This name can be used to generate local links from one section to another. It will be used to create a named anchor at the beginning of the section.

The type of a section can be any prefix of "text" (for text sections) or "code" (for code sections). Any other form will be considered an error. When other sections will be added, this requirement will change.

The optional part of a header follows after the closing bracket and continues until the end of the line. It will be parsed by the sections' specific parser, not by the global one.

This is an example of a file seen at this level of detail:

Implicit text section, unnamed.
Text section, anonymous.
Code section, named first.

B.1.4. Text section structure

Text starting from here is WIP. Don't take as granted!

As of this version, each text section is outputted verbatim in a paragraph tag with a justified text-align style.

Right now, there are no extra arguments and section names are ignored.

B.2. The command line


C. Extending HaCoTeB