Release v0.11.1

v0.11.1 Improved tests by capturing stderr

In previous versions, the warnings emitted by acronymsdown were directly
printed to the terminal.
This messed up the reporter of 'testthat' (which expects to control what
is printed or not; warnings were printed between tests lines...).

Also, we could not run tests on these warnings.
For example, when a key is missing, it is perfectly normal for
acronymsdown to print a warning (it is even expected).

This version brings a new capture-error behaviour to solve these
problems. The capture-error is only enabled if a specific shared library
is compiled (the newly introduced dup2_helper).

End users do not need to compile this library: it is only useful for
testing purposes. Acronymsdown runs will full features without it.

Release v0.11

v0.11 New custom output format, more tests

The main addition of this new version is the 'acronymsdown::md_format'
output format, which is based on 'rmarkdown::md_document'.
This format includes a few default pandoc parameters to help us run
automated tests.

In particular, it allows specifying parameters programmatically (at
runtime), rather than hardcoding them into the Rmd files.
This is particularly useful for the '--atx-headers' parameter, which
ensures that we obtain the same output on older versions of pandoc
(pre 2.11.2), but was deprecated and generates warnings if we include
it by default.

This custom format allowed us to test acronymsdown on more versions
of pandoc (we were limited to >= 2.11.2 previously).

Release v0.10.4

Fixed the release workflow

Release v0.10.2

Bugfix for some old pandoc versions

Newest versions of pandoc have introduced a breaking change that
we must take into account. However, the version detection
mechanism itself was not working on older versions of pandoc.

This new version of acronymsdown introduces a new mechanism to
correctly detect the version of pandoc, including on older
versions.

Release v0.10.0

v0.10.0 Option 'insert_loa'

* Refactored the 'insert_beginning' option into a new 'insert_loa'.
This option has a more intuitive name, and offers more behaviours
('beginning', 'end', 'false' vs 'true', 'false').
* Added 3 tests for the various behaviours of this option.
* Updated documentation.

Release v0.9.1

v0.9.1 Bugfixes, better tests and documentation

This minor version includes:

* A new, better documentation based on pkgdown (using several vignettes to explain the various options and mechanisms, rather than a single Readme).
* An improved 'warn' function in the Lua Filter.
* Less warnings in 'R CMD CHECK'.
* A new workflow to automatically run tests whenever a commit is pushed.
* Several bugfixes to make acronymsdown compatible with pandoc 2.17 and higher.
  The testing workflow revealed that previous versions were not compatible with pandoc 2.17.0.1.
  (This should include pandoc 2.17 as well)

Release v0.9.0

Replaced the "inexisting_keys" option by "non_existing".

Release v0.8.0

Added new option 'insert_links' + various fixes, tests

* Added a new option 'insert_links', which controls
whether to insert a link to the acronym's definition,
when replacing an acronym.
The default behaviour, when the option is unspecified,
is 'true', i.e., the same as previously: links are
automatically inserted.
An important use-case for this option is when the
List of Acronyms itself is disabled: links would
not work.

* Refactored the way tests are launched by 'testthat',
such that a better, more readable, report is printed.

* Added the 'wrap=preserve' Pandoc argument to the
tests, which makes Pandoc keep the line breaks as-is
from the input file. This makes it easier to write
expected files in the tests, since we only have to
focus on the content itself, and not worry about
where we should insert a line break to be consistent
with Pandoc.

* Changed the return value of the 'add_filter()'
function in the R code, from a vector of 2 strings
('--lua-filter', '/path/to/parse-acronyms.lua')
to a single string
('--lua-filter=path/to/parse-acronyms.lua').
This allows adding other Pandoc arguments in the YAML
metadata, by using a list for example.

* Improved documentation for 'add_filter()',
please see the help for more details
('?acronymsdown::add_filter' in a R console).

* Fixed a bug in the error handling of a non-existing
acronym key, where a nil value was written to
stderr.

Release v0.7.0

v0.7.0 Added ability to choose a style

Instead of always replacing with "long name (short name)"
for the first occurrence, or "short name" for the next occurrences,
we now propose several styles, inspired from the glossaries
LaTeX package.
This is controlled by the new "style" option.

Release v0.6.0

v0.6.0 Added ability to load acronyms from external files

* Files that contain acronyms' definitions can be specified
  using the 'acronyms.fromfile' field in the YAML metadata.
* The acronyms' database merges definitions from the
  metadata in the Markdown document itself, and
  optional external files.
* These external files have a so-called YAML structure,
  but are in reality treated as Markdown, in which
  only the metadata is kept.
* Added tests 06 and 07 which correspond to this feature.