We now pull the filtered blocks and speaker notes out at the top of
the `blocksToSlide` function, and then make SpeakerNotes into a
parameter of the `blocksToSlide'` subfunction. The output is the same,
but the logic should be easier to follow now.
We used to keep a map of the slideId-to-notes for each slide. Since we
now extract them at the slide level, this is overcomplicated, and we
can just extract them before converting a slide and then clear the
state after.
Previously, we had treated it as a `Maybe`. But there is no difference
between not having speaker notes and having empty speaker notes. So we
make the SpeakerNotes newtype into a monoid, and test for memptiness
instead of testing for Just/Nothing.
Internal change: We were getting bad results with the empty text box
created by the conversion of notes into an empty paragraph. So we now
convert the notes at the `blocksToSlide` stage, by `walkM`ing a
function over the blocks that returns `()`, and then filters the notes
out. This avoids the need to have a return value for speaker notes,
and thus avoids the empty TextBox.
Together with the previous commit (256f14c5a), closes#4477.
Some blocks automatically split slides (imgs, tables, `column`
divs). We assume that any speaker notes immediately following these
are connected to these elements, and keep them with the related
blocks, splitting after them.
* Annotate gridTable code with comments and abstract small functions
* Don't wrap lines in tables when `--wrap=none`. Instead, expand cells, even if
it results in cells that don't respect relative widths or surpass page column width.
* This change affects RST, Markdown, and Haddock writers.
In the reader, we use the new Table type in Haddock.
Note that tables with col/rowspans will not translate
well into Pandoc.
In the writer, we now render tables always as grid tables,
since Haddock supports these.
and not in the title. If it's in the title, then we get
a titlebar on slides with the `plain` attribute, when
the id is non-null. This fixes a regression from 1.9.x.
Closes#4307.
This patch fixes some cases where the JATS writer was introducing
semantically significant whitespace by indenting and wrapping tags.
Note that the JATS spec has a content model for `<p>` tags of `(#PCDATA | ...`.
Any tag where `#PCDATA` children are possible should not have any
indentation. The same is true for `<th>`, `<td>`, `<term>`, `<label>`.
Make sure there are no empty slides in the pptx output. Because of the
way that slides were split, these could be accidentally produced by
comments after images.
When animations are added, there will be a way to add an empty slide
with either incremental lists or pauses.
Test outputs checked with MS PowerPoint (Office 2013, Windows 10,
VBox). Both files have expected output and are not corrupted.
The name of the Lua script which is executed is made available in the
global Lua variable `PANDOC_SCRIPT_FILE`, both for Lua filters and
custom writers.
Closes: #4393
