Running filters consecutively

[avidseeker]

Here is a defaults file I use for proofreading with scripts taken from https://github.com/pandoc/lua-filters

from: markdown
to: markdown
output-file: '/dev/null'
verbosity: INFO
filters:
  - spell.lua
  - wc.lua

• Expected: Show output of spell.lua, then show wordcount.
• Result: Only output of the first filter I put is shown
• Note: both filters work individually, but when I put then consecutively only the first works.

[tarleb]

Both filters explicitly call os.exit(0), i.e., they cause pandoc to terminate. That means that those filters cannot be used in combination like that. You can think of them as custom writers. In fact, with today's pandoc capabilities, it would be conceptually cleaner to rewrite them as writers.

However, the word-count filter has a way to allow further processing by setting:

metadata:
  wordcount: process-anyway

With that you can switch the filter order to ['wc.lua', 'spell.lua'] and then have both filters run consecutively.