#!/bin/bash
# Generate HTML from post Markdown source
#
#  Copyright (C) 2019 Mike Gerwitz
#
#  This program is free software: you can redistribute it and/or modify
#  it under the terms of the GNU General Public License as published by
#  the Free Software Foundation, either version 3 of the License, or
#  (at your option) any later version.
#
#  This program is distributed in the hope that it will be useful,
#  but WITHOUT ANY WARRANTY; without even the implied warranty of
#  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
#  GNU General Public License for more details.
#
#  You should have received a copy of the GNU General Public License
#  along with this program.  If not, see <http://www.gnu.org/licenses/>.
#
# This script accepts the file name rather than data on stdin because the
# filename encodes the post date.
#
# Note that the `pagetitle' is set to "ignoreme"---it is not used, but is
# needed to suppress the warning pandoc produces without suppressing all
# warnings.
#
# Pandoc is used to generate the HTML and includes a (mostly) static header
# and footer.  Note that this duplicates the date logic in `post2meta',
# because that must be run on this output, but the post must also contain
# the date, and we want to do all HTML processing now.
##

declare -r srcref=https://forge.mikegerwitz.com/mikegerwitz/thoughts/commits/branch/master

set -euo pipefail

# Pandoc output format and extensions.
declare -ra ext=(
  markdown
  smart
  footnotes
  gfm_auto_identifiers
  fancy_lists
  startnum
  tex_math_dollars
)


# Convert extensions to `+'-delimited string.
pexts()
{
  local IFS=+
  echo "${ext[*]}"
}


# Wrap h1 in an hgroup along with the post date.
#
# Sometimes this script is used on things that aren't posts (e.g. normal
# pages), in which case a date will be unavailable and the output will be
# unchanged.
hgroup-wrap()
{
  local -r date=${1?Missing date}
  local -r file=${2?Missing file}

  # Abort if this is not a date prefix
  [[ $date =~ [0-9]{4}-[0-9]{2}-[0-9]{2} ]] || {
    cat
    return
  }

  local -r repo_href="$srcref/$file"

  local anchor
  printf -vanchor '<a class="muted" href="%s">%s</a>' \
         "$repo_href" \
         "$date"

  sed '/^<h1/{
         i<hgroup>
         a<h2 class="date">'"$anchor"'</h2></hgroup>
       }'
}


# Pre-format Markdown files before they get to Pandoc
#
# These may be able to be implemented as Pandoc filters, but I haven't had
# the time to research that yet.  This is actually a fitting real-world
# demonstration of incremental development / MVP that I'm writing about at
# the time that this comment was written (MyCustomBB Part I)!
prefmt()
{
  awk '
    triml {
      gsub( /^ +/, "" )
      triml = 0
    }

    # ties
    { $0 = gensub( /([^\\])~/, "\\1 ", "g" ) }
    { $0 = gensub( /\\~/, "~", "g" ) }

    # TeX-style newline removal
    /%$/ {
      gsub( /%$/, "" )
      printf "%s", $0
      triml = 1
      next
    }

    # reference to the content of the last fence (e.g. if a previous
    # code block is HTML and we want to render it as an example)
    /^```/ {
      gather = !gather
      if ( gather ) gblock = ""
    }
    !/^```/ && gather { gblock = gblock $0 "\n" }
    !gather && /^ *@LASTFENCE@$/ { print gblock; next }

    { print }
  '
}


# Generate HTML from post.  Note that `pagetitle' is set just to suppress
# Pandoc warnings about it missing; it is unused.
main()
{
  local -r file=${1?Missing file name}
  local -r base=$( basename "$file" .md )
  local -r date=${base:0:10}

  pandoc -f"$( pexts )" -thtml5 \
         --standalone --template src/pandoc.tpl \
         --metadata pagetitle:ignoreme \
         --base-header-level=1 \
         -B <( src/mkheader post @__PAGE_TITLE__@ ) \
         -A src/footer.tpl.htm \
         < <( prefmt < "$file" ) \
    | src/h12title @__PAGE_TITLE__@ \
    | hgroup-wrap "$date" "$file"
}


main "$@"