I did a half-hearted port from Markdown to Orgmode for the underlying text of this site. And I also did a half-hearted port to using Hugo page bundles (in anticipating of getting webmentions working).
But the time has come for me to fix things up properly.
First of all, I need to covert my text files from being formated using markdown to be using Orgmode. Enter pandoc.
But that mucks up my frontmatter (that's really important to Hugo). So, enter sed.
The two scripts that do the heavy lifting, are a shell script ConvertToOrg.sh
for INFILE in $(grep -L "^#+date" */*.org)
do
echo "Covertingt from Markdown: " $INFILE
export OUTFILE=$INFILE.txt
pandoc -f markdown -t org -o $OUTFILE $INFILE
echo "Fixing frontmatter in: " $OUTFILE
sed -f fixFrontmatter.sed $OUTFILE > $INFILE
done
and the sed script, fixFrontmatter.sed, looks like this:
/./{H;$!d} ;
x;
/+++/,2{
s/^[[:space:]]//;
s/[[:space:]]/ /g;
s/\+\+\+// ;
s/ *title = "\([^"]*\)"/#+title: \1\n/ ;
s/ *date = "\([^"]*\)"/#+date: \1\n#+lastmod: \1\n/ ;
s/ *categories[^"]*"\([^"]*\)"]/#+categories[]: \1\n#+tags[]: \n/ ;
s/ *draft = "\([^"]*\)"/#+draft: \1\n/ ;
s/ +++ /\n/ ;
}
/:PROPERTIES:/d
/:END:/d
/:CUSTOM_ID:/d
Getting the multi-line replaces working in sed took some time, but I've come to kinda of like sed now.
All in all, it's all a bit of a kludge and it's not 100%. I'm now going through all my old posts to check and touch-up the converted files.
Update 2021-12-30 Thu
This update to my site has also see me strip out the pages that detail my publications, my teaching (courses), and the details of my individual runs.
That'll cause a little link rot. But I doubt many/any folk will notices.
