Emacs and Sed things I tend to forget

notes-on-tools.md

Emacs

• Good summary on special characters in Emacs minibuffers and in general: http://ergoemacs.org/emacs/emacs_line_ending_char.html
• On replacement by newline: https://stackoverflow.com/a/613029
• toggle-truncate-lines and visual-line-mode
• Join particular windows: C-u C-x 0
• Org mode table commands: https://orgmode.org/manual/Built_002din-Table-Editor.html
  • in particular M + S + Left to delete a column

sed

• removing text between two patternsstart and end (inclusive): sed '/start:/,/end/d' to leave the lines with delimiters in place, sed '/start:/,/end/{/start/!{/end/!d;};}' https://stackoverflow.com/a/6287940
• How sed works at all: https://www.gnu.org/software/sed/manual/html_node/Execution-Cycle.html#Execution-Cycle
• Multi-line technique, an example:

grep -r -e 'deriving.*Generic' DA/ASX/ -B1 | \
  sed -e '/--/d' \
      -e '$!N;s@\(DA/ASX/[^.]*\.daml\)-\([^\n]*\)\n\1:\(.*\)$@\1 --- \2 \3@g' \
      -e 's@\(.*\) --- *newtype \(.*\) *= *\2 \(.*\) deriving (\(.*\))$@\1, \2, \3, \4@g'

1. grep for things that derive Generic (and other stuff), with one line before as context
2. remove -- from between matches
3. apply multiline match
   • N to take the next line into the buffer
   • then match the file name (for context with -, then again for match with :) and the line contents,
   • and print the file name \1 and the two lines \2 \3
4. then do a match on the now single-line occurrence of the deriving