I found myself writing a parser with lots of matches in sequence, and
wanted something a bit smarter. This adds two bits of smarts: a shorter
name and string handling makes something like @seq(m('{'), @string,
m('}'))[1] much easier to read and debug.