how can I identify a single line no longer than e.g. 60 characters
preceded and followed by a blank line via regexs. This way I want to
identify section headings. What I did was mark every blank line with
%s/^$/>>>/ and than chomp the CR %s/\n//g and if the text
between the >>>'s isn't longer than 60 characters put it into
\section{}, and replace every >>> with \r\r. But in larger files
this takes a while. Is there a smarter solution to the problem?
You might try:
/^\s*\n\zs.\{1,60}\ze\n\s*$
/^\n\zs.\{1,60}\ze\n$
which should find what you describe (the former allows a "blank"
line to contain whitespace while the latter looks for a truely
"blank" line with zero whitespace).
It breaks down as
^ the start of line
\s* (the optional whitespace, as described)
\n a newline
\zs treat this as the beginning of the match
.\{1,60} 1-60 non-newline characters
\ze you could optionally use this here
to treat this as the end of the match
\n another newline
\s* (more optional whitespace, as described)
$ the end of a line
The above regexp(s) (with or without the "\s*" or the "\ze") can
also be used in ":s" commands to do things like
:%s/^\s*\n\zs.\{1,60}\ze\n\s*$/\U&
to uppercase all your titles, or
:%s/^\s*\n\zs.\{1,60}\ze\n\s*$/\=submatch(0)."\n".substitute(submatch(0),
'.', '=', 'g')
which will underline all your hits with a row of "=" the same
length as your section-title.
Just a few ideas,
-tim
