example from ch.16 "learn vimscript the hard way"

user_16711485 0
Viewed 75

I'm trying to complete an exercise from https://learnvimscriptthehardway.stevelosh.com/chapters/16.html

The sample text to be worked on is:

Topic One

=========

This is some text about topic one. 

It has multiple paragraphs.


Topic Two

=========

This is some text about topic two. It has only one paragraph.

The mapping to delete the heading of Topic One or Topic Two (depending on which body the cursor is placed in) and enter insert mode is:

:onoremap ih :<c-u>execute "normal! ?^==\\+$\r:nohlsearch\rkvg_"<cr>

Enter 'cih' in the body of either text below the headings and respective heading will be erased and the cursor will be placed there ready to go, in insert mode. Great mapping--but, I'm trying to understand what's happening with \+$.

When I omit \+$ and use this mapping:

:onoremap ih :<c-u>execute "normal! ?^==\r:nohlsearch\rkvg_"<cr>

it works fine, seemingly identically to the other mapping. So what is the use of the \+$?

Here is how Mr. Losh explains it:

The first piece,

?^==\+$

performs a search backwards for any line that consists of two or more equal signs and nothing else. This will leave our cursor on the first character of the line of equal signs."

But what does \+$ accomplish? I've tried to enter it manually in command but I just get an error sound. It works as intended as part of the full function, though. but like I said, when I remove it and run the full command without, it works fine.

There's something I'm missing about the necessity of that '+$'... Maybe it has to do with the "two or more equal signs and nothing else"?

vim
2 Answers

The author's command:

?^==\+$

searches backward for a line consisting exclusively of 2 or more equal signs:

  • ^ anchors the pattern to the beginning of the line,
  • = matches a literal equal sign,
  • ^= thus matches a literal equal sign at the beginning of the line,
  • = matches a second equal sign,
  • \+ matches one or more of the preceding atom, as many as possible,
  • =\+ thus matches one or more equal sign, as many as possible,
  • $ anchors the pattern to the end of the line,

so the pattern above is going to match any of the following lines:

==
===
=============
etc.

but not lines like:

==foo
==      <- six spaces
etc.

which is exactly the goal of that exercice.

Your command, on the other hand:

?^==

searches backward for a sequence of two equal signs at the beginning of a line:

  • ^ anchors the pattern to the beginning of the line,
  • == matches two literal equal signs,

so your pattern is going to match the same lines as above:

==
===
=============
etc.

but also lines like:

==foo
==      <- six spaces
etc.

because it is not strict enough.

Your pattern would definitely be good enough if used manually to jump to one of those underlines because it gets the job done with minimal typing. But the goal, here, is to make a mapping. Those things have to be generalised to be reliable, which pretty much requires a level of explicitness and precision your pattern lacks.

In short, Steve's pattern checks all the boxes while yours doesn't: it is explicit and precise while yours is implicit and imprecise.

user_5468610

The \+$ is part of the regular expression matching a line of only equals signs. Without it, your mapping would recognize, for example,

This is not a heading
=This is not an underline

as a heading.

The \+ means "At least two of the previous character (=)". The $ means End of line, so there cannot be anything after the equals signs.

user_40153720
Related