Re: [dev] [sandy] [PATCH] VIM key bindings.

From: Maxime Coste <frrrwww_AT_gmail.com>
Date: Thu, 10 Jul 2014 23:26:08 +0100

On Thu, Jul 10, 2014 at 11:45:07PM +0300, Dimitris Zervas wrote:
> On July 10, 2014 11:29:59 PM EEST, Evan Gates <evan.gates_AT_gmail.com> wrote:
> >> I will agree that it's super easy to implement and understand and it
> >covers most needs.
> >> But how about search?
> >> Is it fast?
> >
> >What about structural regular expressions as found in sam that aren't
> >limited to lines?
> >-emg
> Oh, I've got a lot more to learn...
> and how will you pass the whole text?

It is very fast to search or use regexes with that, as long as your algorithm
supports strings in multiple part (PCRE does, I am not familiar enough with
Posix regex API to tell if you can easily use multiple strings in them). But
that is a requirement for any buffer storage strategy besides a big malloc.

In C++, I basically use a std::vector<std::string> so I do not have to care
about maintaining the sizes and capacities myself, but in C a buffer would
be:

struct line_t
{
    char* data;
    size_t size;
    size_t capacity;
};

struct buffer_t
{
    line_t* lines;
    size_t size;
    size_t capacity;
};

Accessing a character is just buffer.lines[line].data[column], iterating
on character is just incrementing column, checking if column == line.size
and if so setting column to 0 and incrementing line.

Cheers,

Maxime.
Received on Fri Jul 11 2014 - 00:26:08 CEST