Re: [PATCH] userdiff: add support for Fountain documents

From: Zoë Blade <zoe@bytenoise.co.uk>
Date: Jul 19, 2015, 12:30 UTC
Message-ID: <562D4AFF-BFC1-4234-8362-656D1F54DFD0@bytenoise.co.uk>
In-Reply-To: <xmqq380m1k3g.fsf@gitster.dls.corp.google.com>

On 17 Jul 2015, at 23:43, Junio C Hamano <gitster@pobox.com> wrote:

Show 31 quoted lines

> * Although uppercase is recommended for Scene Headings to increase
>   readability, it is not required.
> 
> * A line beginning with any of the following, followed by either a
>   dot or a space, is considered a Scene Heading (unless the line is
>   preceded by an exclamation point !). Case insensitive.
> 
>      INT
>      EXT
>      EST
>      INT./EXT
>      INT/EXT
>      I/E
> 
> * You can "force" a Scene Heading by starting the line with a
>   single period.
> 
> * Scene Headings can optionally be appended with Scene
>   Numbers. Scene numbers are any alphanumerics (plus dashes and
>   periods), wrapped in #.
> 
> So, it appears wrong to insist on capital letters in the patterns.
> The pattern in the patch does not even accept punctuations on the
> line other than apostrophe.  I won't judge if it is OK to limit to
> US-ASCII ;-)
> 
> IPATTERNS("fountain",
>    "^([.][^.]|(INT|EXT|EST|INT./EXT|INT/EXT|I/E)[. ]",
>    "[^ \t-]+"),
> 
> or something like this, perhaps?

Good points, thanks!

This regex should be a bit sturdier:

$ cat scenes.txt int. yes - day INT. YES - DAY #1A# EXT. YES - DAY .YES TOO !EXT. NO INT/EXT YES - DAY INT./EXT YES - DAY I/E YES - DAY no NO NO. !.NO. int yes - day est yes - day !EXT. NO - DAY

$ grep -E "^((\.|(([Ii][Nn][Tt]|[Ee][Ss][Tt]|[Ee][Xx][Tt])?\\.?|[Ii]([Nn][Tt])?\.?/[Ee]([Xx][Tt])?\.?) ).+)$" scenes.txt int. yes - day INT. YES - DAY #1A# EXT. YES - DAY .YES TOO INT/EXT YES - DAY INT./EXT YES - DAY I/E YES - DAY int yes - day est yes - day

Revised version of patch incoming...