Re: Byte compilation error.

[Christopher Roy Bratusek <zanghar freenet de>]

Am Sat, 05 Jun 2010 09:24:54 -0500
schrieb Jeremy Hankins <nowan nowan org>:

> Teika Kazura <teika lavabit com> writes:
> > On Sat, 29 May 2010 09:06:42 -0500, Jeremy Hankins wrote:
> 
> >> The unfinished part has nothing to do with prompt.jl; I can apply
> >> that anytime we decide it's time.  In fact, the sawfish that I run
> >> myself has this patch applied and has for months.  The problem is
> >> that it breaks backward-compatibility since it changes the way the
> >> prompt code is called.
> >
> > We're not in hurry at all, but if you could do it, I'd appreciate it
> > deeply. Is it easy for you to write a rough documentation on the
> > usage? If it's supplied, then it's easy to update existent codes, so
> > there'll be no problem.
> 
> Hmm..  Currently there is no documentation for prompt, only the source.
> I've been meaning to write up some more thorough docs on using the
> prompt module, maybe in the form of a tutorial.  But here's very rough
> documentation of the changes:
> 
> In the past prompt has worked by allowing a variety of helper functions
> and variables to be over-ridden by let statements at the time that
> prompt is called.  With my changes in place these things are not passed
> implicitly via let statements, but passed explicitly in the call to
> prompt.  Here's an example from prompt-extras.jl that I think captures
> all the changes: the prompt-for-file function presented as a diff
> against the old way of doing things.
> 
>  
> +(define filename-history (prompt-make-history))
>  (define (prompt-for-file #!optional title existing start default)
>    "Prompt for a file, if EXISTING is t only files which exist are
>  allowed to be entered."
> @@ -90,14 +100,17 @@ allowed to be entered."
>    (setq start (if (stringp start)
>                   (expand-file-name start)
>                 (file-name-as-directory default-directory)))
> -  (let* ((prompt-completion-fun prompt-complete-filename)
> -        (prompt-validation-fun (and existing prompt-validate-filename))
> -        (prompt-abbrev-fun prompt-abbreviate-filename)
> -        (str (prompt title start)))
> +  (let ((str (prompt #:title title
> +                     #:start start
> +                     #:completion-fun prompt-complete-filename
> +                     #:validation-fun (and existing prompt-validate-filename)
> +                     #:abbrev-fun prompt-abbreviate-filename
> +                     #:history filename-history)))
>      (when (and (string= str "") default)
>        (setq str default))
>      str))
> 
> 
> The prompt-make-history function is new; it creates a data structure
> suitable for holding prompt history.  By creating a new history value
> specifically for prompt-for-file the prompt history for files is kept
> separately from, e.g., the history of prompts for sawfish commands.
> It's a fluid rather than a special variable (in C terms, it's more like
> a pointer than a global variable), so the reference must be passed to
> prompt when prompt is called.  There is a default history defined in
> prompt.jl so it's not necessary to supply a new history if you don't
> want to.
> 
> The other big difference is the let*/let statement.  The
> prompt-complete-filename, prompt-validate-filename, and
> prompt-abbreviate-filename are helper functions defined elsewhere in
> prompt-extras.jl, they haven't changed.  But in the new version they're
> passed explicitly to prompt, along with title (the message prompting for
> input, e.g., "Enter filename:"), start (any initial text, e.g., "~/")
> and history (which holds the history of filename input).
> 

Do you have a full patch (in your git?), as this looks interesting.

Ah… and yes… I'm b.a.c.k™ :D

Chris

-- 

Re: Schöner abstürzen mit Kernel-Modul provoke_crash
Score: 3 Von Cyril am Sa, 30. Januar 2010 um 09:15

Ich bin mir sicher, Microsoft hat noch aus den Windows 9x-Zeiten ein Patent
auf solche Abstürze, das gibt bestimmt eine Klage. :-D