From: Johannes Schindelin <johannes.schindelin@gmx.de>
Date: Tue, 06 Nov 2007 23:45:41 GMT
Subject: Re: [PATCH 3/3] pretty=format: Avoid some expensive calculations when not needed
Message-ID: <Pine.LNX.4.64.0711062343050.4362@racer.site>
In-Reply-To: <4730F5FA.3030705@lsrfire.ath.cx>

Hi,

On Wed, 7 Nov 2007, René Scharfe wrote:

> By the way, the more intrusive surgery required when using strbuf_expand()
> leads to even faster operation.  Here my measurements of most of Paul's
> test cases (best of three runs):
>
> [...]

impressive timings.  Although I wonder where the time comes from, as the 
other substitutions should not be _that_ expensive.

In any case, your approach seems much more sensible, now that we have 
strbuf.

> diff --git a/strbuf.h b/strbuf.h
> index cd7f295..95071d5 100644
> --- a/strbuf.h
> +++ b/strbuf.h
> @@ -102,6 +102,9 @@ static inline void strbuf_addbuf(struct strbuf *sb, struct strbuf *sb2) {
>  	strbuf_add(sb, sb2->buf, sb2->len);
>  }
>  
> +typedef void (*expand_fn_t) (struct strbuf *sb, const char *placeholder, void *context);
> +extern void strbuf_expand(struct strbuf *sb, const char *fmt, const char **placeholders, expand_fn_t fn, void *context);

I wonder if it would even faster (but maybe not half as readable) if 
expand_fd_t got the placeholder_index instead of the placeholder.

Ciao,
Dscho