Re: PATCH: zasprintf

["Bart Schaefer" <schaefer@xxxxxxxxxxxxxxxxxxxxxxx>]

On Sep 17, 11:04am, Clint Adams wrote:
} Subject: Re: PATCH: zasprintf
}
} > I don't believe there's any use zsh could make of asprintf()
} > that isn't better solved another way.
} 
} If by "better" you mean "avoiding asprintf and dealing with a specialized
} case."

I do, mostly because I don't think there are very many specialized cases.
There are 96 uses of sprintf() in zsh, nearly all of them used to convert
integers to strings (which hardly requires a dynamically allocated buffer).
Several of the remaining ones are used for path concatenation.  None have
particularly complex format strings.

I simply think that asprintf() is over-generalized.

} > I don't have any such confidence that asprintf() would be cleaner or more
} > efficient than, say, a realloc'ing version of tricat()
} 
} Why, other than portability?

For one thing, because it can't realloc().  That means that if you use it
in a loop, you have to allocate and free the buffer every time around,
even if you're always writing approximately the same size string.

For another thing:

} Speaking of the real tricat, this eliminates bin_dot's dependence on
} PATH_MAX:

This is a case where a zsh-heap-allocating version of tricat() would be
slightly preferable to a permanent-heap-allocating version.

-- 
Bart Schaefer                                 Brass Lantern Enterprises
http://www.well.com/user/barts              http://www.brasslantern.com

Zsh: http://www.zsh.org | PHPerl Project: http://phperl.sourceforge.net