Re: lynx-dev HTSprintf0 (was lynx 2.8.3dev.4 for Win32 patch)

[T.E.Dickey]

> Btw. my libc6 has functions that already do the basically the same thing 
> as HTSprintf0: 
>  
>      [ #include <stdio.h> ] 
>        int asprintf(char **strp, const char *format, ...); 
>  
> and 
>      [ #include <stdarg.h> ] 
>        int  vasprintf(char  **strp,  const  char *format, va_list 
>        ap); 
>  
> I don't know how widespread they are or whether they are GNU inventions. 
> If they are available on a system they should probably be faster. 
> [ I've actually replaced HTSprintf0 (not HTSprintf) with this a while 
> ago and it seemed to work here.  Some special considerations were 
> necessary for --enable-find-leaks.  It may not be worth the trouble, 
> since it won't cover all systems. ] 

I think it's only glibc2 - someone asked for help on a problem recently
where he'd used it, and (as part of making the program more portable so
I could test it on Solaris) I couldn't find other systems that used it.

So I'm aware of it (perhaps asprintf is faster -- if we were to measure
it and find that it's appreciably faster w/o any obvious way to polish
HTSprintf0, then it would be ok to conditionally build with it -- bear
in mind that if it's not standardized, it'll end up like the various
incompatible versions of snprintf).

>    Klaus 


-- 
Thomas E. Dickey
address@hidden
http://www.clark.net/pub/dickey