Re: Portable replacement function for `trunc'?

[Benoit SIGOURE]

On Oct 4, 2007, at 3:21 AM, Bruno Haible wrote:

> Hello Benoit,
>
> > I was looking for a portable replacement of `trunc' for systems where
> > it is not be available.  I saw that gnulib had that (http://
> > www.gnu.org/software/gnulib/manual/html_node/trunc.html)
>
> Well, there is a misunderstanding here. That doc says that
> 1) gnulib has no module supporting this function,
> 2) all known portability problems are in the "not fixed by gnulib"  
> category.

Sorry but the doc is very misleading then.  Here is what happened to  
me: I went to the gnulib manual to find whether it has a replacement  
for `trunc'.  I ended up on the page above by simply search for  
"trunc" in the index page.  That page is located in the section  
"Function Substitutes" which starts by stating "This chapter  
describes which functions [..] are substituted by Gnulib".  The fact  
that it said "Gnulib module: —" sounded weird to me, but I found that  
since this was a single and simple function, it could be that there  
is no need to make a whole module around it.  The "Portability  
problems not fixed by Gnulib:" simply told me that the replacement  
function wasn't working on the architectures listed but since I don't  
target these for the time being, it didn't disturb me.

If the function is not available, why is it there in the first  
place?  If the intent of the doc is to list all functions with  
possible portability issues (sic!), why not write something more  
explicit such as "Gnulib module: @emph{not available}."?

> > invoked path/to/gnulib-tool --import trunc.  It did what it was meant
> > to do but there was still no trace of `trunc' in my source tree
>
> Now wonder...
>
> But since trunc() is a particularly easy function to implement, I  
> implemented
> this now. Also updating the doc.

I'll give it a try, thanks.

--
Benoit Sigoure aka Tsuna
EPITA Research and Development Laboratory

PGP.sig
Description: This is a digitally signed message part