Re: [PATCH v2 1/2] string: Add strprefix(), and deprecate str_has_prefix()

[Alejandro Colomar <alx+linux-hardening@kernel.org>]

[-- Attachment #1: Type: text/plain, Size: 2753 bytes --]

> Date: 2026-11-08 12:49:08+0100
> From: Alejandro Colomar <alx@kernel.org>
>
> Hi Maciej,
> 
> > Date: 2026-11-08 11:05:51+0000
> > From: "Maciej W. Rozycki" <macro@orcam.me.uk>
> >
> > On Sat, 7 Nov 2026, Alejandro Colomar wrote:
> > 
> > > Hi Maciej,
> > > 
> > > > Date: 2026-11-07 02:44:23+0000
> > > > From: "Maciej W. Rozycki" <macro@orcam.me.uk>
> > > >
> > > > On Thu, 5 Nov 2026, Alejandro Colomar wrote:
> > > > 
> > > > > Do you think strafterprefix() would be better?  Do you have any other
> > > > > suggestions?
> > > > 
> > > >  FWIW strsuffix() seems an obvious choice to me.  After all what the API 
> > > > does is request the string's suffix beyond the specified leading part.
> > > 
> > > Actually, strsuffix() is something else:
> > > 
> > > 	char *
> > > 	strsuffix_(const char *str, const char *suffix)
> > > 	{
> > > 		ptrdiff_t  off;
> > > 
> > > 		off = strlen(str) - strlen(suffix);
> > > 		if (off < 0)
> > > 			return NULL;
> > > 
> > > 		if (!streq(str + off, suffix))
> > > 			return NULL;
> > > 
> > > 		return str + off;
> > > 	}
> > > 
> > > It finds the start of a suffix in a string.  That is:
> > > 
> > > 	strcmp(strsuffix("foobar", "bar"), "bar") == 0;
> > 
> >  Hmm, I'd have called that strskipto().  If that's not possible for some 
> > reason (I can't see it used in Linux), then strskip() or strskipover() 
> > could be used for this API.

Another issue with the name strskipto() is that the name doesn't make it
clear whether the second argument is a string, or a list of characters.

That is, would strskipto() something similar to strpbrk(3)?
It's not clear from the name.  From strprefix() and strsuffix(), the
meaning of the second argument is clear from the function name.


Cheers,
Alex

> 
> strsuffix() exists in many projects, and in all of them it seems to have
> these semantics (or very similar ones).
> 
> Just look at a Debian code search of those names.
> 
> The strprefix() and strsuffix() pair of functions are widely
> implemented, and while the exact details of the return value differ
> slightly (some return just a boolean, others return a pointer that can
> be used as a boolean too), the overall semantics of them are quite
> consistent.  strsuffix() has less drift than strprefix().
> 
> If we added strsuffix() here with the semantics commonly attributed to
> strprefix(), and then had strskipto() with the semantics commonly
> attributed to strsuffix(), I'd expect programmers to eventually get
> confused and use the wrong one. 
> 
> 
> Have a lovely day!
> Alex
> 
> > 
> >  FWIW,
> > 
> >   Maciej
> 
> -- 
> <https://www.alejandro-colomar.es>



-- 
<https://www.alejandro-colomar.es>

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]