James Carlson wrote:

<snip>

Note: I am replying only to strsep() related discussion for now.

> 754: you should have your sponsor (at least) update those CRs to
> reference this file as being another copy.  If I were an RTI Advocate
> (I'm not), I'd probably be less generous: I'd say that if you know
> about a problem (lack of common C library function), and you're adding
> to the problem (yet another copy), then you're now on the hook to fix
> it, even if you didn't cause it in the first place.  (But, like I
> said, I don't have that authority.  :->)

The plan is to wait for strsep() to be integrated into libc and then 
proceed with putback of CR 6691168.

Just to be clear: I am not in favor of providing another copy of 
strsep() (or any other similar function). The strsep() is included only 
for testing purposes.

I have talked with other engineer and there seems to be a queue of 
people lining up for strsep() in libc.

> Ditching strsep and using strtok would avoid that particular problem.

The catch with this approach is that strsep() makes syntax parsing (with 
error detection) easier (than with strtok()).

>> The fix incorporates a function called strsep() which is about to get in
>> libc in the future but is not available so far (see CR6487100 &
>> 4383867). So don't wonder about the implementation of strsep() inside
>> the netcat.c.
> 
> Strange ... one of those two CRs appears to be a duplicate.

I have just closed CR 6487100 as a dup and provided a link to a comment 
about native implementation (str[c]spn()) which was part of 6487100.

The point of that comment is that neither strtok() or strsep() have 
flaws so str[c]spn() (which does a better job) should be used instead in 
internal code. Maybe another thing worth investigating in the context of 
this RFE.


v.
_______________________________________________
opensolaris-code mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/opensolaris-code

Reply via email to