Re: Code reviews [Was: FWTK and smap/smapd]

[ark () eltex ru]

-----BEGIN PGP SIGNED MESSAGE-----

Carson Gaspar <carson () taltos org> said :

> > It depends on what the maximum buflen is, isn't it?
> >
> > If we have a small buffer that we know does fit send queue we may
> > assume that if write count is not equal to expected size something
> > wrong happens with the socket and we may shut the connection down
> > diagnosing network error, am i wrong?
> >
> > If the buffer size can really grow that much we need something like
> > net_write, sure.
>
> Ah, but I cleverly failed to tell you:
>
> - How much data you were dealing with
> - Whether the fd was blocking or non-blocking

Then i should implement the most reliable algorithm.
 
> In my opinion, anyone who _ever_ assumes write(fd, buf, n) will only ever 
> return n or -1 is writing bogus code. Always program defensively - you may 
> be a genius, but the person who mangles your code later won't be.

Hmm actually i still think that handling partial write of, say, 1024 byte buffer
as socket error does not break things. It is bad style, though, and should be
avoided.

                                     _     _  _  _  _      _  _
 {::} {::} {::}  CU in Hell          _| o |_ | | _|| |   / _||_|   |_ |_ |_
 (##) (##) (##)        /Arkan#iD    |_  o  _||_| _||_| /   _|  | o |_||_||_|
 [||] [||] [||]            Do i believe in Bible? Hell,man,i've seen one!

-----BEGIN PGP SIGNATURE-----
Version: PGP 6.5.1i

iQCVAwUBPT2Ex6H/mIJW9LeBAQH0JgP9ED2QwJbrC4aY6wnCM71dJ6SB+gICRBhS
u4u8e6BeiOOdOY10IHtaPLRPNdR5ZWSGvz3/F3XoZggWBNn3wGDfAsIZiodfz32t
HqUeVHMD1WUJrUJHxQd7PZrqslsghVwsu8IPBbDQEDH2zutXitF+y02fOU4dHENR
cnXk0UW2WiI=
=SJdp
-----END PGP SIGNATURE-----
_______________________________________________
firewall-wizards mailing list
firewall-wizards () honor icsalabs com
http://honor.icsalabs.com/mailman/listinfo/firewall-wizards