> 4) Added HX_SOCKOPT_READBUF_FLAG socket option. Value can be one of 
> following.
> 
> HX_SOCK_READBUF_SIZE_DEFAULT
> HX_SOCK_READBUF_SIZE_COPY // copy to precisely sized buffer after read
> HX_SOCK_READBUF_SIZE_PEEK // use FIONBIO or equivalent (not implemented)
> 
> Note that if HX_SOCK_READBUF_SIZE_PEEK  were to work reliably across 
> platforms HX_SOCK_READBUF_SIZE_COPY  may not be needed. However, I'm not at 
> this point sure it is reliably supported and efficient.

Wasn't it agreed that the COPY option is undesirable?  Is it going to
actually be used anywhere?  If not, I don't see a reason for it.

Why not test FIONBIO on the major platforms instead of adding an option that
will probably not be used?

-- 
Old timer, n.:
        One who remembers when charity was a virtue and not an organization.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: Digital signature
Url : http://lists.helixcommunity.org/pipermail/common-dev/attachments/20040809/49203c35/attachment.bin