Re: A really little OCI-Problem

From: Scott Urman <surman_at_oracle.com>
Date: 1995/12/12
Message-ID: <4aku5t$t5v_at_inet-nntp-gw-1.us.oracle.com>#1/1


In article <4ah5p1$dp1_at_news.rz.uni-passau.de>, berndl_at_piligrim.uni-passau.de (Klaus Berndl) writes:
|> A short problem about OCI:
|>
|> The 'odefin'-call requires a buffer and a bufferlength to store fetched datas.
|> The bufferlength is a integer-parameter, which is on some systems limited to
|> 2 bytes. To use greater than 64K buffers, you must write the bufferlength
|> in the first 4 bytes of buffer, and pass -1 as bufferlength!
|>
|> Now my question: Must i code the bufferlength in LOW-HIGH or HIGH-LOW order in the
|> first 4 bytes of buffer. To my knowledge, this is related to the processor-architecture (i`m working on SUN-SPARCS with Solaris 2.3). But in which order the 'odefin'-function requires my bufferlength?
|>
|> Many Thanks and best regards,
|>
|> KLaus Berndl berndl_at_fmi.uni-passau.de
|>
|>
It would be related to your processor. Just put the length into a 4 byte int, and memcpy it into the first 4 bytes of the data. However an int is represented on your system is the correct one. Received on Tue Dec 12 1995 - 00:00:00 CET

Original text of this message