Re: Questions about S/36 files

On 05-Nov-2013 16:47 -0800, Jerry C. Adams wrote:

I remember reading and using the "System Data Areas" manuals on the
S/36. The main part that I used (or, more correctly at this distance,
recall using) was the chapters on the F5 records, which defined the
VTOC key into the data. I have never seen/found anything equivalent
for the AS/400/iSeries/System i, but it would seem that there has to
be an "index" that tells the OS where to start reading (address of
the first record), fields, etc. I had to "blackmail" an SE in order
to get my copy so maybe I'll have to try something similar - except
there are no SEs now. <<SNIP>>

The IBM i ensures integrity of the data by preventing direct access to the data. The OS DB feature merely provides access to the data in buffers, and only via the _supported methods_ provided by the OS DB feature. Thus a program can not address the data directly. FWiW, not even the OS DB component can get to the data directly, as it also has only some specific _supported methods_ provided by the LIC DB feature; many of those methods are /the same/ as those made available to the user programs via the OS DB [directly and the languages] such as Open, Read, Write, et al.

It is simple enough to get the address to the data [space header] and then navigate from there to the first record via the STRSST D/A/D, if you are really curious. Dump the member, the MI "cursor" object type [the address, that of the *MEM object, can be obtained from the DMPOBJ of the *FILE], and it will reveal the address of DataSPace object, and then dump that MI object type at that address, and the first and any additional Data segment addresses will be shown. The path can be found directly from the *QDDS address in the DMPOBJ output, but the OS DB can only /see/ the address to the dataspace as found in the associated space of the member [for integrity, the OS can touch\see only the associated space of the object rather than the object itself].