Shmuel (Seymour J.) Metz writes: > The mapping for ADATA has a number of fields described as offset without > saying what they are offsets against. It appears as if the answer varies > depending on the record type. Is there a guide to parsing ADATA that spells > it out?
Although the comments in the ASMADATA macro are intended to be helpful, they do not provide full documentation. The official interface information for the ADATA records is in the section "Associated data file output" in the Programmer's Guide, which has separate topics for each record type: https://www.ibm.com/docs/en/hla-and-tf/1.6?topic=output-associated-data-file Unless stated otherwise, offsets are normally relative to the start of the record. This means the complete record as written to the data set, including the 4-byte V-format length prefix. (On VM/CMS, the length prefix is removed by the QSAM simulation routines and the length is stored outside the record, so if the file is read using CMS access methods, the offset is 4 more than the offset from the start of the CMS record). If you spot any case where the documentation does not match the implementation, please let us know. Jonathan Scott, HLASM IBM Hursley, UK
