BHCS File and record formats

The following are the file and record formats used by the Barr Host Communications Suite. Click the following links for descriptions of the formats.

• .tnj, .bnj, .anj, .xnj, .lnj formats

• Barr Archive format version 1

• Barr Archive format version 2

• VBM format

• S/370 format

See also:

• Data formats and conversions

• Using the Viewer

.tnj, .bnj, .anj, .xnj, .lnj formats

These formats are used internally by Barr Systems and are subject to change without notification.

When in the Barr spooler, the Barr host formats (.tnj, .bnj, .anj, .xnj, and .lnj) have the same basic layout. The only difference is the data they contain. Text (EBCDIC) jobs are labeled .tnj, binary jobs are labeled .bnj, AFP jobs are labeled .anj, Xerox Metacode jobs are labeled .xnj, and Xerox LCDS Acronym for Line Condition Data Stream. An LCDS print job or print file is line data (text) with some text-based Xerox commands included, such as DJDE commands. jobs are labeled .lnj. The "nj" represents "NJE-wrapped," which means the individual records are formatted similar to the way records are formatted in the NJE protocol. Our records are stored in a subset of the NJE format consisting of data, logical record length (LRECL) bytes (including trailing blanks), record length (RECL) bytes, and a sub-record control byte (SRCB).

You can click in the Data Type column and change the file format.

To view the records:

1. Add the Data Type column to the Spool Window.

2. If the data type is .tnj or .bnj, open the file with the Viewer utility. If the data type is .anj, you must open the file using a hexadecimal editor.

3. On the menu bar, select View | Binary | EBCDIC. Use the examples and following table to interpret the record.

Unspanned record structure:

Spanned record structure (first segment):

SEGL = amount of data bytes + 2 bytes (LRECL bytes)

For more information, see IBM's Network Job Entry Formats and Protocols manual.

Spanned record support allows records larger than 255 bytes to be segmented into smaller records. The spanned record bits indicate whether the record is unspanned or part of a larger spanned record. The first segment of a spanned record has a slightly different format than an unspanned record. The middle and last segments have the same format as an unspanned record, except carriage control bytes are not possible. If the total length of the entire spanned record is not filled by the segments, the remaining bytes are padded with spaces.

Unspanned record example:

The SRCB A0 indicates this is an unspanned record with ASA carriage control (of F1 skip to channel 1). The RECL 04 indicates the record length is 4 bytes. The LRECL 06 indicates the record length will be 6 once the trailing spaces are added back to the end of the record. (This record contains 2 trailing spaces.) The first data byte, F1, is the carriage control type followed by the data "ABC".

Spanned record example:

The first record is the first segment of a spanned record. The SRCB 98 indicates this is a spanned record with machine carriage control (of 89 skip to channel 1). The total record size is 16 bytes. The LRECL in each segment indicates the bytes contained in that segment. The LRECL in the first segment is 05, indicating it contains 5 of the 16 bytes. The LRECL in the second segment is 07, indicating it contains 7 bytes. The LRECL in the last segment is 04 indicating it contains 4 bytes. The first data byte, 89, is the carriage control type followed by the data "123456789012345". Typically, you would not span a 16-byte record across 3 records, it is shown here for simplicity.

Barr Archive format version 1

Lengths are encoded in big endian Storing numbers in such a way that the most significant byte is placed first. format.

SIGNATURE BLOCK

##BARR## Barr Data File VXX.XX ....... ##\x01a\1\1

Identifies the stream that follows as having the Barr Archive format version 1.

JOB HEADER BLOCK

DATA SET BLOCK

TRAILER BLOCK

Barr Archive format version 2

Record lengths are encoded in big endian Storing numbers in such a way that the most significant byte is placed first. format.

Record format

Each tagged record in a Barr Archive Version 2 file begins with a 4 byte tag name that identifies the record type, followed by a 4-byte record length, followed by the data associated with the record.

The OPEN, CLOS, DS_z, and OL_z record types always have a zero record length. This causes the next record to start immediately after their record length part. The Signature Record is always the first record in the file and has an assumed length of 36 bytes. There is no record length part in the Signature Record so the next record starts immediately after the last byte of the Signature Record.

The following is an alphabetical list of record name tags.

Banners

Leading Banner and Trailing Banner records are optional. Banners consist of a single record.

Data sets

Data sets can consist of multiple DS records. Each DS record is concatenated To join sequentially (for example, joining the characters in a character string). to the current data set. The end of a data set is marked by the DS_z data set End tag.

FCBs

Forms Control Buffers are optional. An FCB consists of a single record. Only data sets of the "NJE-wrapped EBCDIC text" type are compatible with FCBs. Other types of data sets will not have FCBs present in their files. Data sets of unknown type are assumed to be of the "NJE-wrapped EBCDIC text" type and therefore can include FCB records.

Overlays

Overlay records are optional. Overlays can consist of more than one record, so the first OL1 tag (or OL2 or OL4) can be followed by another OL1 tag for the same data set. Each overlay record is concatenated to the current overlay. The end of an overlay is marked by the OL_z (Overlay End) tag. The OL3 tag is currently not used.

The fourth byte of an Overlay tag name can be A, B, N, or T. These letters indicate the type of the original overlay, as in the following table:

An overlay tag with the value OL1T indicates that Overlay 1 is NJE-wrapped EBCDIC text. When the fourth character of an overlay tag is the space character, the Overlay is of an unknown type.

Sequenced list of records. A file created with the Archive Version 2 format will have sections in the following order:

VBM format

VBM is a block variable format similar to the mainframe format RECFM=VBM. The blocks and records must adhere to a specific structure.

Block structure � A data block can contain one or more records. The block begins with a two-byte count of the bytes in the block, with the most significant byte (MSB) value listed first followed by the least significant byte (LSB). Two null bytes follow the count bytes. The block size includes the block length bytes, null bytes, and sum of all the record lengths.

Note that as long as the record structure is correct (see below), it is not necessary to block the records (that is, the block structure can be omitted). The Barr Host Communications Suite will automatically detect when the block structure is included. When the block structure is used, the format is referred to as BDW Block Descriptor Words. When the block structure is not used, the format is referred to as RDW Record Descriptor Words.

Record structure � Each record begins with a two-byte count of the bytes in the record and two null bytes, followed by a carriage control byte. The record size includes five bytes of format information (the length bytes, null bytes, and carriage control byte) and all the data bytes. Carriage control bytes must be expressed as machine carriage control.

S/370 format

Files in S/370 format must have a four-byte format header at the beginning of the file. The format header bytes must have specific hexadecimal values. Data records follow the file format header and are usually in EBCDIC because most channel printers require EBCDIC data. Records must contain count bytes indicating the record length and channel command characters. Additionally, most records contain print data. Follow the guidelines below when you create data records.

Include a record length count at the beginning and end of the record. Both length counts must have the same value and must include the number of data bytes plus one byte for carriage control.

Include only one channel command byte per record. Use machine channel commands.

Format header

Records

*The beginning and ending length bytes have the same values.

For S/370 record format, you must give the record length counts in INTEL format, where the most significant byte (MSB) follows the least significant byte (LSB). For example, you would represent the length value 0150 as 50 01.

In this example, the file contains a header and three records. Two of the records contain print text.