Translate tables: Difference between revisions

From m204wiki
Jump to navigation Jump to search
mNo edit summary
No edit summary
Line 138: Line 138:
You can then modify the source and reload the table using the [[JANUS LOADXT]] command.
You can then modify the source and reload the table using the [[JANUS LOADXT]] command.


Note: Because the STANDARD table in the example above may already be a modified version of the STANDARD table distributed with Janus TCP/IP Base, if you definitely want to copy the distributed STANDARD table, issue the USE and [[JANUS DISPLAYXT]] commands in a “private” Online where you are confident the table has not been modified.
Note: Because the STANDARD table in the example above may already be a modified version of the STANDARD table distributed with Janus TCP/IP Base, if you definitely want to copy the distributed STANDARD table, issue the USE and [[JANUS DISPXT]] commands in a “private” Online where you are confident the table has not been modified.


==References==
==References==


See: [[List_of_Janus_commands|Janus command list]],  [[JANUS LOADXT]], [[JANUS DISPLAYXT]]
See: [[List_of_Janus_commands|Janus command list]],  [[JANUS LOADXT]], [[JANUS DISPXT]]


[[Category:Janus Web Server|Translate tables]]
[[Category:Janus Web Server|Translate tables]]
[[Category:Janus commands|Translate tables]]
[[Category:Janus commands|Translate tables]]
[[Category:Janus TCP/IP Base]]
[[Category:Janus TCP/IP Base]]

Revision as of 14:13, 27 March 2013

Janus Connectivity products like Janus Web Server and Janus SOAP typically communicate with applications on remote computers. In many cases, these remote computers use different encodings for symbols, numerals, and letters or “characters.” A character usually occupies a single byte of memory on any computer, but the character displayed for a particular bit value may vary. For example, in an ASCII encoding, the numeral “1” is represented by the hexadecimal value X'31'. On a computer that uses an EBCDIC encoding (IBM mainframes), the representation is X'F1'. Therefore a numeral “1” transmitted from an ASCII-based computer to an EBCDIC-based computer, or vice versa, will be misinterpreted. To avoid such misinterpretation, a method of converting a byte from one encoding to another is needed. Janus translate tables were implemented to solve this problem.

Versions and capability

In Janus release 5.6 and earlier, a fixed translate table provided the most common type of translation: that is, it converted ASCII to EBCDIC when receiving, and it converted EBCDIC to ASCII when sending. In most cases, this is adequate for any of the Janus products. There are, however, many different types of EBCDIC and ASCII character sets, designed to display foreign language characters, currency, and other special symbols. To ensure these characters are transmitted and received correctly, a more flexible approach is required.

In Janus 6.0, the translate table facility was expanded to allow simultaneous use of many different translate tables, and further, to allow user-written or customized translate tables.

Using Translate Tables

To use Janus translate tables, you must do one of the following if you do not want to use the standard tables:

  • Code your own translate table source.
  • Use an external translate table.
  • Use or modify a Sirius-provided translate table.

The JANUS LOADXT command then validates the new table, converts it to an internal format, and adds or replaces any existing table with the same name.

To use the most up-to-date Sirius-provided translation file, specify the following to load and name the Unicode table that handles Janus SOAP XML document translations (you must be running Sirius Mods version 7.6 or higher):

JANUS LOADXT xtabname UNICODE

See JANUS LOADXT.

To view the translations in effect after issuing this command:

  1. Determine the codepage in use at your site for translations between Unicode and EBCDIC:

    UNICODE Display Table Standard

    The result of this UNICODE command (described further in the Janus SOAP Reference Manual) will begin with something like:

    * The following commands produce the current Unicode translation tables: UNICODE Table Standard Base Codepage pppp

    where pppp is the codepage in use.
  2. Using the above codepage, issue:

    UNICODE Display Codepage pppp

    The result will be many lines of single-character translations, one line for each character translated between EBCDIC and Unicode. This format differs entirely

    from that described below in “Defining translate tables”.

To provide backward compatibility with earlier releases, two additional translate tables are loaded during Model 204 initialization. The standard table, named “STANDARD,” is an exact copy of the translate table used with earlier Janus releases. This is the default translate table for all Janus ports, unless you define Web ports with DBCS support. In this case, the “JAPANIBM” translate table is used. Both these tables are built-in to the Sirius Mods code. As described in the next section, “Defining translate tables”, the JANUS DISPXT and JANUS LOADXT commands let you view (or copy for modification purposes) and reload these translate tables.

Defining translate tables

You can define your own translate tables and load them with the JANUS LOADXT command (“JANUS LOADXT” on page 139). These tables can be defined in a Model 204 procedure file or group, or in a sequential file or MVS Partitioned Data Set (PDS).

A translate table must define a 256-byte EBCDIC-to-ASCII table, followed by a 256-byte ASCII-to-EBCDIC table. Comment lines are allowed and can begin with a semicolon (;) or an asterisk (*).

Comments are allowed on the same line as definition lines, provided they are preceded by either of the two comment characters.

You can define any number of bytes of the translate table on a single line, but it is easier to read and maintain if you define 16 bytes per line. Certain characters must be translated to their equivalent values in EBCDIC and ASCII for basic Janus functions to work correctly. Therefore, these characters are reserved for use by Janus, and they may not be translated to any other characters. These reserved characters are the following:

0-9 a-z A-Z . period < less than ( left parenthesis + plus & ampersand  ! exclamation * asterisk ) right parenthesis  ; semicolon / forward slash , comma  % percent > greater than  ? question mark  : colon # pound sign ' single quote = equal sign " double quote

JANUS LOADXT will validate any table it attempts to load, and if these characters are not preserved when translating to ASCII and back, the table will not be loaded.

Each character in the translate table is coded in hexadecimal format and must be two bytes long. Each two-byte hexadecimal string must be separated by one or more blanks.

Following is an abbreviated example of translate table source code. Each of the rows of hexadecimal characters define the EBCDIC encoding for an ASCII character at that offset in the table. So for example, to find out how this table would translate an ASCII 2 (hexadecimal X'32'), you locate offset X'32' in the table, which is in the fourth row, third pair of characters. Note that the top and right-side comments help to locate offsets quickly. At offset X'32' in this table, you find the value X'F2', which is the EBCDIC encoding for the numeral 2.

Of course, to send characters, you would want to translate EBCDIC encoding back into ASCII. Therefore, an EBCDIC to ASCII section must also be coded. This is not shown in the example.

 ;  ; ASCII-to-EBCDIC table for English (US) CECP Code Page 037  ; 00 01 02 03 04 05 06 07 08 09 0A 0B 0C 0D 0E 0F  ; 00 01 02 03 37 2D 2E 2F 16 05 25 0B 0C 0D 0E 0F ; 00 ; 10 11 12 13 3C 3D 32 26 18 19 3F 27 1C 1D 1E 1F ; 10 ; 40 5A 7F 7B 5B 6C 50 7D 4D 5D 5C 4E 6B 60 4B 61 ; 20 ; F0 F1 F2 F3 F4 F5 F6 F7 F8 F9 7A 5E 4C 7E 6E 6F ; 30 ; 7C C1 C2 C3 C4 C5 C6 C7 C8 C9 D1 D2 D3 D4 D5 D6 ; 40 ; D7 D8 D9 E2 E3 E4 E5 E6 E7 E8 E9 BA E0 BB B0 6D ; 50 ; 79 81 82 83 84 85 86 87 88 89 91 92 93 94 95 96 ; 60 ; 97 98 99 A2 A3 A4 A5 A6 A7 A8 A9 C0 4F D0 A1 07 ; 70 ; 68 DC 51 42 43 44 47 48 52 53 54 57 56 58 63 67 ; 80 ; 71 9C 9E CB CC CD DB DD DF EC FC 70 B1 80 BF FF ; 90 ; 45 55 CE DE 49 69 9A 9B AB AF 5F B8 B7 AA 8A 8B ; A0 ; 2B 2C 09 21 28 65 62 64 B4 38 31 34 33 4A B2 24 ; B0 ; 22 17 29 06 20 2A 46 66 1A 35 08 39 36 30 3A 9F ; C0 ; 8C AC 72 73 74 0A 75 76 77 23 15 14 04 6A 78 3B ; D0 ; EE 59 EB ED CF EF A0 8E AE FE FB FD 8D AD BC BE ; E0 ; CA 8F 1B B9 B6 B5 E1 9D 90 BD B3 DA FA EA 3E 41 ; F0 ;

You can specify existing external translate table source when loading translate tables.

For example, IBM supplies many translate tables with its TCP/IP product under MVS. You can use these tables directly with JANUS LOADXT. To use these tables, you must first allocate the dataset and member you want to load. Use the Model 204 ALLOCATE command to do this:

ALLOCATE IBMXTAB WITH SCOPE=SYSTEM DSN=TCPIP.AEZAXLT1 - MEMBER=US OLD SEQUENTIAL VOLUME=OS260R

Then use the JANUS LOADXT command to load the translate table:

JANUS LOADXT name DDNAME IBMXTAB

You can obtain source copies of any previously loaded translate table (including internal tables) by using the JANUS DISPXT command and the Model 204 USE command. For example, to obtain a copy of the current translate table “STANDARD,” issue the following commands:

USE ddname JANUS DISPXT STANDARD

You can then modify the source and reload the table using the JANUS LOADXT command.

Note: Because the STANDARD table in the example above may already be a modified version of the STANDARD table distributed with Janus TCP/IP Base, if you definitely want to copy the distributed STANDARD table, issue the USE and JANUS DISPXT commands in a “private” Online where you are confident the table has not been modified.

References

See: Janus command list, JANUS LOADXT, JANUS DISPXT