To interpret commands correctly the punctuation characters below are on different code points depending on the EBCDIC CCSID used to enter these values.
CRITICAL PUNCTUATION CHARACTERS: ! $ # @ [ \\ ] ^ ` { | } ~
These critical characters are interpreted normally dependent on the environment variable LANG. If the environment variable LANG is not defined then the default CCSID IBM1047 (Open Systems Latin-1) is used. Below is the current list of supported CCSIDs on EBCDIC systems.
SUPPORTED EBCDIC CODE PAGES FOR COMMAND ENTRY:
"IBM-1140","IBM-1141","IBM-1142","IBM-1143",
"IBM-1144","IBM-1145","IBM-1146","IBM-1147",
"IBM-1148","IBM-1149","IBM-1153","IBM-1154",
"IBM-1156","IBM-1122","IBM-1047","IBM-924",
"IBM-500","IBM-273","IBM-037","IBM-875","IBM-424",
"IBM-277","IBM-278","IBM-280","IBM-284","IBM-285",
"IBM-297","IBM-871","IBM-870","IBM-1025","IBM-1112",
"IBM-1157"
You can define the code page explicitly (LANG=de_DE.IBM-1141) or only the language code (LANG=de_DE, LANG=C). If only the language code defined then the CCSID is derived from the language code (DE=IBM-1141, US=IBM-1140, C=IBM-1047, ...).
If possible, these critical characters are also converted for print outs. At output it is not possible to convert anything correctly, because some strings for print out are coming from other sources (like system messages and others). Only all known literals are converted, for unknown variables such a conversion is not possible and CLE/P® expects that such strings are encoded in the correct system code page, but there is no guaranty for this.
To be independent of the environment escaping of the CLP strings is possible for the critical punctuation characters on EBCDIC systems. See list below:
! = &EXC; - Exclamation mark
$ = &DLR; - Dollar sign
# = &HSH; - Hashtag (number sign)
@ = &ATS; - At sign
[ = &SBO; - Square bracket open
\ = &BSL; - Backslash
] = &SBC; - Square bracket close
^ = &CRT; - Caret (circumflex)
` = &GRV; - Grave accent
{ = &CBO; - Curly bracket open
| = &VBR; - Vertical bar
} = &CBC; - Curly bracket close
~ = &TLD; - Tilde
A escape sequence starts with the ampersand (&) followed be 3 character (not case sensitive) and is terminated with semicolon (;). With an X and 2 hexadecimal digit the definition of a binary byte value are possible. If such a sequence in the string required then the ampersand must be typed twice (&&). To mark a part or the of whole string in a certain CCSID the escape sequence must contain 1 to 6 digits. Few samples can be found below:
&1047; all in 1047 &0; reset to system code page &1140; no in 1140
At the beginning the system code page is active except the environment variable CLP_STRING_CCSID is defined. The defined CCSID is valid till the next CCSID definition. An unsupported CCSID (0) can be used to restore the interpretation to the system code page.
This is no character set conversion, only the UNICODE code points 0-127 which are on different code points in the different EBCDIC code pages are replaced. All other higher code points (e.g. German Umlauts ('ä','ü','ö')) are not touched.
The partial CCSID conversion are mainly useful for application programming interfaces. At compile time the CCSID for literals must be defined. This CCSID could be differ from the system CCSID (local character set) of variable parameter. In such a case a application can mark the literal part in the CCSID used for literals at compile time, and the variable part could be conform to the CCSID defined over the LANG variable. See the C example below:
snprintf(acCmd,sizeof(acCmd),"&1047;get.file='&0;%s&1047'",pcFilename);
C-Code are normally in 1047 and if no literal conversion defined then the literals also in 1047. The file name is a parameter in local character set. In Cobol the default code page for literals is 1140. To define the CCSID for CLP strings from outside, the environment variable CLP_STRING_CCSID can be defined. This is useful if an application cannot recompile with the correct code page for the literals.
On ASCII/UTF-8 platforms a miss-interpretation of punctuation characters smaller than 128 is not possible. On such platforms the LANG variable is not used for command interpretation or printouts. The escape and CCSID sequences are supported but have no effect on these platforms. This helps to develop platform independent code for CLP strings.