For very large Unicode code points, the ASC function returns a negative value that is signed 2's complement

May 10, 2023
Knowledge

Title

URL Name

asc-function-overflows-and-returns-negative-value-for-large-maxint32-unicode-codepoints

Article Number

000230600

Environment

Product: OpenEdge Version: 11.7.x, 12.2.x OS: All supported platforms

Question/Problem Description

For very large Unicode code points, the ASC function returns a negative value that is signed 2's complement
For higher blocks of Unicode, e.g. Dingbats, and Playing cards, the ASC function returns a negative value that is signed 2's complement

Steps to Reproduce

Run the below sample code, MESSAGE ASC('𓅓', 'UTF-8', 'UTF-8') SKIP(1) SESSION:cpinternal //ASC returns -258767469 that is signed 2's complement for 𓅓 VIEW-AS ALERT-BOX INFORMATION BUTTONS OK.

Clarifying Information

For very large Unicode code points (e.g.https://www.fileformat.info/info/unicode/char/13153/index.htm), the ASC function returns a signed 2's complement (in this example -258767469). The ABL support int as signed int, so the value we are getting is technically correct when using a (signed) int value. The codepoints specified by UTF-8 define the bytes for a given character, but it is not defining the actual integer value, because the actual value will look different depending on whether we are using a signed or unsigned int. Both are correct from the codepoint perspective, that is, they both resolve to the same hexa value. More importantly, if we feed that negative value to the CHR() function, it will give you the same (correct) character if using the same codepages, so the fact is we have signed integers in the ABL.

Error Message

Defect Number

Enhancement Number

Cause

Resolution

If returning signed 2's complement is not feasible then use the workaround below.

Workaround

Checking if the output is signed 2's complement or write a method to convert the 2-component value back to a positive int64.
Another workaround is using the utf-32 instead of utf-8 for the ASC/CHR function. That way you get a positive value, and one that directly corresponds to the Unicode codepoint notation when converted to hexadecimal.

MESSAGE ASC('A', 'UTF-32', SESSION:CPINTERNAL) SKIP
0x0041 SKIP(1) /* 65 in hex */ "~u0041" SKIP
CHR(0x0041, 'UTF-32') VIEW-AS ALERT-BOX.

Notes

Keyword Phrase

Last Modified Date

5/17/2023 4:38 AM

For very large Unicode code points, the ASC function returns a negative value that is signed 2's complement

Files(0)

Was this article helpful?

Was this article helpful?

For very large Unicode code points, the ASC function returns a negative value that is signed 2's complement

Information

Related Articles