Standardize rule/pattern quoting

• 12/09/03 13:42:27 schererm changed notes2
• 12/09/03 13:42:27 schererm moved from incoming to formatting
• 02/06/04 12:59:12 andy sent reply 1
• 02/06/04 13:02:31 andy changed notes2
• 02/17/04 03:09:09 weiv changed notes2
• 02/17/04 20:06:17 weiv changed notes2
• 02/18/04 13:59:43 schererm changed notes2
• 07/19/04 18:39:03 schererm changed notes2

Consolidating with bug 901, which was...

BACKGROUND

Several components of ICU4C/J are data driven. They take patterns (short strings, less than a line of text) or rules (typically one or more lines of text) that parameterize them. Examples are SimpleDateFormat, DecimalFormat, RuleBasedCollator, RuleBasedTransliterator, RuleBasedBreakIterator, and other classes. In these rules, a class of characters ('specials') must be escaped or quoted if they are to be used as literals. Quoting mechanisms include single quotes and backslash escapes (including \uXXXX).

Rule-based components typically return their rule string via methods such as toPattern(), getRules(), etc. The returned string should be directly usable to create new objects equivalent to the originals. For this to work, special characters must be escaped in the returned rule string.

RFE

Add a quoting function that escapes and/or quotes special characters. This function should take a string, a set of characters that need quoting, and return the quoted string.

C

/**

• Quote specials in-place. Return actual length (may be > capacity).
• @param specials set of characters that need quoting */

int32_t u_quoteSpecials(UChar* string, int32_t length, int32_t capacity,

UUnicodeSet* specials);

This might be a little cleaner:

/**

• Quote specials. Return result length (may be > capacity). */

int32_t u_quoteSpecials(const UChar* string, int32_t length,

UChar* result, int32_t resultCapacity, UUnicodeSet* specials);

Java

/**

• Quote specials. */

UnicodeString quoteSpecials(UnicodeString string,

UnicodeSet specials);