wpp: Fix expansion of macro bodies following the parsing of a numerical digit.

The current regular expression causes all letters and spaces (among other characters) following the appearance of a digit to be classed as a literal, including C identifiers which may need to be expanded. The expression was intended to catch the remaining characters that were not covered by the first two rules ([^a-zA-Z0-9'"#/\\\n]+ and {cident}), but the [^'"#/\\\n] expression caught {cident} as well. While one solution would have been just to catch the expression that match [a-zA-Z0-9]* that don't match {cident}, i.e. [0-9][a-zA-Z0-9]*, in the interests of avoiding unnecessary multiple LITERALs being generated and then combined during parsing the expression also includes the first expression, making it [0-9][a-zA-Z0-9]*[^a-zA-Z0-9'"#/\\\n]*.

wpp: Fix expansion of macro bodies following the parsing of a numerical digit.
3fecf9dd · Rob Shearman · Alexandre Julliard · 22699818 · 3fecf9dd
Commit 3fecf9dd authored Jan 18, 2010 by Rob Shearman Committed by Alexandre Julliard Jan 19, 2010
Show whitespace changes
Inline Side-by-side

Showing with 1 addition and 1 deletion

ppl.l libs/wpp/ppl.l +1 -1

No files found.
--- a/libs/wpp/ppl.l
+++ b/libs/wpp/ppl.l
@@ -527,7 +527,7 @@ void pp_writestring(const char *format, ...)
 <pp_mbody>{cident}		ppy_lval.cptr = pp_xstrdup(ppy_text); return tIDENT;
 <pp_mbody>\#\#			return tCONCAT;
 <pp_mbody>\#			return tSTRINGIZE;
-<pp_mbody>[0-9][^'"#/\\\n]*	ppy_lval.cptr = pp_xstrdup(ppy_text); return tLITERAL;
+<pp_mbody>[0-9][a-zA-Z0-9]*[^a-zA-Z0-9'"#/\\\n]*	ppy_lval.cptr = pp_xstrdup(ppy_text); return tLITERAL;
 <pp_mbody>(\\\r?)|(\/[^/*'"#\\\n]*)	ppy_lval.cptr = pp_xstrdup(ppy_text); return tLITERAL;
 <pp_mbody>\\\r?\n{ws}+		newline(0); ppy_lval.cptr = pp_xstrdup(" "); return tLITERAL;
 <pp_mbody>\\\r?\n		newline(0);