Pulled from regcomp.c.
*/
PERL_STATIC_INLINE I32
-S_regcurly(pTHX_ const char *s)
+S_regcurly(pTHX_ const char *s,
+ const bool rbrace_must_be_escaped /* Should the terminating '} be
+ preceded by a backslash? This
+ is an abnormal case */
+ )
{
PERL_ARGS_ASSERT_REGCURLY;
while (isDIGIT(*s))
s++;
}
- if (*s != '}')
- return FALSE;
- return TRUE;
+
+ return (rbrace_must_be_escaped)
+ ? *s == '\\' && *(s+1) == '}'
+ : *s == '}';
}
/* XXX Add documentation after final interface and behavior is decided */
*/
STATIC char
-S_grok_bslash_c(pTHX_ const char source, const bool utf8, const bool output_warning)
+S_grok_bslash_c(pTHX_ const char source, const bool output_warning)
{
U8 result;
- if (utf8) {
- /* Trying to deprecate non-ASCII usages. This construct has never
- * worked for a utf8 variant. So, even though are accepting non-ASCII
- * Latin1 in 5.14, no need to make them work under utf8 */
- if (! isASCII(source)) {
- Perl_croak(aTHX_ "Character following \"\\c\" must be ASCII");
+ if (! isPRINT_A(source)) {
+ const char msg[] = "Character following \"\\c\" must be printable ASCII";
+ if (! isASCII(source)) {
+ Perl_croak(aTHX_ "%s", msg);
+ }
+ else if (output_warning) { /* Unprintables can be removed in v5.22 */
+ Perl_ck_warner_d(aTHX_ packWARN2(WARN_DEPRECATED, WARN_SYNTAX), "%s",
+ msg);
}
}
+ else if (source == '{') {
+ assert(isPRINT_A(toCTRL('{')));
- result = toCTRL(source);
- if (! isASCII(source)) {
- Perl_ck_warner_d(aTHX_ packWARN2(WARN_DEPRECATED, WARN_SYNTAX),
- "Character following \"\\c\" must be ASCII");
+ /* diag_listed_as: Use "%s" instead of "%s" */
+ Perl_croak(aTHX_ "Use \"%c\" instead of \"\\c{\"", toCTRL('{'));
}
- else if (! isCNTRL(result) && output_warning) {
- if (source == '{') {
- Perl_ck_warner_d(aTHX_ packWARN2(WARN_DEPRECATED, WARN_SYNTAX),
- "\"\\c{\" is deprecated and is more clearly written as \";\"");
- }
- else {
- U8 clearer[3];
- U8 i = 0;
- if (! isWORDCHAR(result)) {
- clearer[i++] = '\\';
- }
- clearer[i++] = result;
- clearer[i++] = '\0';
-
- Perl_ck_warner(aTHX_ packWARN(WARN_SYNTAX),
- "\"\\c%c\" is more clearly written simply as \"%s\"",
- source,
- clearer);
- }
+
+ result = toCTRL(source);
+ if (output_warning && ! isCNTRL_L1(result)) {
+ /* We use isCNTRL_L1 above and not simply isCNTRL, because on EBCDIC
+ * machines, things like \cT map into a C1 control. */
+ U8 clearer[3];
+ U8 i = 0;
+ if (! isWORDCHAR(result)) {
+ clearer[i++] = '\\';
+ }
+ clearer[i++] = result;
+ clearer[i++] = '\0';
+
+ Perl_ck_warner(aTHX_ packWARN(WARN_SYNTAX),
+ "\"\\c%c\" is more clearly written simply as \"%s\"",
+ source,
+ clearer);
}
return result;
return TRUE;
}
+STATIC char*
+S_form_short_octal_warning(pTHX_
+ const char * const s, /* Points to first non-octal */
+ const STRLEN len /* Length of octals string, so
+ (s-len) points to first
+ octal */
+) {
+ /* Return a character string consisting of a warning message for when a
+ * string constant in octal is weird, like "\078". */
+
+ const char * sans_leading_zeros = s - len;
+
+ PERL_ARGS_ASSERT_FORM_SHORT_OCTAL_WARNING;
+
+ assert(*s == '8' || *s == '9');
+
+ /* Remove the leading zeros, retaining one zero so won't be zero length */
+ while (*sans_leading_zeros == '0') sans_leading_zeros++;
+ if (sans_leading_zeros == s) {
+ sans_leading_zeros--;
+ }
+
+ return Perl_form(aTHX_
+ "'%.*s' resolved to '\\o{%.*s}%c'",
+ (int) (len + 2), s - len - 1,
+ (int) (s - sans_leading_zeros), sans_leading_zeros,
+ *s);
+}
+
/*
* Local variables:
* c-indentation-style: bsd