is_utf8_string

Message-ID: <slrn8ldfpn.h5k.simon@justanother.perlhacker.org>

p4raw-id: //depot/cfgperl@6243

diff --git a/embed.h b/embed.h
index 4c95dc2..fa199fb 100644 (file)
--- a/embed.h
+++ b/embed.h
@@ -305,6 +305,7 @@
 #define to_uni_title_lc                Perl_to_uni_title_lc
 #define to_uni_lower_lc                Perl_to_uni_lower_lc
 #define is_utf8_char           Perl_is_utf8_char
+#define is_utf8_string         Perl_is_utf8_string
 #define is_utf8_alnum          Perl_is_utf8_alnum
 #define is_utf8_alnumc         Perl_is_utf8_alnumc
 #define is_utf8_idfirst                Perl_is_utf8_idfirst
@@ -1758,6 +1759,7 @@
 #define to_uni_title_lc(a)     Perl_to_uni_title_lc(aTHX_ a)
 #define to_uni_lower_lc(a)     Perl_to_uni_lower_lc(aTHX_ a)
 #define is_utf8_char(a)                Perl_is_utf8_char(aTHX_ a)
+#define is_utf8_string(a,b)    Perl_is_utf8_string(aTHX_ a,b)
 #define is_utf8_alnum(a)       Perl_is_utf8_alnum(aTHX_ a)
 #define is_utf8_alnumc(a)      Perl_is_utf8_alnumc(aTHX_ a)
 #define is_utf8_idfirst(a)     Perl_is_utf8_idfirst(aTHX_ a)
@@ -3445,6 +3447,8 @@
 #define to_uni_lower_lc                Perl_to_uni_lower_lc
 #define Perl_is_utf8_char      CPerlObj::Perl_is_utf8_char
 #define is_utf8_char           Perl_is_utf8_char
+#define Perl_is_utf8_string    CPerlObj::Perl_is_utf8_string
+#define is_utf8_string         Perl_is_utf8_string
 #define Perl_is_utf8_alnum     CPerlObj::Perl_is_utf8_alnum
 #define is_utf8_alnum          Perl_is_utf8_alnum
 #define Perl_is_utf8_alnumc    CPerlObj::Perl_is_utf8_alnumc

diff --git a/embed.pl b/embed.pl
index 5b63a35..96603be 100755 (executable)
--- a/embed.pl
+++ b/embed.pl
@@ -1620,6 +1620,7 @@ Ap        |U32    |to_uni_upper_lc|U32 c
 Ap     |U32    |to_uni_title_lc|U32 c
 Ap     |U32    |to_uni_lower_lc|U32 c
 Ap     |int    |is_utf8_char   |U8 *p
+Ap     |bool   |is_utf8_string |U8 *s|STRLEN len
 Ap     |bool   |is_utf8_alnum  |U8 *p
 Ap     |bool   |is_utf8_alnumc |U8 *p
 Ap     |bool   |is_utf8_idfirst|U8 *p

diff --git a/global.sym b/global.sym
index 8a9b008..1f03877 100644 (file)
--- a/global.sym
+++ b/global.sym
@@ -185,6 +185,7 @@ Perl_to_uni_upper_lc
 Perl_to_uni_title_lc
 Perl_to_uni_lower_lc
 Perl_is_utf8_char
+Perl_is_utf8_string
 Perl_is_utf8_alnum
 Perl_is_utf8_alnumc
 Perl_is_utf8_idfirst

diff --git a/objXSUB.h b/objXSUB.h
index ad5aa09..e30258b 100644 (file)
--- a/objXSUB.h
+++ b/objXSUB.h
@@ -707,6 +707,10 @@
 #define Perl_is_utf8_char      pPerl->Perl_is_utf8_char
 #undef  is_utf8_char
 #define is_utf8_char           Perl_is_utf8_char
+#undef  Perl_is_utf8_string
+#define Perl_is_utf8_string    pPerl->Perl_is_utf8_string
+#undef  is_utf8_string
+#define is_utf8_string         Perl_is_utf8_string
 #undef  Perl_is_utf8_alnum
 #define Perl_is_utf8_alnum     pPerl->Perl_is_utf8_alnum
 #undef  is_utf8_alnum

diff --git a/perlapi.c b/perlapi.c
index ccb7c8f..29428d5 100755 (executable)
--- a/perlapi.c
+++ b/perlapi.c
@@ -1326,6 +1326,13 @@ Perl_is_utf8_char(pTHXo_ U8 *p)
     return ((CPerlObj*)pPerl)->Perl_is_utf8_char(p);
 }
 
+#undef  Perl_is_utf8_string
+bool
+Perl_is_utf8_string(pTHXo_ U8 *s, STRLEN len)
+{
+    return ((CPerlObj*)pPerl)->Perl_is_utf8_string(s, len);
+}
+
 #undef  Perl_is_utf8_alnum
 bool
 Perl_is_utf8_alnum(pTHXo_ U8 *p)

diff --git a/pod/perlapi.pod b/pod/perlapi.pod
index 1e48809..86ad5bd 100644 (file)
--- a/pod/perlapi.pod
+++ b/pod/perlapi.pod
@@ -2282,19 +2282,19 @@ false, defined or undefined.  Does not handle 'get' magic.
 =for hackers
 Found in file sv.h
 
-=item svtype
+=item SvTYPE
 
-An enum of flags for Perl types.  These are found in the file B<sv.h> 
-in the C<svtype> enum.  Test these flags with the C<SvTYPE> macro.
+Returns the type of the SV.  See C<svtype>.
+
+       svtype  SvTYPE(SV* sv)
 
 =for hackers
 Found in file sv.h
 
-=item SvTYPE
-
-Returns the type of the SV.  See C<svtype>.
+=item svtype
 
-       svtype  SvTYPE(SV* sv)
+An enum of flags for Perl types.  These are found in the file B<sv.h> 
+in the C<svtype> enum.  Test these flags with the C<SvTYPE> macro.
 
 =for hackers
 Found in file sv.h
@@ -2939,6 +2939,16 @@ Converts the specified character to uppercase.
 =for hackers
 Found in file handy.h
 
+=item U8 *s
+
+Returns true if first C<len> bytes of the given string form valid a UTF8
+string, false otherwise.
+
+       bool_utf8_string        U8 *s(STRLEN len)
+
+=for hackers
+Found in file utf8.c
+
 =item utf8_to_bytes
 
 Converts a string C<s> of length C<len> from UTF8 into ASCII encoding.

diff --git a/proto.h b/proto.h
index c1d9a66..fc34840 100644 (file)
--- a/proto.h
+++ b/proto.h
@@ -367,6 +367,7 @@ PERL_CALLCONV U32   Perl_to_uni_upper_lc(pTHX_ U32 c);
 PERL_CALLCONV U32      Perl_to_uni_title_lc(pTHX_ U32 c);
 PERL_CALLCONV U32      Perl_to_uni_lower_lc(pTHX_ U32 c);
 PERL_CALLCONV int      Perl_is_utf8_char(pTHX_ U8 *p);
+PERL_CALLCONV bool     Perl_is_utf8_string(pTHX_ U8 *s, STRLEN len);
 PERL_CALLCONV bool     Perl_is_utf8_alnum(pTHX_ U8 *p);
 PERL_CALLCONV bool     Perl_is_utf8_alnumc(pTHX_ U8 *p);
 PERL_CALLCONV bool     Perl_is_utf8_idfirst(pTHX_ U8 *p);

diff --git a/utf8.c b/utf8.c
index 9bb89a4..666ec34 100644 (file)
--- a/utf8.c
+++ b/utf8.c
@@ -134,6 +134,30 @@ Perl_is_utf8_char(pTHX_ U8 *s)
     return len;
 }
 
+/*
+=for apidoc Am|bool_utf8_string|U8 *s|STRLEN len
+
+Returns true if first C<len> bytes of the given string form valid a UTF8
+string, false otherwise.
+
+=cut
+*/
+
+bool 
+Perl_is_utf8_string(pTHX_ U8 *s, STRLEN len)
+{
+    U8* x=s;
+    U8* send=s+len;
+    int c;
+    while (x < send) {
+        c = is_utf8_char(x);
+        x += c;
+        if (!c || x > send)
+            return 0;
+    }
+    return 1;
+}
+
 UV
 Perl_utf8_to_uv(pTHX_ U8* s, I32* retlen)
 {