Remove some special EBCDIC code

The 'variant_byte_number' function was written to find the byte number
in a word of the first byte whose meaning varies depending on if the
string it is part of is encoded in UTF-8 or not.  On ASCII machines,
that is simply when the upper bit is set.  On EBCDIC machines, there is
no similar pattern, so this function hasn't been compiled on those.

A long time ago, I realized that this function could also handle binary
data by coercing that binary data into having the form of having that
bit set or not depending on the pattern being looked for, and then
calling that function.

But I actually hadn't realized until now that it was binary data not
tied to a character set that was being worked on.  This commit rectifies
that.  A new alias is added for that function that emphasizes that it
works on binary data, the function is now compiled for EBCDIC, and the
EBCDIC-only code that avoided using it is now removed.

Author: khwilliamson

embed.fnc

@@ -4010,6 +4010,8 @@ CRTdip	|UV	|valid_utf8_to_uv					\
 CRTdmp	|UV	|valid_utf8_to_uvchr					\
 				|NN const U8 *s 			\
 				|NULLOK STRLEN *retlen
+CRTip	|unsigned int|variant_byte_number				\
+				|PERL_UINTMAX_T word
 Adp	|int	|vcmp		|NN SV *lhv				\
 				|NN SV *rhv
 Adpr	|void	|vcroak 	|NULLOK const char *pat 		\
@@ -4108,10 +4110,6 @@ TXp	|void	|set_padlist	|NN CV *cv				\
 : Used in sv.c
 p	|void	|dump_sv_child	|NN SV *sv
 #endif
-#if !defined(EBCDIC)
-CRTip	|unsigned int|variant_byte_number				\
-				|PERL_UINTMAX_T word
-#endif
 #if defined(F_FREESP) && !defined(HAS_CHSIZE) && !defined(HAS_TRUNCATE)
 ARdp	|I32	|my_chsize	|int fd 				\
 				|Off_t length

embed.h

@@ -848,6 +848,7 @@
 # define valid_identifier_sv(a)                 Perl_valid_identifier_sv(aTHX_ a)
 # define valid_utf8_to_uv                       Perl_valid_utf8_to_uv
 # define Perl_valid_utf8_to_uvchr               valid_utf8_to_uvchr
+# define variant_byte_number                    Perl_variant_byte_number
 # define vcmp(a,b)                              Perl_vcmp(aTHX_ a,b)
 # define vcroak(a,b)                            Perl_vcroak(aTHX_ a,b)
 # define vdeb(a,b)                              Perl_vdeb(aTHX_ a,b)
@@ -874,9 +875,6 @@
 #   define pad_setsv(a,b)                       Perl_pad_setsv(aTHX_ a,b)
 #   define pad_sv(a)                            Perl_pad_sv(aTHX_ a)
 # endif
-# if !defined(EBCDIC)
-#   define variant_byte_number                  Perl_variant_byte_number
-# endif
 # if defined(F_FREESP) && !defined(HAS_CHSIZE) && !defined(HAS_TRUNCATE)
 #   define my_chsize(a,b)                       Perl_my_chsize(aTHX_ a,b)
 # endif

inline.h

@@ -2008,13 +2008,20 @@ Perl_single_1bit_pos32(U32 word)
 
 }
 
-#ifndef EBCDIC
+/* Returns the byte number of the lowest numbered-byte whose uppermost bit is
+ * set */
+#define first_upper_bit_set_byte_number(word) Perl_variant_byte_number(word)
 
 PERL_STATIC_INLINE unsigned int
 Perl_variant_byte_number(PERL_UINTMAX_T word)
 {
-    /* This returns the position in a word (0..7) of the first variant byte in
-     * it.  This is a helper function.  Note that there are no branches */
+    /* This returns the position in a word (0..7) of the first byte whose
+     * uppermost bit is set.  On ASCII boxes, this is equivalent to the first
+     * byte whose representation is different in UTF-8 vs not, hence the name
+     * and text in the comments.  It was only later that this was used for
+     * binary data, not tied to the character set.
+     *
+     * This is a helper function.  Note that there are no branches */
 
     /* Get just the msb bits of each byte */
     word &= PERL_VARIANTS_WORD_MASK;
@@ -2023,7 +2030,7 @@ Perl_variant_byte_number(PERL_UINTMAX_T word)
      * word */
     assert(word);
 
-#  if BYTEORDER == 0x1234 || BYTEORDER == 0x12345678
+#if BYTEORDER == 0x1234 || BYTEORDER == 0x12345678
 
     /* Bytes are stored like
      *  Byte8 ... Byte2 Byte1
@@ -2036,7 +2043,7 @@ Perl_variant_byte_number(PERL_UINTMAX_T word)
      * to 0..7 */
     return (unsigned int) ((word + 1) >> 3) - 1;
 
-#  elif BYTEORDER == 0x4321 || BYTEORDER == 0x87654321
+#elif BYTEORDER == 0x4321 || BYTEORDER == 0x87654321
 
     /* Bytes are stored like
      *  Byte1 Byte2  ... Byte8
@@ -2069,11 +2076,10 @@ Perl_variant_byte_number(PERL_UINTMAX_T word)
     /* If all else fails, it's better to return something than just random */
     return 0;
 
-#  endif
+#endif
 
 }
 
-#endif
 #if defined(PERL_CORE) || defined(PERL_EXT)
 
 /*

proto.h

@@ -653,15 +653,9 @@ S_find_span_end(U8 * s, const U8 * send, const U8 span_byte)
                 continue;
             }
 
-            /* Here, at least one byte in the word isn't 'span_byte'. */
-
-#ifdef EBCDIC
-
-            break;
-
-#else
-
-            /* This xor leaves 1 bits only in those non-matching bytes */
+            /* Here, at least one byte in the word isn't 'span_byte'.
+             *
+             * This xor leaves 1 bits only in those non-matching bytes */
             span_word ^= * (PERL_UINTMAX_T *) s;
 
             /* Make sure the upper bit of each non-matching byte is set.  This
@@ -671,10 +665,7 @@ S_find_span_end(U8 * s, const U8 * send, const U8 span_byte)
             span_word |= span_word << 4;
 
             /* That reduces the problem to what this function solves */
-            return s + variant_byte_number(span_word);
-
-#endif
-
+            return s + first_upper_bit_set_byte_number(span_word);
         } while (s + PERL_WORDSIZE <= send);
     }
 
@@ -789,20 +780,11 @@ S_find_span_end_mask(U8 * s, const U8 * send, const U8 span_byte, const U8 mask)
                 continue;
             }
 
-#ifdef EBCDIC
-
-            break;
-
-#else
-
             masked ^= span_word;
             masked |= masked << 1;
             masked |= masked << 2;
             masked |= masked << 4;
-            return s + variant_byte_number(masked);
-
-#endif
-
+            return s + first_upper_bit_set_byte_number(masked);
         } while (s + PERL_WORDSIZE <= send);
     }