Mercurial > hg > xemacs-beta

diff src/text.c @ 5774:7a538e1a4676

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Use skip_ascii() in no_conversion_convert() when encoding. src/ChangeLog addition: 2013-12-19 Aidan Kehoe <kehoea@parhasard.net> * text.c: * text.h: * text.h (skip_ascii): Move skip_ascii (), the very fast inline function from the bytecount-to-charcount code, to text.h, to allow the coding systems to use it too as needed. * file-coding.c (no_conversion_convert): Use skip_ascii() as appropriate here, halving the time taken to write large files in my tests (again, relevant to VM buffers, but not a panacea to our issues with them.)
author Aidan Kehoe <kehoea@parhasard.net>
date Thu, 19 Dec 2013 18:13:11 +0000
parents 56144c8593a8
children 0cb4f494a548
line wrap: on
line diff
--- a/src/text.c	Tue Dec 17 20:49:52 2013 +0200
+++ b/src/text.c	Thu Dec 19 18:13:11 2013 +0000
@@ -2204,79 +2204,6 @@
 
 #ifdef MULE
 
-#ifdef EFFICIENT_INT_128_BIT
-# define STRIDE_TYPE INT_128_BIT
-# define HIGH_BIT_MASK \
-    MAKE_128_BIT_UNSIGNED_CONSTANT (0x80808080808080808080808080808080)
-#elif defined (EFFICIENT_INT_64_BIT)
-# define STRIDE_TYPE INT_64_BIT
-# define HIGH_BIT_MASK MAKE_64_BIT_UNSIGNED_CONSTANT (0x8080808080808080)
-#else
-# define STRIDE_TYPE INT_32_BIT
-# define HIGH_BIT_MASK MAKE_32_BIT_UNSIGNED_CONSTANT (0x80808080)
-#endif
-
-#define ALIGN_BITS ((EMACS_UINT) (ALIGNOF (STRIDE_TYPE) - 1))
-#define ALIGN_MASK (~ ALIGN_BITS)
-#define ALIGNED(ptr) ((((EMACS_UINT) ptr) & ALIGN_BITS) == 0)
-#define STRIDE sizeof (STRIDE_TYPE)
-
-/* Skip as many ASCII bytes as possible in the memory block [PTR, END).
-   Return pointer to the first non-ASCII byte.  optimized for long
-   stretches of ASCII. */
-inline static const Ibyte *
-skip_ascii (const Ibyte *ptr, const Ibyte *end)
-{
-  const unsigned STRIDE_TYPE *ascii_end;
-
-  /* Need to do in 3 sections -- before alignment start, aligned chunk,
-     after alignment end. */
-  while (!ALIGNED (ptr))
-    {
-      if (ptr == end || !byte_ascii_p (*ptr))
-	return ptr;
-      ptr++;
-    }
-  ascii_end = (const unsigned STRIDE_TYPE *) ptr;
-  /* This loop screams, because we can detect ASCII
-     characters 4 or 8 at a time. */
-  while ((const Ibyte *) ascii_end + STRIDE <= end
-	 && !(*ascii_end & HIGH_BIT_MASK))
-    ascii_end++;
-  ptr = (Ibyte *) ascii_end;
-  while (ptr < end && byte_ascii_p (*ptr))
-    ptr++;
-  return ptr;
-}
-
-/* Skip as many ASCII bytes as possible in the memory block [END, PTR),
-   going downwards.  Return pointer to the location above the first
-   non-ASCII byte.  Optimized for long stretches of ASCII. */
-inline static const Ibyte *
-skip_ascii_down (const Ibyte *ptr, const Ibyte *end)
-{
-  const unsigned STRIDE_TYPE *ascii_end;
-
-  /* Need to do in 3 sections -- before alignment start, aligned chunk,
-     after alignment end. */
-  while (!ALIGNED (ptr))
-    {
-      if (ptr == end || !byte_ascii_p (*(ptr - 1)))
-	return ptr;
-      ptr--;
-    }
-  ascii_end = (const unsigned STRIDE_TYPE *) ptr - 1;
-  /* This loop screams, because we can detect ASCII
-     characters 4 or 8 at a time. */
-  while ((const Ibyte *) ascii_end >= end
-	 && !(*ascii_end & HIGH_BIT_MASK))
-    ascii_end--;
-  ptr = (Ibyte *) (ascii_end + 1);
-  while (ptr > end && byte_ascii_p (*(ptr - 1)))
-    ptr--;
-  return ptr;
-}
-
 /* Function equivalents of bytecount_to_charcount/charcount_to_bytecount.
    These work on strings of all sizes but are more efficient than a simple
    loop on large strings and probably less efficient on sufficiently small