This is the mail archive of the guile@sourceware.cygnus.com mailing list for the Guile project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

[PATCH] scsh block comment fix + small speedup


OK, I looked up the scsh manual, it says block comments start with
#! and end with !# on a line by itself. This patch fixes the block
comment parsing. While I was at it, I also rewrote scm_flush_ws() to
make regular comments parse faster. On my P90, it shaves .02 seconds
off guile startup time.

This patch is for 1.3.4, as I can't get the CVS version to compile
(autoconf problems). I believe it will apply to the CVS version without
any problem though.

-- 
C. Ray C. aka Christopher Cramer
crayc@pyro.net
http://www.pyro.net/~crayc/
diff -ru old/guile-1.3.4/libguile/ports.c guile-1.3.4/libguile/ports.c
--- old/guile-1.3.4/libguile/ports.c	Mon Sep 20 16:34:57 1999
+++ guile-1.3.4/libguile/ports.c	Sat Apr 29 19:43:25 2000
@@ -703,14 +703,9 @@
   int c;
   scm_port *pt = SCM_PTAB_ENTRY (port);
 
-  if (pt->rw_active == SCM_PORT_WRITE)
-    {
-      /* may be marginally faster than calling scm_flush.  */
-      scm_ptobs[SCM_PTOBNUM (port)].flush (port);
-    }
+  if (pt->rw_active == SCM_PORT_WRITE) scm_flush(port);
   
-  if (pt->rw_random)
-    pt->rw_active = SCM_PORT_READ;
+  if (pt->rw_random) pt->rw_active = SCM_PORT_READ;
 
   if (pt->read_pos >= pt->read_end)
     {
diff -ru old/guile-1.3.4/libguile/read.c guile-1.3.4/libguile/read.c
--- old/guile-1.3.4/libguile/read.c	Mon Sep 20 16:34:57 1999
+++ guile-1.3.4/libguile/read.c	Sun Apr 30 18:12:08 2000
@@ -55,6 +55,7 @@
 #include "read.h"
 
 
+static void skip_scsh_block_comment (SCM port);
 
 SCM_SYMBOL (scm_keyword_prefix, "prefix");
 
@@ -122,41 +123,102 @@
   return SCM_CHARS (*tok_buf);
 }
 
+/* yes, I know this is ugly -crc */
+#define fill_if_needed(x,bp,be,eoferr)                          \
+  if (bp >= be)                                                 \
+    {                                                           \
+      if (scm_fill_input (port) == EOF)                         \
+	{                                                       \
+	  if (eoferr)                                           \
+	    scm_wta (SCM_UNDEFINED, "end of file in ", eoferr); \
+	  return EOF;                                           \
+	}                                                       \
+      bp = SCM_PTAB_ENTRY(x)->read_pos;                         \
+      be = SCM_PTAB_ENTRY(x)->read_end;                         \
+    }
 
+#define update_port(x,bp)          \
+  SCM_PTAB_ENTRY(x)->read_pos = bp;
 
+#define fill_regardless(x,bp,be,eoferr)                     \
+  if (scm_fill_input (port) == EOF)                         \
+    {                                                       \
+      if (eoferr)                                           \
+	scm_wta (SCM_UNDEFINED, "end of file in ", eoferr); \
+      return EOF;                                           \
+    }                                                       \
+  bp = SCM_PTAB_ENTRY(x)->read_pos;                         \
+  be = SCM_PTAB_ENTRY(x)->read_end;
+  
 int 
-scm_flush_ws (port, eoferr)
-     SCM port;
-     const char *eoferr;
+scm_flush_ws (SCM port, const char *eoferr)
 {
-  register int c;
+  int c;
+  scm_port *pt = SCM_PTAB_ENTRY (port);
+  const unsigned char *bp;
+  unsigned char *be;
+  int line;
+  int col;
+
+  bp = pt->read_pos;
+  be = pt->read_end;
+  line = SCM_LINUM (port);
+  col = SCM_COL (port);
+
   while (1)
-    switch (c = scm_getc (port))
-      {
-      case EOF:
-      goteof:
-	if (eoferr)
-	  scm_wta (SCM_UNDEFINED, "end of file in ", eoferr);
-	return c;
-      case ';':
-      lp:
-	switch (c = scm_getc (port))
-	  {
-	  case EOF:
-	    goto goteof;
-	  default:
-	    goto lp;
-	  case SCM_LINE_INCREMENTORS:
-	    break;
-	  }
-	break;
-      case SCM_LINE_INCREMENTORS:
-      case SCM_SINGLE_SPACES:
-      case '\t':
-	break;
-      default:
-	return c;
-      }
+    {
+      fill_if_needed(port,bp,be,eoferr)
+      switch (c = *bp++)
+	{
+	case ';':
+	  /* only works if SCM_LINE_INCREMENTORS is '\n' */
+	  fill_if_needed(port,bp,be,eoferr)
+	  while ((bp = memchr(bp, '\n', be - bp)) == NULL)
+	    {
+	      fill_regardless(port,bp,be,eoferr)
+	    }
+	  bp++;
+	  line++;
+	  col = 0;
+	  break;
+	case SCM_LINE_INCREMENTORS:
+	  line++;
+	  col = 0;
+	  break;
+	case SCM_SINGLE_SPACES:
+	  col++;
+	  break;
+	case '\t':
+	  col += 8 - col % 8;
+	  break;
+	case '#':
+	  fill_if_needed(port,bp,be,eoferr)
+	  if (*bp++ == '!')
+	    {
+	      update_port(port,bp)
+	      SCM_COL(port) = col + 2;
+	      SCM_LINUM(port) = line;
+
+	      skip_scsh_block_comment(port);
+
+	      bp = pt->read_pos;
+	      be = pt->read_end;
+	      line = SCM_LINUM (port);
+	      col = SCM_COL (port);
+	    } else {
+	      update_port(port,--bp)
+	      SCM_COL(port) = col;
+	      SCM_LINUM(port) = line;
+	      return c;
+	    }
+	  break;
+	default:
+	  update_port(port,bp)
+	  SCM_COL(port) = col;
+	  SCM_LINUM(port) = line;
+	  return c;
+	}
+    }
 }
 
 
@@ -242,8 +304,7 @@
    newline/exclamation-point/sharp-sign/newline sequence.  */
 
 static void
-skip_scsh_block_comment (port)
-     SCM port;
+skip_scsh_block_comment (SCM port)
 {
   /* Is this portable?  Dear God, spare me from the non-eight-bit
      characters.  But is it tasteful?  */
@@ -357,14 +418,6 @@
 	  scm_ungetc (c, port);
 	  c = '#';
 	  goto num;
-
-	case '!':
-	  /* start of a shell script.  Parse as a block comment,
-	     terminated by !#, just like SCSH.  */
-	  skip_scsh_block_comment (port);
-	  /* EOF is not an error here */
-	  c = scm_flush_ws (port, (char *)NULL);
-	  goto tryagain_no_flush_ws;
 
 	case '*':
 	  j = scm_read_token (c, tok_buf, port, 0);
diff -ru old/guile-1.3.4/libguile/read.h guile-1.3.4/libguile/read.h
--- old/guile-1.3.4/libguile/read.h	Sat Feb  6 04:30:48 1999
+++ guile-1.3.4/libguile/read.h	Sun Apr 30 15:30:00 2000
@@ -53,6 +53,9 @@
  * be incremented for the purposes of error reporting.  This feature
  * is only used for scheme code loaded from files.
  *
+ * WARNING: if you change SCM_LINE_INCREMENTORS you will break
+ * scm_flush_ws() in read.c -crc
+ *
  * SCM_WHITE_SPACES are other characters which should be treated like spaces
  * in programs.
  */

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]