>From 86d0a1f5ba2f195c3c398573cf8d377ea8111456 Mon Sep 17 00:00:00 2001
From: Assaf Gordon <assafgordon@gmail.com>
Date: Thu, 6 Jan 2022 12:24:40 -0700
Subject: [PATCH 4/9] cut: extract 'cut -D -f' to a separate function

---
 src/cut.c | 170 +++++++++++++++++++++++++++++-------------------------
 1 file changed, 93 insertions(+), 77 deletions(-)

diff --git a/src/cut.c b/src/cut.c
index 84caad091..369c47856 100644
--- a/src/cut.c
+++ b/src/cut.c
@@ -124,6 +124,9 @@ static bool adv_mode;
    and do not sort the output list */
 static bool allow_duplicates;
 
+/* Minor optimization: save a pointer to the last field pair sentinel
+   (which is always added by set_fields() */
+static struct field_range_pair *last_frp;
 
 /* For long options that have no equivalent short option, use a
    non-character as a pseudo short option, starting with CHAR_MAX + 1.  */
@@ -441,6 +444,87 @@ cut_fields (FILE *stream)
     }
 }
 
+static bool
+cut_adv_fields (char* linebuf, size_t len)
+{
+  static char **fieldpos = NULL ;
+  static idx_t alloc_flds = 0;
+
+  /* Split into fields */
+  char *p = linebuf;
+  size_t l = len;
+  idx_t fld = 0 ;
+  while (true)
+    {
+      char *endp = memchr (p, delim, l);
+
+      /* NUL-terminate the field if not the last */
+      if (endp)
+        *endp = '\0';
+
+      //fprintf(stderr,"Field %ld: '%s'\n", fld, p);
+
+      /* Store this field */
+      if (fld >= alloc_flds)
+        fieldpos = xpalloc (fieldpos, &alloc_flds, 10, -1, sizeof(char*));
+      fieldpos[fld] = p;
+      fld++;
+
+      if (!endp)
+        break;
+
+      l -= (endp-p+1);
+      p = endp+1;
+    }
+
+
+  bool output = false;
+
+  if (fld>1)
+    {
+      /* Iterate the requested field LIST, and print accordingly */
+      for (struct field_range_pair* r = frp; r != last_frp ; ++r)
+        {
+          /* If open-ended range, print up to the available fields */
+          uintmax_t hi = (r->hi == UINTMAX_MAX) ? fld : r->hi;
+
+          for (uintmax_t i = r->lo - 1 ; i < hi ; ++i )
+            {
+#if 0
+              fprintf(stderr,"Requested field: %zu\n", i);
+              fprintf(stderr,"Field %zu: '%s'\n", i, fieldpos[i-1]);
+#endif
+
+              if (i >=fld)
+                break;
+
+              if (output)
+                fwrite (output_delimiter_string, sizeof (char),
+                        output_delimiter_length, stdout);
+
+              fputs (fieldpos[i], stdout);
+              output = true;
+            }
+        }
+    }
+
+  /* Print non-delimited lines */
+  if (!output && fld==1)
+    {
+      if (!suppress_non_delimited)
+        {
+          fputs(linebuf, stdout);
+          output = true;
+        }
+    }
+
+  IF_LINT (free (fieldpos));
+  IF_LINT (fieldpos = NULL);
+  IF_LINT (alloc_flds = 0);
+
+  return output || fld>1;
+}
+
 static void
 cut_adv (FILE *stream)
 {
@@ -448,14 +532,6 @@ cut_adv (FILE *stream)
   size_t bufsize = 0;
   ssize_t len;
 
-  char **fieldpos = NULL ;
-  idx_t alloc_flds = 0;
-
-  /* Minor optimization: save a pointer to the last field pair sentinel
-     (which is always added by set_fields() */
-  struct field_range_pair *last_frp = frp;
-  while ( ! (last_frp->hi==UINTMAX_MAX && last_frp->lo==UINTMAX_MAX) )
-    ++last_frp;
 
   while (true)
     {
@@ -478,81 +554,15 @@ cut_adv (FILE *stream)
           --len;
         }
 
-      /* Split into fields */
-      char *p = linebuf;
-      size_t l = len;
-      idx_t fld = 0 ;
-      while (true)
-        {
-          char *endp = memchr (p, delim, l);
-
-          /* NUL-terminate the field if not the last */
-          if (endp)
-            *endp = '\0';
-
-          //fprintf(stderr,"Field %ld: '%s'\n", fld, p);
-
-          /* Store this field */
-          if (fld >= alloc_flds)
-            fieldpos = xpalloc (fieldpos, &alloc_flds, 10, -1, sizeof(char*));
-          fieldpos[fld] = p;
-          fld++;
-
-          if (!endp)
-            break;
-
-          l -= (endp-p+1);
-          p = endp+1;
-        }
-
 
-      bool first = true;
-
-      if (fld>1)
-        {
-          /* Iterate the requested field LIST, and print accordingly */
-          for (struct field_range_pair* r = frp; r != last_frp ; ++r)
-            {
-              /* If open-ended range, print up to the available fields */
-              uintmax_t hi = (r->hi == UINTMAX_MAX) ? fld : r->hi;
-
-              for (uintmax_t i = r->lo - 1 ; i < hi ; ++i )
-                {
-#if 0
-                  fprintf(stderr,"Requested field: %zu\n", i);
-                  fprintf(stderr,"Field %zu: '%s'\n", i, fieldpos[i-1]);
-#endif
-
-                  if (i >=fld)
-                    break;
-
-                  if (!first)
-                    fwrite (output_delimiter_string, sizeof (char),
-                            output_delimiter_length, stdout);
-
-                  fputs (fieldpos[i], stdout);
-                  first = false;
-                }
-            }
-        }
-
-      /* Print non-delimited lines */
-      if (first && fld==1)
-        {
-          if (!suppress_non_delimited)
-            {
-              fputs(linebuf, stdout);
-              putchar (line_delim);
-            }
-          continue;
-        }
+      bool output = cut_adv_fields (linebuf, len);
 
       //fprintf(stderr,"end of line\n");
-      putchar (line_delim);
+      if (output)
+        putchar (line_delim);
 
     }
 
-  free (fieldpos);
   free (linebuf);
 }
 
@@ -724,6 +734,12 @@ main (int argc, char **argv)
               | (complement ? SETFLD_COMPLEMENT : 0)
               | (allow_duplicates ? SETFLD_NO_SORT : 0) );
 
+  /* Minor optimization: keep a pointer to the sentinel (last) pair */
+  last_frp = frp;
+  while ( ! (last_frp->hi==UINTMAX_MAX && last_frp->lo==UINTMAX_MAX) )
+    ++last_frp;
+
+
   if (!delim_specified)
     delim = '\t';
 
-- 
2.20.1