missing-values: Make mv_add_str() easier to use.

Both existing callers of mv_add_str() were applying their own padding to
the string value.  This commit moves the padding into mv_add_str() itself.

This commit also fixes a minor to-do item in the MISSING VALUES command
whereby truncation of too-long missing values was done bytewise rather than
on the basis of full characters.

diff --git a/src/data/missing-values.c b/src/data/missing-values.c
index 0fc9d01af01ae5a7842aebf3a8285217485de3ca..9aba57e0e05fb6249764d2c5f778fb6337275a82 100644 (file)
--- a/src/data/missing-values.c
+++ b/src/data/missing-values.c
@@ -1,5 +1,5 @@
 /* PSPP - a program for statistical analysis.
-   Copyright (C) 2005, 2009, 2011 Free Software Foundation, Inc.
+   Copyright (C) 2005, 2009, 2011, 2013 Free Software Foundation, Inc.
 
    This program is free software: you can redistribute it and/or modify
    it under the terms of the GNU General Public License as published by
@@ -23,6 +23,7 @@
 
 #include "data/variable.h"
 #include "libpspp/assertion.h"
+#include "libpspp/cast.h"
 #include "libpspp/str.h"
 
 /* Types of user-missing values.
@@ -158,19 +159,23 @@ mv_add_value (struct missing_values *mv, const union value *v)
   NOT_REACHED ();
 }
 
-/* Attempts to add S to the set of string missing values MV.  S
-   must contain exactly as many characters as MV's width.
-   Returns true if successful, false if MV has no more room for
+/* Attempts to add S, which is LEN bytes long, to the set of string missing
+   values MV.  Returns true if successful, false if MV has no more room for
    missing values or if S is not an acceptable missing value. */
 bool
-mv_add_str (struct missing_values *mv, const uint8_t s[])
+mv_add_str (struct missing_values *mv, const uint8_t s[], size_t len)
 {
   union value v;
   bool ok;
 
   assert (mv->width > 0);
+  while (len > mv->width)
+    if (s[--len] != ' ')
+      return false;
+
   value_init (&v, mv->width);
-  memcpy (value_str_rw (&v, mv->width), s, mv->width);
+  buf_copy_rpad (CHAR_CAST (char *, value_str_rw (&v, mv->width)), mv->width,
+                 CHAR_CAST (char *, s), len, ' ');
   ok = mv_add_value (mv, &v);
   value_destroy (&v, mv->width);
 

diff --git a/src/data/missing-values.h b/src/data/missing-values.h
index 4d046faec3a1be411c4b5be5b2f0e93881b0e182..511ebd7ddd37273a9f2df1b6ad1ac0c31227e792 100644 (file)
--- a/src/data/missing-values.h
+++ b/src/data/missing-values.h
@@ -1,5 +1,5 @@
 /* PSPP - a program for statistical analysis.
-   Copyright (C) 2005, 2009 Free Software Foundation, Inc.
+   Copyright (C) 2005, 2009, 2013 Free Software Foundation, Inc.
 
    This program is free software: you can redistribute it and/or modify
    it under the terms of the GNU General Public License as published by
@@ -94,7 +94,7 @@ void mv_get_range (const struct missing_values *, double *low, double *high);
 
 /* Adding and modifying discrete values. */
 bool mv_add_value (struct missing_values *, const union value *);
-bool mv_add_str (struct missing_values *, const uint8_t[]);
+bool mv_add_str (struct missing_values *, const uint8_t[], size_t len);
 bool mv_add_num (struct missing_values *, double);
 void mv_pop_value (struct missing_values *, union value *);
 bool mv_replace_value (struct missing_values *, const union value *, int idx);

diff --git a/src/data/sys-file-reader.c b/src/data/sys-file-reader.c
index 681e6f531b152a61cc8b7a030893fe715a2c063a..c4162498724b14474ee8d0da456b2f3c5c07ebc1 100644 (file)
--- a/src/data/sys-file-reader.c
+++ b/src/data/sys-file-reader.c
@@ -1085,11 +1085,7 @@ parse_variable_records (struct sfm_reader *r, struct dictionary *dict,
               value_init_pool (r->pool, &value, width);
               value_set_missing (&value, width);
               for (i = 0; i < rec->missing_value_code; i++)
-                {
-                  uint8_t *s = value_str_rw (&value, width);
-                  memcpy (s, rec->missing + 8 * i, MIN (width, 8));
-                  mv_add_str (&mv, s);
-                }
+                mv_add_str (&mv, rec->missing + 8 * i, MIN (width, 8));
             }
           var_set_missing_values (var, &mv);
         }

diff --git a/src/language/dictionary/missing-values.c b/src/language/dictionary/missing-values.c
index a906b8a8f3cacbd889bce5b9be004357e687dc39..a369c2ea6999a802f00b9d8b723f244d2f4a2ffe 100644 (file)
--- a/src/language/dictionary/missing-values.c
+++ b/src/language/dictionary/missing-values.c
@@ -1,5 +1,5 @@
 /* PSPP - a program for statistical analysis.
-   Copyright (C) 1997-9, 2000, 2006, 2009, 2010, 2011 Free Software Foundation, Inc.
+   Copyright (C) 1997-9, 2000, 2006, 2009, 2010, 2011, 2013 Free Software Foundation, Inc.
 
    This program is free software: you can redistribute it and/or modify
    it under the terms of the GNU General Public License as published by
@@ -96,12 +96,16 @@ cmd_missing_values (struct lexer *lexer, struct dataset *ds)
             }
           else
             {
+              const char *encoding = dict_get_encoding (dict);
+
               mv_init (&mv, MV_MAX_STRING);
               while (!lex_match (lexer, T_RPAREN))
                 {
-                  uint8_t value[MV_MAX_STRING];
-                  char *dict_mv;
-                  size_t length;
+                  const char *utf8_s;
+                  size_t utf8_trunc_len;
+                  size_t utf8_len;
+
+                  char *raw_s;
 
                   if (!lex_force_string (lexer))
                     {
@@ -109,24 +113,24 @@ cmd_missing_values (struct lexer *lexer, struct dataset *ds)
                       break;
                     }
 
-                  dict_mv = recode_string (dict_get_encoding (dict), "UTF-8",
-                                           lex_tokcstr (lexer),
-                                           ss_length (lex_tokss (lexer)));
-                  length = strlen (dict_mv);
-                  if (length > MV_MAX_STRING)
-                    {
-                      /* XXX truncate graphemes not bytes */
-                      msg (SE, _("Truncating missing value to maximum "
-                                 "acceptable length (%d bytes)."),
-                           MV_MAX_STRING);
-                      length = MV_MAX_STRING;
-                    }
-                  memset (value, ' ', MV_MAX_STRING);
-                  memcpy (value, dict_mv, length);
-                  free (dict_mv);
-
-                  if (!mv_add_str (&mv, value))
+                  /* Truncate the string to fit in 8 bytes in the dictionary
+                     encoding. */
+                  utf8_s = lex_tokcstr (lexer);
+                  utf8_len = ss_length (lex_tokss (lexer));
+                  utf8_trunc_len = utf8_encoding_trunc_len (utf8_s, encoding,
+                                                            MV_MAX_STRING);
+                  if (utf8_trunc_len < utf8_len)
+                    msg (SE, _("Truncating missing value to maximum "
+                               "acceptable length (%d bytes)."),
+                         MV_MAX_STRING);
+
+                  /* Recode to dictionary encoding and add. */
+                  raw_s = recode_string (encoding, "UTF-8",
+                                         utf8_s, utf8_trunc_len);
+                  if (!mv_add_str (&mv, CHAR_CAST (const uint8_t *, raw_s),
+                                   strlen (raw_s)))
                     ok = false;
+                  free (raw_s);
 
                   lex_get (lexer);
                   lex_match (lexer, T_COMMA);