MDEV-15592 - Column COMPRESSED should select a 'high order' datatype

Compressed blob columns didn't accept data at their capacity. E.g. storing
255 bytes to TINYBLOB results in "Data too long" error.

Now it is allowed assuming compression method was able to produce shorter
string (so that both metadata and compressed data fits blob) and
column_compression_threshold is lower than blob.

If no compression was performed, we still have to reserve additional byte
for metadata and thus we perform normal data truncation and return it's
status.

mysql-test/main/column_compression.result

@@ -1336,15 +1336,33 @@ a	LENGTH(a)
 DROP TABLE t1;
 CREATE TABLE t1(a TINYTEXT COMPRESSED);
 SET column_compression_threshold=300;
+INSERT INTO t1 VALUES(REPEAT('a', 254));
+INSERT INTO t1 VALUES(REPEAT(' ', 254));
 INSERT INTO t1 VALUES(REPEAT('a', 255));
 ERROR 22001: Data too long for column 'a' at row 1
 INSERT INTO t1 VALUES(REPEAT(' ', 255));
 Warnings:
 Note	1265	Data truncated for column 'a' at row 1
+INSERT INTO t1 VALUES(REPEAT('a', 256));
+ERROR 22001: Data too long for column 'a' at row 1
+INSERT INTO t1 VALUES(REPEAT(' ', 256));
+Warnings:
+Note	1265	Data truncated for column 'a' at row 1
+Note	1265	Data truncated for column 'a' at row 1
+INSERT INTO t1 VALUES(REPEAT('a', 257));
+ERROR 22001: Data too long for column 'a' at row 1
+INSERT INTO t1 VALUES(REPEAT(' ', 257));
+Warnings:
+Note	1265	Data truncated for column 'a' at row 1
+Note	1265	Data truncated for column 'a' at row 1
 SET column_compression_threshold=DEFAULT;
-SELECT a, LENGTH(a) FROM t1;
-a	LENGTH(a)
-                                                                                                                                                                                                                                                              	254
+SELECT LEFT(a, 10), LENGTH(a) FROM t1 ORDER BY 1;
+LEFT(a, 10)	LENGTH(a)
+          	254
+          	254
+          	254
+          	254
+aaaaaaaaaa	254
 DROP TABLE t1;
 # Corner case: VARCHAR(255) COMPRESSED must have 2 bytes pack length
 CREATE TABLE t1(a VARCHAR(255) COMPRESSED);
@@ -1360,6 +1378,32 @@ SELECT a, LENGTH(a) FROM t1;
 a	LENGTH(a)
 aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa	255
 DROP TABLE t1;
+CREATE TABLE t1(a VARCHAR(65531) COMPRESSED);
+SET column_compression_threshold=65537;
+INSERT INTO t1 VALUES(REPEAT('a', 65530));
+INSERT INTO t1 VALUES(REPEAT(' ', 65530));
+INSERT INTO t1 VALUES(REPEAT('a', 65531));
+INSERT INTO t1 VALUES(REPEAT(' ', 65531));
+INSERT INTO t1 VALUES(REPEAT('a', 65532));
+ERROR 22001: Data too long for column 'a' at row 1
+INSERT INTO t1 VALUES(REPEAT(' ', 65532));
+Warnings:
+Note	1265	Data truncated for column 'a' at row 1
+INSERT INTO t1 VALUES(REPEAT('a', 65533));
+ERROR 22001: Data too long for column 'a' at row 1
+INSERT INTO t1 VALUES(REPEAT(' ', 65533));
+Warnings:
+Note	1265	Data truncated for column 'a' at row 1
+SET column_compression_threshold=DEFAULT;
+SELECT LEFT(a, 10), LENGTH(a) FROM t1 ORDER BY 1, 2;
+LEFT(a, 10)	LENGTH(a)
+          	65530
+          	65531
+          	65531
+          	65531
+aaaaaaaaaa	65530
+aaaaaaaaaa	65531
+DROP TABLE t1;
 #
 # MDEV-14929 - AddressSanitizer: memcpy-param-overlap in
 #              Field_longstr::compress
@@ -1419,3 +1463,12 @@ COLUMN_NAME	CHARACTER_MAXIMUM_LENGTH	CHARACTER_OCTET_LENGTH
 a	10	10
 b	10	30
 DROP TABLE t1;
+#
+# MDEV-15592 - Column COMPRESSED should select a 'high order' datatype
+#
+CREATE TABLE t1(a TINYTEXT COMPRESSED);
+INSERT INTO t1 VALUES(REPEAT('a', 255));
+SELECT LEFT(a, 10), LENGTH(a) FROM t1;
+LEFT(a, 10)	LENGTH(a)
+aaaaaaaaaa	255
+DROP TABLE t1;

mysql-test/main/column_compression.test

@@ -64,11 +64,19 @@ DROP TABLE t1;
 
 CREATE TABLE t1(a TINYTEXT COMPRESSED);
 SET column_compression_threshold=300;
+INSERT INTO t1 VALUES(REPEAT('a', 254));
+INSERT INTO t1 VALUES(REPEAT(' ', 254));
 --error ER_DATA_TOO_LONG
 INSERT INTO t1 VALUES(REPEAT('a', 255));
 INSERT INTO t1 VALUES(REPEAT(' ', 255));
+--error ER_DATA_TOO_LONG
+INSERT INTO t1 VALUES(REPEAT('a', 256));
+INSERT INTO t1 VALUES(REPEAT(' ', 256));
+--error ER_DATA_TOO_LONG
+INSERT INTO t1 VALUES(REPEAT('a', 257));
+INSERT INTO t1 VALUES(REPEAT(' ', 257));
 SET column_compression_threshold=DEFAULT;
-SELECT a, LENGTH(a) FROM t1;
+SELECT LEFT(a, 10), LENGTH(a) FROM t1 ORDER BY 1;
 DROP TABLE t1;
 
 --echo # Corner case: VARCHAR(255) COMPRESSED must have 2 bytes pack length
@@ -80,6 +88,22 @@ SET column_compression_threshold=DEFAULT;
 SELECT a, LENGTH(a) FROM t1;
 DROP TABLE t1;
 
+CREATE TABLE t1(a VARCHAR(65531) COMPRESSED);
+SET column_compression_threshold=65537;
+INSERT INTO t1 VALUES(REPEAT('a', 65530));
+INSERT INTO t1 VALUES(REPEAT(' ', 65530));
+INSERT INTO t1 VALUES(REPEAT('a', 65531));
+INSERT INTO t1 VALUES(REPEAT(' ', 65531));
+--error ER_DATA_TOO_LONG
+INSERT INTO t1 VALUES(REPEAT('a', 65532));
+INSERT INTO t1 VALUES(REPEAT(' ', 65532));
+--error ER_DATA_TOO_LONG
+INSERT INTO t1 VALUES(REPEAT('a', 65533));
+INSERT INTO t1 VALUES(REPEAT(' ', 65533));
+SET column_compression_threshold=DEFAULT;
+SELECT LEFT(a, 10), LENGTH(a) FROM t1 ORDER BY 1, 2;
+DROP TABLE t1;
+
 
 --echo #
 --echo # MDEV-14929 - AddressSanitizer: memcpy-param-overlap in
@@ -113,6 +137,7 @@ INSERT INTO t1 VALUES('a');
 SET column_compression_threshold=DEFAULT;
 DROP TABLE t1;
 
+
 --echo #
 --echo # MDEV-15938 - TINYTEXT CHARACTER SET utf8 COMPRESSED truncates data
 --echo #
@@ -136,3 +161,12 @@ FROM INFORMATION_SCHEMA.COLUMNS
 WHERE TABLE_SCHEMA='test' AND TABLE_NAME='t1' AND COLUMN_NAME IN ('a','b')
 ORDER BY COLUMN_NAME;
 DROP TABLE t1;
+
+
+--echo #
+--echo # MDEV-15592 - Column COMPRESSED should select a 'high order' datatype
+--echo #
+CREATE TABLE t1(a TINYTEXT COMPRESSED);
+INSERT INTO t1 VALUES(REPEAT('a', 255));
+SELECT LEFT(a, 10), LENGTH(a) FROM t1;
+DROP TABLE t1;

sql/field.cc

@@ -7920,10 +7920,13 @@ void Field_varstring::hash(ulong *nr, ulong *nr2)
   Compress field
 
   @param[out]    to         destination buffer for compressed data
-  @param[in,out] to_length  in: size of to, out: compressed data length
+  @param[in]     to_length  size of to
   @param[in]     from       data to compress
   @param[in]     length     from length
+  @param[in]     max_length truncate `from' to this length
+  @param[out]    out_length compessed data length
   @param[in]     cs         from character set
+  @param[in]     nchars     copy no more than "nchars" characters
 
   In worst case (no compression performed) storage requirement is increased by
   1 byte to store header. If it exceeds field length, normal data truncation is
@@ -7947,42 +7950,57 @@ void Field_varstring::hash(ulong *nr, ulong *nr2)
   followed by compressed data.
 */
 
-int Field_longstr::compress(char *to, uint *to_length,
+int Field_longstr::compress(char *to, uint to_length,
                             const char *from, uint length,
+                            uint max_length,
+                            uint *out_length,
                             CHARSET_INFO *cs, size_t nchars)
 {
   THD *thd= get_thd();
-  char *buf= 0;
+  char *buf;
+  uint buf_length;
   int rc= 0;
 
   if (String::needs_conversion_on_storage(length, cs, field_charset) ||
-      *to_length <= length)
+      max_length < length)
   {
-    if (!(buf= (char*) my_malloc(*to_length - 1, MYF(MY_WME))))
+    set_if_smaller(max_length, static_cast<ulonglong>(field_charset->mbmaxlen) * length + 1);
+    if (!(buf= (char*) my_malloc(max_length, MYF(MY_WME))))
     {
-      *to_length= 0;
+      *out_length= 0;
       return -1;
     }
 
-    rc= well_formed_copy_with_check(buf, *to_length - 1, cs, from, length,
-                                    nchars, true, &length);
-    from= buf;
+    rc= well_formed_copy_with_check(buf, max_length, cs, from, length,
+                                    nchars, true, &buf_length);
+  }
+  else
+  {
+    buf= const_cast<char*>(from);
+    buf_length= length;
   }
 
-  if (length == 0)
-    *to_length= 0;
-  else if (length >= thd->variables.column_compression_threshold &&
-      (*to_length= compression_method()->compress(thd, to, from, length)))
+  if (buf_length == 0)
+    *out_length= 0;
+  else if (buf_length >= thd->variables.column_compression_threshold &&
+      (*out_length= compression_method()->compress(thd, to, buf, buf_length)))
     status_var_increment(thd->status_var.column_compressions);
   else
   {
     /* Store uncompressed */
     to[0]= 0;
-    memcpy(to + 1, from, length);
-    *to_length= length + 1;
+    if (buf_length < to_length)
+      memcpy(to + 1, buf, buf_length);
+    else
+    {
+      /* Storing string at blob capacity, e.g. 255 bytes string to TINYBLOB. */
+      rc= well_formed_copy_with_check(to + 1, to_length - 1, cs, from, length,
+                                      nchars, true, &buf_length);
+    }
+    *out_length= buf_length + 1;
   }
 
-  if (buf)
+  if (buf != from)
     my_free(buf);
   return rc;
 }
@@ -8036,10 +8054,12 @@ int Field_varstring_compressed::store(const char *from, size_t length,
                                       CHARSET_INFO *cs)
 {
   ASSERT_COLUMN_MARKED_FOR_WRITE_OR_COMPUTED;
-  uint to_length= (uint)MY_MIN(field_length, field_charset->mbmaxlen * length + 1);
-  int rc= compress((char*) get_data(), &to_length, from, (uint) length, cs,
-                   (to_length - 1) / field_charset->mbmaxlen);
-  store_length(to_length);
+  uint compressed_length;
+  int rc= compress((char*) get_data(), field_length, from, (uint) length,
+                   Field_varstring_compressed::max_display_length(),
+                   &compressed_length, cs,
+                   Field_varstring_compressed::char_length());
+  store_length(compressed_length);
   return rc;
 }
 
@@ -8648,7 +8668,10 @@ int Field_blob_compressed::store(const char *from, size_t length,
                                  CHARSET_INFO *cs)
 {
   ASSERT_COLUMN_MARKED_FOR_WRITE_OR_COMPUTED;
-  uint to_length= (uint)MY_MIN(max_data_length(), field_charset->mbmaxlen * length + 1);
+  uint compressed_length;
+  uint max_length= max_data_length();
+  uint to_length= (uint) MY_MIN(max_length,
+                                field_charset->mbmaxlen * length + 1);
   String tmp(from, length, cs);
   int rc;
 
@@ -8658,9 +8681,9 @@ int Field_blob_compressed::store(const char *from, size_t length,
   if (value.alloc(to_length))
     goto oom;
 
-  rc= compress((char*) value.ptr(), &to_length, tmp.ptr(), (uint) length, cs,
-               (uint) length);
-  set_ptr(to_length, (uchar*) value.ptr());
+  rc= compress((char*) value.ptr(), to_length, tmp.ptr(), (uint) length,
+               max_length, &compressed_length, cs, (uint) length);
+  set_ptr(compressed_length, (uchar*) value.ptr());
   return rc;
 
 oom:

sql/field.h

@@ -1790,8 +1790,10 @@ class Field_longstr :public Field_str
                                          const Item *item) const;
   bool cmp_to_string_with_stricter_collation(const Item_bool_func *cond,
                                              const Item *item) const;
-  int compress(char *to, uint *to_length,
+  int compress(char *to, uint to_length,
                const char *from, uint length,
+               uint max_length,
+               uint *out_length,
                CHARSET_INFO *cs, size_t nchars);
   String *uncompress(String *val_buffer, String *val_ptr,
                      const uchar *from, uint from_length);