adlint 2.6.2 → 2.6.10

data/ChangeLog

@@ -1,3 +1,152 @@
+Thu Nov  8 13:01:03 2012  Yutaka Yanoh  <yanoh@users.sourceforge.net>
+
+	* release.ga : 2.6.10
+	  - Fix code checks so that duplicative messages of the same ID at the
+	    same location can be unique till the function step-in analysis is
+	    supported.
+	  - Fix W1069 detection to output a warning message at correct location
+	    when an incomplete if-else statements chain appears in a complete
+	    if-else-statement.
+	  - Fix W0708 detection not to warn when an address of the controlling
+	    variable is passes to a function in the controlling part of
+	    for-statement.
+	  - Fix W0534, W0585, W0611 and w0708 detections to correctly deduct
+	    the controlling variable.
+	  - Add missing notification of the controlling expression value
+	    reference.
+	  - Fix W0460 detection to output outermost variable name when the
+	    referencing variable is an array element or a member of struct or
+	    union.
+
+Wed Nov  7 15:17:28 2012  Yutaka Yanoh  <yanoh@users.sourceforge.net>
+
+	* release.rc : 2.6.9
+	  - Fix W0460 detection to output outermost variable name when the
+	    referencing variable is an array element or a member of struct or
+	    union.
+	  - Add missing notification of the controlling expression value
+	    reference and sequence-point at the end of the 1st expression
+	    evaluation in the conditional-expression.
+
+Wed Nov  7 14:10:54 2012  Yutaka Yanoh  <yanoh@users.sourceforge.net>
+
+	* features/code_check/W0599.feature : Add W0599 specification to
+	  reproduce a known bug.
+
+	* lib/adlint/c/interp.rb : Add missing notification of the controlling
+	  expression value reference and sequence-point at the end of the 1st
+	  expression evaluation in the conditional-expression.
+
+Wed Nov  7 13:37:45 2012  Yutaka Yanoh  <yanoh@users.sourceforge.net>
+
+	* lib/adlint/exam/c_builtin/c_check.rb : Fix W0460 detection to output
+	  outermost variable name when the referencing variable is an array
+	  element or a member of struct or union.
+
+Wed Nov  7 11:19:41 2012  Yutaka Yanoh  <yanoh@users.sourceforge.net>
+
+	* release.rc : 2.6.7
+	  - Move value reference notification of whole controlling expression
+	    to notification suppressible section.
+
+Wed Nov  7 11:15:59 2012  Yutaka Yanoh  <yanoh@users.sourceforge.net>
+
+	* features/code_check/W0599.feature : Add W0599 specification to
+	  reproduce a known bug.
+
+	* lib/adlint/c/ctrlexpr.rb : Move value reference notification of whole
+	  controlling expression to notification suppressible section.
+
+Wed Nov  7 10:01:25 2012  Yutaka Yanoh  <yanoh@users.sourceforge.net>
+
+	* release.rc : 2.6.5
+	  - Add missing sequence-point notification after referencing value of
+	    whole controlling expression in the controlling-statement.
+
+Wed Nov  7 09:57:52 2012  Yutaka Yanoh  <yanoh@users.sourceforge.net>
+
+	* features/code_check/W0599.feature : Add W0599 specification to
+	  reproduce a known bug.
+
+	* lib/adlint/c/interp.rb : Add missing sequence-point notification
+	  after referencing value of whole controlling expression in the
+	  controlling-statement.
+
+Tue Nov  6 16:48:25 2012  Yutaka Yanoh  <yanoh@users.sourceforge.net>
+
+	* release.rc : 2.6.3
+	  - Fix code checks so that duplicative messages of the same ID at the
+	    same location can be unique till the function step-in analysis is
+	    supported.
+	  - Fix W1069 detection to output a warning message at correct location
+	    when an incomplete if-else statements chain appears in a complete
+	    if-else-statement.
+	  - Fix W0708 detection not to warn when an address of the controlling
+	    variable is passes to a function in the controlling part of
+	    for-statement.
+	  - Fix W0534, W0585, W0611 and w0708 detections to correctly deduct
+	    the controlling variable.
+	  - Add missing notification of the controlling expression value
+	    reference.
+
+Tue Nov  6 16:17:03 2012  Yutaka Yanoh  <yanoh@users.sourceforge.net>
+
+	* features/code_check/W1073.feature : Add W1073 specification to
+	  reproduce a known bug.
+
+	* lib/adlint/c/interp.rb : Add missing notification of the controlling
+	  expression value reference.
+	* lib/adlint/c/ctrlexpr.rb : Ditto.
+
+Tue Nov  6 13:35:28 2012  Yutaka Yanoh  <yanoh@users.sourceforge.net>
+
+	* features/code_check/W0611.feature : Add W0611 specification.
+	* MANIFEST : Ditto.
+
+	* lib/adlint/exam/c_builtin/c_check.rb : Fix W0611 detection to
+	  correctly deduct controlling variable of the iteration-statement.
+
+Tue Nov  6 11:46:49 2012  Yutaka Yanoh  <yanoh@users.sourceforge.net>
+
+	* features/code_check/W0534.feature : Add W0534 specification.
+	* MANIFEST : Ditto.
+
+	* lib/adlint/exam/c_builtin/c_check.rb : Fix W0534 detection not to
+	  warn about const variable appears in the controlling expression.
+
+Tue Nov  6 10:57:45 2012  Yutaka Yanoh  <yanoh@users.sourceforge.net>
+
+	* features/code_check/W0585.feature : Add W0585 specification.
+	* MANIFEST : Ditto.
+
+	* lib/adlint/exam/c_builtin/c_check.rb : Fix W0585 detection not to
+	  warn about const variable appears in the controlling expression.
+
+Mon Nov  5 15:17:58 2012  Yutaka Yanoh  <yanoh@users.sourceforge.net>
+
+	* features/code_check/W0708.feature : Add W0708 specification.
+	* MANIFEST : Ditto.
+
+	* lib/adlint/exam/c_builtin/c_check.rb : Fix W0708 detection not to
+	  warn when an address of the controlling variable is passes to a
+	  function in the controlling part of for-statement.
+
+Fri Nov  2 17:21:20 2012  Yutaka Yanoh  <yanoh@users.sourceforge.net>
+
+	* features/code_check/W1069.feature : Add W1069 specification to
+	  reproduce a known bug.
+
+	* lib/adlint/exam/c_builtin/c_check_shima.rb : Fix W1069 detection to
+	  output a warning message at correct location when an incomplete
+	  if-else statements chain appears in a complete if-else-statement.
+
+Thu Nov  1 17:17:10 2012  Yutaka Yanoh  <yanoh@users.sourceforge.net>
+
+	* lib/adlint/exam/c_builtin/c_check.rb : Fix code checks so that
+	  duplicative messages of the same ID at the same location can be
+	  unique till the function step-in analysis is supported.
+	* lib/adlint/exam/c_builtin/c_check_shima.rb : Ditto.
+
 Wed Oct 31 10:16:44 2012  Yutaka Yanoh  <yanoh@users.sourceforge.net>
 
 	* release.ga : 2.6.2

data/MANIFEST

@@ -188,6 +188,7 @@ features/code_check/W0254.feature
 features/code_check/W0425.feature
 features/code_check/W0431.feature
 features/code_check/W0432.feature
+features/code_check/W0460.feature
 features/code_check/W0477.feature
 features/code_check/W0478.feature
 features/code_check/W0479.feature
@@ -206,12 +207,15 @@ features/code_check/W0499.feature
 features/code_check/W0500.feature
 features/code_check/W0501.feature
 features/code_check/W0502.feature
+features/code_check/W0534.feature
 features/code_check/W0573.feature
 features/code_check/W0582.feature
 features/code_check/W0583.feature
 features/code_check/W0584.feature
+features/code_check/W0585.feature
 features/code_check/W0599.feature
 features/code_check/W0606.feature
+features/code_check/W0611.feature
 features/code_check/W0635.feature
 features/code_check/W0641.feature
 features/code_check/W0643.feature
@@ -238,6 +242,7 @@ features/code_check/W0700.feature
 features/code_check/W0703.feature
 features/code_check/W0705.feature
 features/code_check/W0707.feature
+features/code_check/W0708.feature
 features/code_check/W0711.feature
 features/code_check/W0712.feature
 features/code_check/W0713.feature

data/NEWS

@@ -21,6 +21,26 @@
 
 ++
 
+=== \AdLint 2.6.10 is released (2012-11-08)
+
+==== Changes since the 2.6.2 release
+
+* Fix code checks so that duplicative messages of the same ID at the same
+  location can be unique till the function step-in analysis is supported
+* Fix W1069 detection to output a warning message at correct location when an
+  incomplete if-else statements chain appears in a complete if-else-statement
+* Fix W0708 detection not to warn when an address of the controlling variable
+  is passes to a function in the controlling part of for-statement
+* Fix W0534, W0585, W0611 and w0708 detections to correctly deduct the
+  controlling variable
+* Add missing notification of the controlling expression value reference
+* Fix W0460 detection to output outermost variable name when the referencing
+  variable is an array element or a member of struct or union
+
+See the file
+{ChangeLog}[http://adlint.sourceforge.net/pmwiki/upload.d/Main/ChangeLog]
+for more details.
+
 === \AdLint 2.6.2 is released (2012-10-31)
 
 ==== Changes since the 2.6.0 release
@@ -40,10 +60,6 @@
 * Fix the interpreter to place sequence-point at the end of the
   return-statement evaluation
 
-See the file
-{ChangeLog}[http://adlint.sourceforge.net/pmwiki/upload.d/Main/ChangeLog]
-for more details.
-
 === \AdLint 2.6.0 is released (2012-10-22)
 
 ==== Changes since the 2.4.10 release

data/etc/mesg.d/c_builtin/en_US/messages.yml

@@ -49,7 +49,7 @@
 #         - "ERR:X99"
 #       format: "Your custom message for the error of E9999."
 
-version: "2.6.2"
+version: "2.6.10"
 
 message_definition:
   W0001:

data/etc/mesg.d/c_builtin/ja_JP/messages.yml

@@ -49,7 +49,7 @@
 #         - "ERR:X99"
 #       format: "Your custom message for the error of E9999."
 
-version: "2.6.2"
+version: "2.6.10"
 
 message_definition:
   W0001:

data/etc/mesg.d/core/en_US/messages.yml

@@ -49,7 +49,7 @@
 #         - "ERR:X99"
 #       format: "Your custom message for the error of E9999."
 
-version: "2.6.2"
+version: "2.6.10"
 
 message_definition:
   X0001:

data/etc/mesg.d/core/ja_JP/messages.yml

@@ -49,7 +49,7 @@
 #         - "ERR:X99"
 #       format: "Your custom message for the error of E9999."
 
-version: "2.6.2"
+version: "2.6.10"
 
 message_definition:
   X0001:

data/features/code_check/W0460.feature

@@ -0,0 +1,27 @@
+Feature: W0460
+
+  W0460 detects that a value of the variable is not possibly initialized.
+
+  Scenario: array element is not possibly initialized
+    Given a target source named "fixture.c" with:
+      """
+      static int foo(void)
+      {
+          int a[5];
+
+          if (a[1] == 0) { /* W0459 */
+              a[0] = 0;
+          }
+
+          return a[0]; /* W0460 */
+      }
+      """
+    When I successfully run `adlint fixture.c` on noarch
+    Then the output should exactly match with:
+      | mesg  | line | column |
+      | W1076 | 1    | 12     |
+      | W0459 | 5    | 14     |
+      | W0460 | 9    | 13     |
+      | W0629 | 1    | 12     |
+      | W0950 | 3    | 11     |
+      | W0628 | 1    | 12     |

data/features/code_check/W0534.feature

@@ -0,0 +1,73 @@
+Feature: W0534
+
+  W0534 detects that a controlling variable of for-statement is not initialized
+  in the 1st part of for-statement.
+
+  Scenario: the 1st part of for-statement does not initialize the controlling
+            variable
+    Given a target source named "fixture.c" with:
+      """
+      static int bar(int *);
+
+      static int foo(const int num)
+      {
+          int i = 0, j;
+          for (j = 0; (bar(&i)) && (i < num); i++) { /* W0534 */
+              j++;
+          }
+          return j;
+      }
+      """
+    When I successfully run `adlint fixture.c` on noarch
+    Then the output should exactly match with:
+      | mesg  | line | column |
+      | W1076 | 3    | 12     |
+      | W0534 | 6    | 10     |
+      | W0629 | 3    | 12     |
+      | W0425 | 5    | 16     |
+      | W0628 | 3    | 12     |
+
+  Scenario: the 1st part of for-statement is empty
+    Given a target source named "fixture.c" with:
+      """
+      static int bar(int *);
+
+      static int foo(const int num)
+      {
+          int i = 0, j = 0;
+          for (; (bar(&i)) && (i < num); i++) { /* W0534 */
+              j++;
+          }
+          return j;
+      }
+      """
+    When I successfully run `adlint fixture.c` on noarch
+    Then the output should exactly match with:
+      | mesg  | line | column |
+      | W1076 | 3    | 12     |
+      | W0534 | 6    | 10     |
+      | W0629 | 3    | 12     |
+      | W0425 | 5    | 16     |
+      | W0628 | 3    | 12     |
+
+  Scenario: the 1st part of for-statement initializes the controlling variable
+    Given a target source named "fixture.c" with:
+      """
+      static int bar(int *);
+
+      static int foo(const int num)
+      {
+          int i, j = 0;
+          for (i = 0; (bar(&i)) && (i < num); i++) { /* OK */
+              j++;
+          }
+          return j;
+      }
+      """
+    When I successfully run `adlint fixture.c` on noarch
+    Then the output should exactly match with:
+      | mesg  | line | column |
+      | W1076 | 3    | 12     |
+      | W0629 | 3    | 12     |
+      | W0425 | 5    | 12     |
+      | W0628 | 3    | 12     |

data/features/code_check/W0585.feature

@@ -0,0 +1,135 @@
+Feature: W0585
+
+  W0585 detects that a controlling variable of for-statement is not updated in
+  the 3rd part of for-statement.
+
+  Scenario: the 3rd part of for-statement does not increment the controlling
+            variable
+    Given a target source named "fixture.c" with:
+      """
+      static int bar(int *);
+
+      static int foo(const int num)
+      {
+          int i, j = 0;
+          for (i = 0; (bar(&i)) && (i < num); j++) { /* W0585 */
+          }
+          return j;
+      }
+      """
+    When I successfully run `adlint fixture.c` on noarch
+    Then the output should exactly match with:
+      | mesg  | line | column |
+      | W1076 | 3    | 12     |
+      | W0585 | 6    | 42     |
+      | W0629 | 3    | 12     |
+      | W0425 | 5    | 12     |
+      | W0628 | 3    | 12     |
+
+  Scenario: the 3rd part of for-statement is empty
+    Given a target source named "fixture.c" with:
+      """
+      static int bar(int *);
+
+      static int foo(const int num)
+      {
+          int i, j = 0;
+          for (i = 0; (bar(&i)) && (i < num); ) { /* OK */
+              j++;
+          }
+          return j;
+      }
+      """
+    When I successfully run `adlint fixture.c` on noarch
+    Then the output should exactly match with:
+      | mesg  | line | column |
+      | W1076 | 3    | 12     |
+      | W0629 | 3    | 12     |
+      | W0425 | 5    | 12     |
+      | W0628 | 3    | 12     |
+
+  Scenario: the 3rd part of for-statement increments the controlling variable
+    Given a target source named "fixture.c" with:
+      """
+      static int bar(int *);
+
+      static int foo(const int num)
+      {
+          int i, j = 0;
+          for (i = 0; (bar(&i)) && (i < num); i++) { /* OK */
+              j++;
+          }
+          return j;
+      }
+      """
+    When I successfully run `adlint fixture.c` on noarch
+    Then the output should exactly match with:
+      | mesg  | line | column |
+      | W1076 | 3    | 12     |
+      | W0629 | 3    | 12     |
+      | W0425 | 5    | 12     |
+      | W0628 | 3    | 12     |
+
+  Scenario: the 3rd part of c99-for-statement does not increment the
+            controlling variable
+    Given a target source named "fixture.c" with:
+      """
+      static int bar(int *);
+
+      static int foo(const int num)
+      {
+          int j = 0;
+          for (int i = 0; (bar(&i)) && (i < num); j++) { /* W0585 */
+          }
+          return j;
+      }
+      """
+    When I successfully run `adlint fixture.c` on noarch
+    Then the output should exactly match with:
+      | mesg  | line | column |
+      | W1076 | 3    | 12     |
+      | W0585 | 6    | 46     |
+      | W0629 | 3    | 12     |
+      | W0628 | 3    | 12     |
+
+  Scenario: the 3rd part of c99-for-statement is empty
+    Given a target source named "fixture.c" with:
+      """
+      static int bar(int *);
+
+      static int foo(const int num)
+      {
+          int j = 0;
+          for (int i = 0; (bar(&i)) && (i < num); ) { /* OK */
+              j++;
+          }
+          return j;
+      }
+      """
+    When I successfully run `adlint fixture.c` on noarch
+    Then the output should exactly match with:
+      | mesg  | line | column |
+      | W1076 | 3    | 12     |
+      | W0629 | 3    | 12     |
+      | W0628 | 3    | 12     |
+
+  Scenario: the 3rd part of for-statement increments the controlling variable
+    Given a target source named "fixture.c" with:
+      """
+      static int bar(int *);
+
+      static int foo(const int num)
+      {
+          int j = 0;
+          for (int i = 0; (bar(&i)) && (i < num); i++) { /* OK */
+              j++;
+          }
+          return j;
+      }
+      """
+    When I successfully run `adlint fixture.c` on noarch
+    Then the output should exactly match with:
+      | mesg  | line | column |
+      | W1076 | 3    | 12     |
+      | W0629 | 3    | 12     |
+      | W0628 | 3    | 12     |

data/features/code_check/W0599.feature

@@ -7,7 +7,7 @@ Feature: W0599
     Given a target source named "fixture.c" with:
       """
       extern void bar(int, int, int);
-      
+
       static void foo(int a, int b, int *c)
       {
           b = (a + 1) + a++; /* W0599 */
@@ -69,3 +69,103 @@ Feature: W0599
       | W1071 | 1    | 12     |
       | W0629 | 1    | 12     |
       | W0628 | 1    | 12     |
+
+  Scenario: object-specifier as a controlling expression and updating same
+            variable in the branch
+    Given a target source named "fixture.c" with:
+      """
+      static void foo(int i)
+      {
+          if (i) {
+              i++; /* OK */
+          }
+      }
+      """
+    When I successfully run `adlint fixture.c` on noarch
+    Then the output should exactly match with:
+      | mesg  | line | column |
+      | W1076 | 1    | 13     |
+      | W0629 | 1    | 13     |
+      | W0114 | 3    | 5      |
+      | W0628 | 1    | 13     |
+
+  Scenario: assignment-expression as a controlling expression and updating same
+            variable in the branch
+    Given a target source named "fixture.c" with:
+      """
+      extern int bar(void);
+
+      static void foo(int i)
+      {
+          while (i = bar()) {
+              i++; /* OK */
+          }
+      }
+      """
+    When I successfully run `adlint fixture.c` on noarch
+    Then the output should exactly match with:
+      | mesg  | line | column |
+      | W0118 | 1    | 12     |
+      | W1076 | 3    | 13     |
+      | W0629 | 3    | 13     |
+      | W0108 | 5    | 14     |
+      | W0114 | 5    | 5      |
+      | W0628 | 3    | 13     |
+
+  Scenario: assignment-expression as a controlling expression and updating same
+            variable in the branch
+    Given a target source named "fixture.c" with:
+      """
+      extern int bar(int);
+
+      static void foo(void)
+      {
+          int i;
+          for (i = 0; i = bar(i); ) {
+              i++; /* OK */
+          }
+          i++; /* OK */
+      }
+      """
+    When I successfully run `adlint fixture.c` on noarch
+    Then the output should exactly match with:
+      | mesg  | line | column |
+      | W0118 | 1    | 12     |
+      | W1076 | 3    | 13     |
+      | W0708 | 7    | 10     |
+      | W0629 | 3    | 13     |
+      | W0108 | 6    | 19     |
+      | W0114 | 6    | 5      |
+      | W0628 | 3    | 13     |
+
+  Scenario: updating variable in the 1st expression of conditional-expression
+    Given a target source named "fixture.c" with:
+      """
+      static int foo(int i)
+      {
+          return i++ ? i : 0; /* OK */
+      }
+      """
+    When I successfully run `adlint fixture.c` on noarch
+    Then the output should exactly match with:
+      | mesg  | line | column |
+      | W1076 | 1    | 12     |
+      | W0629 | 1    | 12     |
+      | W0114 | 3    | 13     |
+      | W0628 | 1    | 12     |
+
+  Scenario: updating variable in the 1st expression of conditional-expression
+    Given a target source named "fixture.c" with:
+      """
+      static int foo(int i)
+      {
+          return i++ > 0 ? i : 0; /* OK */
+      }
+      """
+    When I successfully run `adlint fixture.c` on noarch
+    Then the output should exactly match with:
+      | mesg  | line | column |
+      | W1076 | 1    | 12     |
+      | W0629 | 1    | 12     |
+      | W0501 | 3    | 20     |
+      | W0628 | 1    | 12     |