+2009-04-15 Bob Duff <duff@adacore.com>
+
+ * sem_ch5.adb (Analyze_Loop_Statement): Don't check for infinite loop
+ warnings unless the loop comes from source, because checking generated
+ loops is a waste of time, and makes it harder to debug
+ Check_Infinite_Loop_Warning.
+
+ * sem_warn.adb (Check_Infinite_Loop_Warning): If the local variable
+ tested in the while loop is a renaming, do not warn. Otherwise, we get
+ false alarms, because it's usually renaming something that we can't
+ deal with (an indexed component, a global variable, ...).
+
+ * gnat_rm.texi: Fix typo
+
+2009-04-15 Thomas Quinot <quinot@adacore.com>
+
+ * sem_ch6.adb: Minor reformatting
+
2009-04-15 Hristian Kirtchev <kirtchev@adacore.com>
* exp_ch7.adb (Expand_Ctrl_Function_Call): Check for the case where the
The fourth form, with an On|Off parameter and a string, is used to
control individual messages, based on their text. The string argument
is a pattern that is used to match against the text of individual
-warning messages (not including the initial "warnings: " tag).
+warning messages (not including the initial "warning: " tag).
-The pattern may contain asterisks which match zero or more characters in
+The pattern may contain asterisks, which match zero or more characters in
the message. For example, you can use
@code{pragma Warnings (Off, "*bits of*unused")} to suppress the warning
message @code{warning: 960 bits of "a" unused}. No other regular
Process_End_Label (Loop_Statement, 'e', Ent);
End_Scope;
Kill_Current_Values;
- Check_Infinite_Loop_Warning (N);
+
+ -- No point in checking for warnings in code we generated
+
+ if Comes_From_Source (N) then
+ Check_Infinite_Loop_Warning (N);
+ end if;
-- Code after loop is unreachable if the loop has no WHILE or FOR
-- and contains no EXIT statements within the body of the loop.
procedure Possible_Freeze (T : Entity_Id);
-- T is the type of either a formal parameter or of the return type.
-- If T is not yet frozen and needs a delayed freeze, then the
- -- subprogram itself must be delayed. If T is the limited view of
- -- of an incomplete type the subprogram must be frozen as well,
- -- because T may depend on local types that have not been frozen yet.
+ -- subprogram itself must be delayed. If T is the limited view of an
+ -- incomplete type the subprogram must be frozen as well, because
+ -- T may depend on local types that have not been frozen yet.
---------------------
-- Possible_Freeze --
procedure Possible_Freeze (T : Entity_Id) is
begin
- if Has_Delayed_Freeze (T)
- and then not Is_Frozen (T)
- then
+ if Has_Delayed_Freeze (T) and then not Is_Frozen (T) then
Set_Has_Delayed_Freeze (Designator);
elsif Is_Access_Type (T)
then
Set_Has_Delayed_Freeze (Designator);
- elsif Ekind (T) = E_Incomplete_Type
- and then From_With_Type (T)
- then
+ elsif Ekind (T) = E_Incomplete_Type and then From_With_Type (T) then
Set_Has_Delayed_Freeze (Designator);
end if;
+
end Possible_Freeze;
-- Start of processing for Check_Delayed_Subprogram
-- Start of processing for Check_Infinite_Loop_Warning
begin
- -- We need a while iteration with no condition actions. Conditions
+ -- We need a while iteration with no condition actions. Condition
-- actions just make things too complicated to get the warning right.
if No (Iter)
Find_Var (Condition (Iter));
- -- Nothing to do if local variable from source not found
+ -- Nothing to do if local variable from source not found. If it's a
+ -- renaming, it is probably renaming something too complicated to deal
+ -- with here.
if No (Var)
or else Ekind (Var) /= E_Variable
or else Is_Library_Level_Entity (Var)
or else not Comes_From_Source (Var)
+ or else Nkind (Parent (Var)) = N_Object_Renaming_Declaration
then
return;
projects / thirdparty / gcc.git / commitdiff
