What is a purpose in Pascal to declare variable that is pointer to a pointer? I have a code in Mac Think Pascal. Here is some parts from the code that I don't understand:
type
my_array = array[1..100] of integer;
my_array_pointer = ^my_array;
my_array_handle = ^my_array_pointer;
...
var
xx : my_array_handle;
...
begin
xx:= my_array_handle(NewHandle( sizeof(my_array)) );
As you see, the last line is an assignment of a type my_array_handle to variable xx. What does it mean? What does NewHandle function do? (This is an internal function of Think Pascal). Actually, I need to convert a Think Pascal program to Windows Pascal. However I cannot find the description of NewHandle function, and don't know how to implement this function using the standard (New(), GetMem() etc) pointer functions.
This is a classic macos feature, not typically something of Pascal.
I don't know exactly, but it had something to do with relocatability of the loaded program in a non PM environment.
Note that the indirect pointer is allocated via an OS function, which probably means that it is allocated in some table that is maintained by the OS. (so that the OS can move/relocate the program?)
In modern Mac (and other) programming this whole principle is alien. Just clean it up.
(added)
If you want to keep these redirections, you could try your luck with something like:
function newhandle( nrbytes:integer):ppointer;
var xx : ppointer;
begin
new(xx);
getmem(xx^,nrbytes);
newhandle:=xx;
end;
I didn't add this originally, but I recommend you simply clean up these anachronistic indirect references and use my_array_pointer based pointers and getmem or new directly.
The meaning of these indirect references has no use on non m68k Classic Mac OS systems, though afaik later PPC versions still somewhat support them. (PPC is always protected mode)
If you really want the gritty details, you probably want to subscribe to the mac-pascal list.
Related
I have a script to work out how much free stack space there is in each FreeRTOS task. GDB’s language is set to auto. The script works fine when the current language is c, but fails when the current language is ada.
I have, in the class Stacks,
tcb_t = gdb.lookup_type("TCB_t")
int_t = gdb.lookup_type("int")
used to:
find {Ada task control block}.Common.Thread,
thread = atcb["common"]["thread"]
convert to a pointer to the FreeRTOS task control block,
tcb = thread.cast(Stacks.tcb_t.pointer()).dereference()
find the logical top of the stack
stk = tcb["pxStack"].cast(Stacks.int_t.pointer())
Now I need to loop logically down the stack until I find an entry not equal to the initialised value,
free = 0
while stk[free] == 0xa5a5a5a5:
free = free + 1
which works fine if the current frame’s language is c, but if it’s ada I get
Python Exception <class 'gdb.error'> not an array or string:
Error occurred in Python command: not an array or string
I’ve traced this to the expression stk[free], which is being interpreted using the rules of the current language (in Ada, array indexing uses parentheses, so it would be stk(free), which is of course illegal since Python treats it as a function call).
I’ve worked round this by
def invoke(self, arg, from_tty):
gdb.execute("set language c")
...
gdb.execute("set language auto")
but it seems wrong not to set the language back to what it was originally.
So,
is there a way of detecting the current GDB language setting from Python?
is there an alternate way of indexing that doesn’t depend on the current GDB language setting?
I am using custom LLVM pass where if I encounter a store to
where the compiler converts the value to a Constant; e.g. there is an explicit store:
X[gidx] = 10;
Then LLVM will generate this error:
aoc: ../../../Instructions.cpp:1056: void llvm::StoreInst::AssertOK(): Assertion `getOperand(0)->getType() == cast<PointerType>(getOperand(1)->getType())->getElementType() && "Ptr must be a pointer to Val type!"' failed.
The inheritance order goes as: Value<-User<-Constant, so this shouldn't be an issue, but it is. Using an a cast on the ConstantInt or ConstantFP has no effect on this error.
So I've tried this bloated solution:
Value *new_value;
if(isa<ConstantInt>(old_value) || isa<ConstantFP>(old_value)){
Instruction *allocInst = builder.CreateAlloca(old_value->getType());
builder.CreateStore(old_value, allocInst);
new_value = builder.CreateLoad(allocResultInst);
}
However this solution creates its own register errors when different type are involved, so I'd like to avoid it.
Does anyone know how to convert a Constant to a Value? It must be a simple issue that I'm not seeing. I'm developing on Ubuntu 12.04, LLVM 3, AMD gpu, OpenCL kernels.
Thanks ahead of time.
EDIT:
The original code that produces the first error listed is simply:
builder.CreateStore(old_value, store_addr);
EDIT2:
This old_value is declared as
Value *old_value = current_instruction->getOperand(0);
So I'm grabbing the value to be stored, in this case "10" from the first code line.
You didn't provide the code that caused this first assertion, but its wording is pretty clear: you are trying to create a store where the value operand and the pointer operand do not agree on their types. It would be useful for the question if you'd provide the code that generated that error.
Your second, so-called "bloated" solution, is the correct way to store old_value into the stack and then load it again. You write:
However this solution creates its own register errors when different type are involved
These "register errors" are the real issue you should be addressing.
In any case, the whole premise of "converting a constant to a value" is flawed - as you have correctly observed, all constants are values. There's no point storing a value into the stack with the sole purpose of loading it again, and indeed the standard LLVM pass "mem2reg" will completely remove such a sequence, replacing all uses of the load with the original value.
I have just started using Ada, and I'm finding the generic package declarations to be rather silly. Maybe I'm not doing it right, so I'm looking for better options.
Take a look at the below example.
package STD_Code_Maps is new
Ada.Containers.Map(Key_Type => STD_Code_Type;
Element_Type => Ada.Strings.Unbounded.Unbounded_String);
STD_Code_Map : STD_Code_Maps.Map;
-- . . .
procedure Do_Something is
Item : Ada.Strings.Unbounded.Unbounded_String;
begin
Item := STD_Code_Maps.Element(STD_Code_Maps.First(STD_Code_Map));
-- Do something with the Item
-- . . .
end Do_Something;
It would be much cleaner to simply be able to write STD_Code_Map.First.Element instead of the godforsaken STD_Code_Maps.Element(STD_Code_Maps.First(STD_Code_Map));
Obviously I'm doing this wrong -- I think. I'm repeating the phrase STD_Code_Map at least thrice over there. I'm all for verbosity and everything, but really the code I'm writing seems bad and silly to me.
I was wondering if there is solution that doesn't require you to rename the package to something like package Map renames STD_Code_Maps; which would of course shorten the code but I don't want to do this on each and every procedure entry. I really think something like STD_Code_Map.First.Element would be much simpler. Can this be done in Ada 2012?
Note: Using the Unbounded_String package by default is also so difficult. Did the standard library designers actually give much thought to the ridiculous and overly long package hierarchy?
Thanks for reading this, and potentially helping me out. I'm new to Ada.
GNAT GPL 2012 and 2013, and FSF GCC 4.7 and 4.8, support the new container indexing scheme of Ada 2012, which means that you can write
Item := STD_Code_Map ({some Cursor});
And you can do this even with the -gnat05 switch to force Ada 2005 mode! (which has to be a bug).
Ada 2005 allows you to call a primitive function of a tagged type using object.function notation, provided that the first operand is of the tagged type; so you can write STD_Code_Map.First as shorthand for STD_Code_Maps.First (STD_Code_Map).
Putting these together, you can write
Item := STD_Code_Map (STD_Code_Map.First);
which is quite short!
The issue has nothing to do with generics. As Simon pointed out, you can say STD_Code_Map.First, since the type of STD_Code_Map is a tagged type and Ada supports this notation for tagged types. On the other hand, the type of STD_Code_Map.First is a Cursor type, which isn't tagged (making it tagged would have caused problems declaring certain operations that take both a Cursor and a Map). But even without the Ada 2012 container indexing that Simon mentioned, you can say
STD_Code_Maps.Element(STD_Code_Map.First);
which is a little better. Besides renaming a package, you can also rename the function:
function Elem (Position : STD_Code_Maps.Cursor) return STD_Code_Maps.Element_Type
renames STD_Code_Maps.Element;
and you can now use just Elem instead of STD_Code_Maps.Element wherever the renaming is directly visible. (You could call it Element if you want to. The renaming name can be the same or it can be different.) This could be helpful if you use that function a lot.
Ada was designed for readability and maintainability. (written once, read & maintained for much longer) This means that it does get a little verbose at times. If you prefer terse & cryptic there are plenty of other languages out there !
If you want to avoid typing STD_Code_Map all the time, just use a use clause:
use STD_Code_Map;
which would mean your code of
Item := STD_Code_Maps.Element(STD_Code_Maps.First(STD_Code_Map));
would become
Item := Element(First(STD_Code_Map));
Getting names nice and readable in Ada can sometimes be tricky. Often times language designers made the task worse than it had to be, by designing Ada standard library packages for use with Ada use clauses, without a thought to how they'd look to some poor sap who either can't or doesn't want to use that feature.
In this case though, there are things you can do on your own end.
For example "_Maps.Map" is redundant, so why not get rid of it from the package name? Why not use names so that you can write:
package STD_Code is new
Ada.Containers.Map(Key_Type => STD_Code_Type;
Element_Type => Ada.Strings.Unbounded.Unbounded_String);
Map : STD_Code.Map;
-- . . .
procedure Do_Something is
Item : Ada.Strings.Unbounded.Unbounded_String;
begin
Item := STD_Code.Element(STD_Code.First(Map));
-- Do something with the Item
-- . . .
end Do_Something;
Now Code looks like a bit of a null-meaning word too. Everything in a program is a code. So I'd consider ditching it as well. Unsually I name my Ada container packages something that says their basic theoretical function (eg: STD_to_String), while objects are more specific nouns.
Also, I should point out that if your map is constant and you can live with names that look like identifiers, often times you can get rid of maps to strings entirely by using an enumerated type and the 'image attribute.
If you're using Ada 2012 the STD_Code_Maps.Element(STD_Code_Maps.First(STD_Code_Map)); could become:
Function Get_First( Map : STD_Code_Maps.Map ) return Unbounded_String is
( STD_Code_Maps.Element(Map.First) );
That is an example of the new Expression-functions, which were mainly introduced for pre- and post-conditions.
Of course, if you're using Ada 2012 then the precise map you'd likely want is Ada.Containers.Indefinite_Ordered_Maps -- the Indefinite_* containers are those that [can] have indefinite elements, like String.
In addition to TED's comment, using an enumeration when the maps are constant have even more advantages, namely case coverage: the compiler will flag as an error any case-statement that doesn't cover all alternatives. So if you add a new code it will flag all the case-statements that you need to modify. (Of course this benefit is lost when you use the others case.)
folks!
I pass a struct full of data to my kernel, and I run into the following difficulty using it (very stripped down):
[edit: mac osx / xcode 3.2 on mac book pro; this compile is obviously for cpu]
typedef struct
{
float xoom;
int sizex;
} varholder;
float zX, xd;
__kernel void Harlan( __global varholder * vh )
{
int X = get_global_id(0), Y = get_global_id(1);
zX = ( ( X - vh->sizex/2 ) / vh->xoom + vh->sizex/2 ); // (a)
xd = zX; // (b) BOOM!!
}
after executing line (a), the line marked (b), a simple assignment, gives "LLVM compiler failed to compile a function".
if, however, we do not execute line (a), then line (b) is fine.
So, through my fiddling around a LOT with this, it seems as if it is the assignment statement (a), which uses a passed-in parameter, that messes up the future access of the variable zX. However, of course I need to be able to use the results of calculations further down the line.
I have zX and xd declared at the file level because my helper functions need them.
Any thoughts?
Thanks!
David
p.s. I'm now registered so will be able to upvote and accept answers, which I am sadly unable to do for the last person who helped me (used same username to register, but can't seem to vote on the old post; sorry!).
No, say it ain't so!
I am sincerely hoping that this is not a "correct" answer to my own question. I found on another forum (though not the same question asked!) the following, and I am afraid that it refers to what I'm trying to do:
(quote)
You're doing something the standard prohibits. Section 6.5 says:
'All program scope variables must be declared in the __constant address space.'
In other words, program scope variables cannot be mutable.
(end quote)
... well, tcha!!!! What an astoundingly inconvenient restriction! I'm sure there's reasoning behind it.
[edit: Not At All inconvenient! it was in fact astonishingly easy to work around, given a fresh start the next morning. (And no alcohol.)]
You guys & dolls all knew this, right, and didn't have the heart to tell me?...
I would like to check if a pointer inside a derived type has already been defined or not. I wrote the following simple code to show you my problem:
program test
implicit none
type y
real(8), pointer :: x(:)
end type y
type(y), pointer :: w(:)
allocate(w(2))
allocate(w(1)%x(2))
write(*,*) associated(w(1)%x), associated(w(2)%x)
end program test
Compiling this code with gFortran 4.4.1 and running it on Ubuntu gives the result:
T F
whereas the same code compiled on Windows Vista with the Intel Fortran compiler 11.0 provides:
T T
The first result (gFortran) is what I am actually expecting. But the fact that the Intel compiler provides a different result makes me fear my code might not be correct. Am I doing something terribly wrong with the pointers in this example? Any idea or explanation?
Many thanks in advance for your help!
You are testing to see if a pointer is associated without explicitly using nullify on the pointers. A great page on common Fortran mistakes remarks (with the code sample removed):
Many people think that the status of a pointer which has never been associated is .not. associated. This is false. (...) When a pointer is declared its status is undefined, and cannot be safely queried with the associated intrinsic.
It looks like the gfortran compiler may be set up to explicitly nullify pointers on declaration - you should probably think of this like the compiler automatically setting declared variables to zero, and not count on that behavior. If you want to be sure, you will nullify it yourself.
Edit:
I'm reading through the Intel compiler guide, and it specifies how to make sure that the pointer is nullified correctly - you can set up your derived type as
type y
real(8), pointer :: x(:) => null()
end type y
Note, however, that it seems like this is limited to Fortran 95, as mentioned in the linked article.