In order to save different types(str or double) of values in a non-typed var in my language,I need a struct like:
Struct BaseVarStruct{
bool _isString;
union {
string _sVal;
double _dVal;
}_val;
};
I try to use StructType* BaseVarType=StructType::create("basevar",type::getInt1Ty(),???),but i dont't know for sure it's a literal or not.In offical class ref,There's a class called"UnionType",but i can't find it .MAYBE union type is not supported in LLVM now？or maybe i can define union type by StructType and give all members a same offset?
The LLVM IR used to have a union type - which was essentially a struct with all fields at offset 0 - but it no longer does (removed in version 2.8). I believe that compiling any language with dynamic types or unions into the strongly-typed IR will require adding bitcasts.
So you could find the requires alignment and create a struct with the fields to fit it (see this related question on how Clang does it), then branch based on _isString, and in each branch target perform a different bitcast. Alternatively, if you know the type expected of the union at each point it is used, you could encode that information in metadata instead of a field, and then compile the the access to the union based on the metadata.
Related
I want to deal with a structure like this struct foo {char *name; char **fields ; size_t nfields};
If I define corresponding structure in Squeak
ExternalStructure subclass: #Foo
instanceVariableNames: ''
classVariableNames: ''
poolDictionaries: ''
category: 'FFI-Tests'.
and define the fields naively with
Foo class>fields
^#(
(name 'char*')
(fields 'char**')
(nfields 'unsigned long')
)
then generate the accessors with Foo defineFields, I get those undifferentiated types for name and fields:
Foo>>name
^ExternalData fromHandle: (handle pointerAt: 1) type: ExternalType char asPointerType
Foo>>fields
^ExternalData fromHandle: (handle pointerAt: 5) type: ExternalType char asPointerType
That is troubling, the second indirection is missing for the fields accessor.
How should I specify fields accessor in the spec?
If not possible, how do I define it manually?
And I have the same problem for this HDF5 function prototype: int H5Tget_array_dims(hid_t tid, hsize_t *dims[])
The following syntax is not accepted:
H5Tget_array_dims: tid with: dims
<cdecl: long 'H5Tget_array_dims'(Hid_t Hsize_t * * )>
The compiler barks argument expected -> before the second *...
I add to resort to void * instead, that is totally bypassing typechecking - less than ideal...
Any idea how to deal correctly with such prototype?
Since Compiler-mt.435, the parser will not complain anymore but call back to ExternalType>>asPointerToPointerType. See source.squeak.org/trunk/Compiler-mt.435.diff and source.squeak.org/FFI/FFI-Kernel-mt.96.diff
At the time of writing this, such pointer-to-pointer type will be treated as regular pointer type. So, you loose the information that the external type actually points to an array of pointers.
When would one need that information?
When coercing arguments in the FFI plugin during the call
When constructing the returned object in the FFI plugin during the call
When interpreting instances of ExternalData from struct fields and FFI call return values
In tools such as the object explorer
There already several kinds of RawBitsArray in Squeak. Adding String and ExternalStructure (incl. packed or union) to the mix, we have all kinds of objects in Squeak to map the inner-most dimension (i.e., int*, char*, void*). ExternalData can represent the other levels of the multi-dimensional array (i.e., int**, char**, void** and so on).
So, there are remaining tasks here:
Store that pointer dimension information maybe in a new external type to be found via ExternalType>>referencedType. We may want to put new information into compiledSpec. See http://forum.world.st/FFI-Plugin-Question-about-multi-dimensional-arrays-e-g-char-int-void-td5118484.html
Update value reading in ExternalArray to unwrap one pointer after the other; and let the code generator for struct-field accessors generate code in a similar fashion.
Extend argument coercing in the plugin to accept arrays of the already supported arrays (i.e. String etc.)
I am currently sending data between my PC and an ARM M4 Microcontroller via UART. I've defined my own protocol where each message looks like this:
[START_CHAR LEN TYPE SEQ DATA CRC]
The START_CHAR and LEN fields help me determine when the data ends, after which I look up the TYPE (constant offset of 3) to figure out what data came in order to unpack it into a message class.
Now I'm looking into flatbuffers and it seems perfect except that I cannot encode the TYPE into the message without including it inside the actual message. Here is what I am trying to do:
namespace FlatMessage;
uint8 const TYPE = 50; // does not compile
table String {
value:string;
}
root_type String;
I could create an Enum but that is messy. Thank you!
[EDIT] I should add that I could just change the protocol to have an END_CHAR but I need to support the TYPE field for legacy reasons.
Well actually, I suppose I would still need the type to figure out how to deserialize it as a flatbuffer.
e.g.
uint8_t *buf = builder.GetBufferPointer(); // I can do this with END_CHAR because I could get the buffer.
auto receive_string = GetString(buf); // But I wouldn't know what the type is. e.g. this could be GetCoolString(buf).
You have a couple of options to store a type with a FlatBuffer:
Prefix a buffer yourself with a type.
Use the file_identifier feature of FlatBuffers, to make it possible to identify the type of FlatBuffer.
Store the type in FlatBuffers itself, by using a union type. Make the root table have a single union field.
I have two functions: one returns a list of fields, the other returns a select query (which selects the corresponding values of the fields).
private List<Field<?>> fields() {
....
}
private Select<?> select() {
...
}
Note that the degree is determined at runtime, it depends on the user input. Hence List<Field<?>> and Select<?>.
It is possible to insert into a table:
context.insertInto(table, fields()).select(select()))
It is not possible to update a table:
context.update(table).set(DSL.row(fields()), select())
Could this functionality be added to jOOQ 3.7?
Which workaround can we use for now?
Nice catch, there's a missing method on the UpdateSetFirstStep DSL API, which accepts RowN as a first argument, the type returned from DSL.row(Collection). This should be fixed for jOOQ 3.7:
https://github.com/jOOQ/jOOQ/issues/4475
As a workaround, and if you can live with the guilt of the hack, you could cast to raw types:
context.update(table).set((Row1) DSL.row(fields()), (Select) select())
You can cast DSL.row(fields()) to Row1, because the internal implementation type returned by DSL.row(fields()) implements all Row[N] types.
In Fortran, is there a way to determine the type of a variable?
A possible use case where the type of a variable would be needed is the following. We pass a variable's type as an argument to a function, to be able to call type-specific code with that function, thus eliminating the need to have separate similar functions for each data type.
Well you might be able to do what you want if you mess about with the KIND intrinsic and POINTERs, but if you are only concerned with the signature of functions and subroutines, leave it to Fortran. If you define
function calc8(arg1)
real(8), intent(in) :: arg1
...
and
function calc4(arg1)
real(4), intent(in) :: arg1
...
in a module, and declare an interface like this
interface calc
module procedure calc8
module procedure calc4
end interface
(Warning, I haven't checked the syntax in detail, that's your responsibility.)
then Fortran will match the call to the right version of the function. Sure, you have to write both versions of the function, but that's really the Fortran 95 way of doing it. This can be quite tedious, I tend to write a generic version and run a sed script to specialise it. It's a bit of a kludge but it works.
If the code of the function is identical apart from the kind of the arguments I sometimes write the function for real(8) (or whatever) and write a version for real(4) which calls the real(8) version wrapped in type conversions.
In Fortran 2003 there are improved ways of defining polymorphic and generic functions, but I haven't really got my head around them yet.
Yes, there are two ways.
The first way does requires you to write separate functions or subroutines for each variable type, but you don't have to call different functions. This may or may not be close enough to what you want. You write the separate routines, then to write an interface to create a generic function or subroutine wrapping these specific subroutines. You don't have to pass the variable type, or do anything special in your call -- it is all done via the declaration and automatically by the compiler from the variable itself, as long as the variables are different enough that the compiler can distinguish them (there are rules about what is required). This is similar to how intrinsic functions work -- you can call sin with a real argument, a double precision real argument or a complex argument and the compiler calls the correct actual function and returns the matching result. High Performance Mark sketched a solution along these lines. For another question, I posted a working example, where the distinguishing feature of the variables was array rank: how to write wrapper for 'allocate'. An advantage of this method is that it is widely support by Fortran compilers.
In Fortran 2003/2008 there are extensive object oriented features. Quoting "Fortran 95/2003 explained" by Metcalf, Reid and Cohen, "To execute alternative code depending on the dynamic type of a polymorphic entity and to gain access to the dynamic parts, the select type construct is provided." The select type statement is a bit similar to a select case statement. This is support by fewer compilers. Again, you don't have to pass the type, since the compiler can figure it you from the variable itself. But it has to be a polymorphic type... Both Intel ifort and gfortran list select type and polymorphic datatypes as supported -- the later with some experimental aspects in gfortran (http://gcc.gnu.org/wiki/Fortran2003). These are recent additions to these compilers.
Here is a piece of code that determines what the type of something is. The action is trivial but you should be able to extend it to your use case.
module element_to_datatype
use iso_fortran_env
use mpi_f08
implicit none
contains
function get_element_datatype(element) result(datatype)
class(*), intent(in) :: element
type(MPI_Datatype) :: datatype
select type(element)
! REAL types
type is ( real(kind=REAL32) )
datatype = MPI_REAL4
type is ( real(kind=REAL64) )
datatype = MPI_REAL8
! COMPLEX types
type is ( complex(kind=REAL32) )
datatype = MPI_COMPLEX8
type is ( complex(kind=REAL64) )
datatype = MPI_COMPLEX16
! INTEGER types
type is ( integer(kind=INT8) )
datatype = MPI_INTEGER1
type is ( integer(kind=INT16) )
datatype = MPI_INTEGER2
type is ( integer(kind=INT32) )
datatype = MPI_INTEGER4
type is ( integer(kind=INT64) )
datatype = MPI_INTEGER8
! OTHER types
type is ( logical )
datatype = MPI_LOGICAL
end select
end function
end module element_to_datatype
A previous answer shows how to do this with interfaces, so I won't repeat that.
I have a C structure that allow users to configure options in an embedded system. Currently the GUI we use for this is custom written for every different version of this configuration structure. What I'd like for is to be able to describe the structure members in some format that can be read by the client configuration application, making it universal across all of our systems.
I've experimented with describing the structure in XML and having the client read the file; this works in most cases except those where some of the fields have inter-dependencies. So the format that I use needs to have a way to specify these; for instance, member A must always be less than or equal to half of member B.
Thanks in advance for your thoughts and suggestions.
EDIT:
After reading the first reply I realized that my question is indeed a little too vague, so here's another attempt:
The embedded system needs to have access to the data as a C struct, running any other language on the processor is not an option. Basically, all I need is a way to define metadata with the structure, this metadata will be downloaded onto flash along with firmware. The client configuration utility will then read the metadata file over RS-232, CAN etc. and populate a window (a tree-view) that the user can then use to edit options.
The XML file that I mentioned tinkering with was doing exactly that, it contained the structure member name, data type, number of elements etc. The location of the member within the XML file implicitly defined its position in the C struct. This file resides on flash and is read by the configuration program; the only thing lacking is a way to define dependencies between structure fields.
The code is generated automatically using MATLAB / Simulink so I do have access to a scripting language to help with the structure creation. For example, if I do end up using XML the structure will only be defined in the XML format and I'll use a script to create the C structure during code generation.
Hope this is clearer.
For the simple case where there is either no relationship or a relationship with a single other field, you could add two fields to the structure: the "other" field number and a pointer to a function that compares the two. Then you'd need to create functions that compared two values and return true or false depending upon whether or not the relationship is met. Well, guess you'd need to create two functions that tested the relationship and the inverse of the relationship (i.e. if field 1 needs to be greater than field 2, then field 2 needs to be less than or equal to field 1). If you need to place more than one restriction on the range, you can store a pointer to a list of function/field pairs.
An alternative is to create a validation function for every field and call it when the field is changed. Obviously this function could be as complex as you wanted but might require more hand coding.
In theory you could generate the validation functions for either of the above techniques from the XML description that you described.
I would have expected you to get some answers by now, but let me see what I can do.
Your question is a bit vague, but it sounds like you want one of
Code generation
An embedded extension language
A hand coded run-time mini language
Code Generation
You say that you are currently hand tooling the configuration code each time you change this. I'm willing to bet that this is a highly repetitive task, so there is no reason that you can't write program to do it for you. Your generator should consume some domain specific language and emit c code and header files which you subsequently build into you application. An example of what I'm talking about here would be GNU gengetopt. There is nothing wrong with the idea of using xml for the input language.
Advantages:
the resulting code can be both fast and compact
there is no need for an interpreter running on the target platform
Disadvantages:
you have to write the generator
changing things requires a recompile
Extension Language
Tcl, python and other languages work well in conjunction with c code, and will allow you to specify the configuration behavior in a dynamic language rather than mucking around with c typing and strings and and and...
Advantages:
dynamic language probably means the configuration code is simpler
change configuration options without recompiling
Disadvantages:
you need the dynamic language running on the target platform
Mini language
You could write your own embedded mini-language.
Advantages:
No need to recompile
Because you write it it will run on your target
Disadvantages:
You have to write it yourself
How much does the struct change from version to version? When I did this kind of thing I hardcoded it into the PC app, which then worked out what the packet meant from the firmware version - but the only changes were usually an extra field added onto the end every couple of months.
I suppose I would use something like the following if I wanted to go down the metadata route.
typedef struct
{
unsigned char field1;
unsigned short field2;
unsigned char a_string[4];
} data;
typedef struct
{
unsigned char name[16];
unsigned char type;
unsigned char min;
unsigned char max;
} field_info;
field_info fields[3];
void init_meta(void)
{
strcpy(fields[0].name, "field1");
fields[0].type = TYPE_UCHAR;
fields[0].min = 1;
fields[0].max = 250;
strcpy(fields[1].name, "field2");
fields[1].type = TYPE_USHORT;
fields[1].min = 0;
fields[1].max = 0xffff;
strcpy(fields[2].name, "a_string");
fields[2].type = TYPE_STRING;
fields[2].min = 0 // n/a
fields[2].max = 0 // n/a
}
void send_meta(void)
{
rs232_packet packet;
memcpy(packet.payload, fields, sizeof(fields));
packet.length = sizeof(fields);
send_packet(packet);
}