www.digitalmars.com         C & C++   DMDScript  

digitalmars.D.learn - D string to C struct fixed-size array

reply bdh <ben.h posteo.net> writes:
Hi,

I'm trying to create bindings to the GraphcicsMagick C library 
which has the following struct defined:

     #define MaxTextExtent 2053

     typedef struct _Image {
       /* other members skipped */
       char filename[MaxTextExtent];
     } Image;


In an `extern (C)` block I've "converted" the struct to D:

     enum MaxTextExtent = 2053;

     struct Image
     {
       /* other members skipped */
       char[MaxTextExtent] filename;
     }


The problem is trying to set the `filename` member since I can't 
directly use a D string.  I've seen a post which is a similar 
issue[0], but adapting the code provides an error:
     static assert: "Cannot put a string into a char[2053]"

I've also tried using `new char[MaxTextExtent]` and looping 
through the string to copy the characters over.  This does 
compile, however, when trying to write the image to disk the 
provided filename is often a "stripped" version of when you want 
(e.g. "image.jpg" becomes "g").

[0]: 
https://forum.dlang.org/thread/lshnzqbfkrhfkliapicm forum.dlang.org
Jan 03 2021
next sibling parent reply rikki cattermole <rikki cattermole.co.nz> writes:
import std;
void main()
{
     int[] a = [1, 2, 3, 4, 5];
     int[3] b;

     b[0 .. 3] = a[1 .. 4];
     b.writeln;
}

Same principle, just remember to null terminate after slicing your 
dynamic array and assigning it to your static array.
Jan 03 2021
parent reply bdh <ben.h posteo.net> writes:
On Sunday, 3 January 2021 at 09:28:55 UTC, rikki cattermole wrote:

 import std;
 void main()
 {
     int[] a = [1, 2, 3, 4, 5];
     int[3] b;

     b[0 .. 3] = a[1 .. 4];
     b.writeln;
 }

 Same principle, just remember to null terminate after slicing 
 your dynamic array and assigning it to your static array.


Thanks for the suggestion, however, it yields the same results as 
my "create new array and loop through string" attempt.  That is 
to say, an incomplete filename is used when writing (saving) the 
image to disk.

If it help, here is some of the actual code (using what I've 
understood from your sample):

     int main(string[] args)
     {
       ulong len;

       Image* image;
       /* a separate structure, which also has a 
char[MaxTextExtent] filename */
       ImageInfo* imageInfo;
       /* another GraphicsMagick struct */
       ExceptionInfo exception;

       InitializeMaigck(null);
       /* creates a ImageInfo with defaults */
       imageInfo = CloneImageInfo(null);

       /* actual code has a check for args.length */
       len = args[1].length;

       /* set the input filename */
       imageInfo.filename[0 .. len] = args[1];
       imageInfo.filename[len] = '\0';

       /* read image from input file */
       image = ReadImage(imageInfo, &exception);

       /* set the output filename */
       len = args[2].length;
       image.filename[0 .. len] = args[2];
       image.filename[len] = '\0';

       /* write (save) image to disk */
       WriteImage(imageInfo, image);

       /* memory cleanup skipped */

       return 0;
     }

The full code is really just a port of the `convert.c` example 
shown on the API page[0],

[0]: http://www.graphicsmagick.org/api/api.html
Jan 03 2021
parent reply rikki cattermole <rikki cattermole.co.nz> writes:
Your definition of Image is probably wrong.

You may have missed a pointer (8 bytes).
Jan 03 2021
parent reply bdh <ben.h posteo.net> writes:
On Sunday, 3 January 2021 at 11:16:25 UTC, rikki cattermole wrote:

 Your definition of Image is probably wrong.

 You may have missed a pointer (8 bytes).


I'm pretty sure it's correct?

Here's the D version:
   
https://repo.or.cz/magickd.git/blob/e5d06e939:/source/magickd/core/c/image.d#l751

Here's the C version:
   
http://hg.code.sf.net/p/graphicsmagick/code/file/a622095da492/magick/image.h#l683
Jan 03 2021
next sibling parent reply Steven Schveighoffer <schveiguy gmail.com> writes:
On 1/3/21 9:17 PM, bdh wrote:

 On Sunday, 3 January 2021 at 11:16:25 UTC, rikki cattermole wrote:

 Your definition of Image is probably wrong.

 You may have missed a pointer (8 bytes).


 
 I'm pretty sure it's correct?
 
 Here's the D version:
 https://repo.or.cz/magickd.git/blob/e5d06e939:/source/magickd
core/c/image.d#l751 
 
 
 Here's the C version:
 http://hg.code.sf.net/p/graphicsmagick/code/file/a622095da492
magick/image.h#l683 
 
 


How to test:

in D-land:

import magickd.core.c.image : Image;

pragma(msg, Image.filename.offsetof);

In C-land

#include <stdio.h>
#include <magick/image.h>

int main()
{
    Image *ptr = NULL;
    printf("%ld", &(ptr->filename)); // may need a different specifier
}

If those are not the same, then your struct is off.

-Steve
Jan 04 2021
parent bdh <ben.h posteo.net> writes:
On Monday, 4 January 2021 at 17:24:27 UTC, Steven Schveighoffer 
wrote:

 On 1/3/21 9:17 PM, bdh wrote:

 On Sunday, 3 January 2021 at 11:16:25 UTC, rikki cattermole 
 wrote:

 Your definition of Image is probably wrong.

 You may have missed a pointer (8 bytes).


 
 [...]
 


 How to test:

 in D-land:

 import magickd.core.c.image : Image;

 pragma(msg, Image.filename.offsetof);

 In C-land

 #include <stdio.h>
 #include <magick/image.h>

 int main()
 {
    Image *ptr = NULL;
    printf("%ld", &(ptr->filename)); // may need a different 
 specifier
 }

 If those are not the same, then your struct is off.

 -Steve


Thanks for this, I wasn't aware of the .offsetof property (I see 
it's mentioned in the Struct and Unions section of the language 
reference).

As rikki guessed, I had missed something (offset was 480 instead 
of 488).  The issue wasn't with the Image struct itself, but with 
the TimerInfo struct (used for the `timer` member of Image) 
missing a "private" unsigned long.

Thanks again.
Jan 04 2021
prev sibling parent reply tsbockman <thomas.bockman gmail.com> writes:
On Monday, 4 January 2021 at 02:17:33 UTC, bdh wrote:

 I'm pretty sure it's correct?

 Here's the D version:
   
 https://repo.or.cz/magickd.git/blob/e5d06e939:/source/magickd/core/c/image.d#l751

 Here's the C version:
   
 http://hg.code.sf.net/p/graphicsmagick/code/file/a622095da492/magick/image.h#l683




 /* XXX: haven't done private members yet, wonder if they're 
 needed to
  * complete this */


Leaving out the private members of a struct changes its size, and 
sometimes its alignment. This will definitely break things if you 
ever allocate an instance yourself (including via the stack).

If you only EVER access that type via pointers to instances 
allocated by an API which internally uses the complete, correct 
definition, and ALL of the private members are at the end of the 
struct, it should work - but those are two big "if"s, and too 
easy to unthinkingly violate later by accident. So, it is best to 
ensure that the struct definition matches precisely on both the D 
and C sides of the API, or at least enough to ensure that the 
size and alignment are the same.
Jan 04 2021
parent bdh <ben.h posteo.net> writes:
On Monday, 4 January 2021 at 18:05:40 UTC, tsbockman wrote:

 /* XXX: haven't done private members yet, wonder if they're 
 needed to
  * complete this */


 Leaving out the private members of a struct changes its size, 
 and sometimes its alignment. This will definitely break things 
 if you ever allocate an instance yourself (including via the 
 stack).

 If you only EVER access that type via pointers to instances 
 allocated by an API which internally uses the complete, correct 
 definition, and ALL of the private members are at the end of 
 the struct, it should work - but those are two big "if"s, and 
 too easy to unthinkingly violate later by accident. So, it is 
 best to ensure that the struct definition matches precisely on 
 both the D and C sides of the API, or at least enough to ensure 
 that the size and alignment are the same.


While the struct that comment is from (ImageInfo) had been fixed 
before your message[0], I do appreciate the information here 
since I wasn't aware of it before and it helped solve the issue 
(see my reply to Steve). Thanks.

[0]: https://repo.or.cz/magickd.git/commit/27ac82b110
Jan 04 2021
prev sibling parent reply Jack <jckj33 gmail.com> writes:
On Sunday, 3 January 2021 at 08:43:34 UTC, bdh wrote:

 Hi,

 I'm trying to create bindings to the GraphcicsMagick C library 
 which has the following struct defined:

 [...]


Do you mean fill .filename member from a D string? something like 
this?


import std.stdio;

 [...]



    struct Image
    {
      /* other members skipped */
      char[MaxTextExtent] filename;
    }
}

 [...]
import std.stdio;

extern(C)
{
 
    enum MaxTextExtent = 2053;



    struct Image
    {
      /* other members skipped */
      char[MaxTextExtent] filename;
    }
}

void main()
{
    import std.string : toStringz;
    import core.stdc.string : memcpy, strlen;

    auto i = Image();
    auto dstr = "hello world!!!"; // D string
    auto cs = toStringz(dstr); // now it's C-null-terminatted 
 string



    // since .filename isn't a pointer but an array, I think
    // you have to use memcpy() here. = operator wouldn't work 
 properly.
    memcpy(&i.filename[0], &cs[0], strlen(cs)+1);

    // casting away arrayness to make it a pointer (that a C's 
 array is after all)
    printf("str = [%s]\n", &i.filename[0]);
}

 [...]


note that even if .filename was a pointer, in order to the C 
converted stirng don't turn into garbage in a GC cycle, you would 
have to either keep reference to dstr around or malloc() and 
memcpy() it. So you need to convert from C to D: you can use 
fromStringz() on D string then to!string to convert to a D 
string, like this:


 import std.conv : to;
 import std.string : fromStrinz;
 string s = to!string(cstr.fromStringz);
Jan 04 2021
parent bdh <ben.h posteo.net> writes:
On Monday, 4 January 2021 at 16:35:23 UTC, Jack wrote:

 Do you mean fill .filename member from a D string? something 
 like this?

 [...]

     // since .filename isn't a pointer but an array, I think
     // you have to use memcpy() here. = operator wouldn't work 
 properly.
     memcpy(&i.filename[0], &cs[0], strlen(cs)+1);


I'm pretty sure I tried using = but it resulted in a type error 
(or array length mismatch).
Also tried memcpy which compiled, though due to the incorrect 
struct definition it appeared to not work.


     // casting away arrayness to make it a pointer (that a C's 
 array is after all)
     printf("str = [%s]\n", &i.filename[0]);


I'll keep this and ...


 note that even if .filename was a pointer, in order to the C 
 converted stirng don't turn into garbage in a GC cycle, you 
 would have to either keep reference to dstr around or malloc() 
 and memcpy()


... this in mind, thanks.
Jan 04 2021