entity MTEXT member text reads only the first character about libredwg

[iisaxin]

我的英语不太好，请尽量仔细的看懂我的话，
我的意思是你的dwg的文件版本太高，
你试试读取r2000版本的文件看看问题是否存在

[rurban]

You are confused. >+ R2007 all text values are Windows unicode (i.e UCS-2),
With the rest you are also apparently confused. internal Dwg_Object * addresses are not offsets into the file.

[IdoGuzi]

Thanks for answering rurban.
trying to get the text as 16 bit chars works now, I'm happy it's not a bug.
Is there a place in the documentation taking about the text format?

As I said in the post I'm on version 0.12.5 and I see the member's type has changed but the comment is the same.
quoting: "byte offset in the file".

libredwg/include/dwg.h

Line 8958 in 76d2f0d

size_t address; /*!< byte offset in the file */

Can you please briefly explain to what this address is pointing.

[rurban]

Dwg_Object *obj->address is indeed the byte offset. But you were talking about the obj ptr, not obj->address.

The docs at https://www.gnu.org/software/libredwg/manual/LibreDWG.html#Types mentions

BITCODE_TU wchar*
length + windows 2-byte wchar string (UCS-2). The default text type since r2007.

I.e. all T types are TU since r2007, else TV,

[IdoGuzi]

From the start I meant the obj.address, probably language barrier (I'm not a native English speaker).
I printed the address like this (looping over all dwg_objects):
printf("size: %d, address: %d\n", object.size, object.address);
and got: 193, 275739
and later ran

int test(void) {
	unsigned int size = 193;
	unsigned int address = 275739;
	char buf[size];
	int fd = open("my_file.dwg", O_RDONLY);
	printf("fd: %d\n", fd);
	int err = pread(fd, (void *)buf, size, address);
	close(fd);
	if (err < 0) {
		printf("failed to read mtext raw data, what: %s\n", strerror(err));
		return err;
	}
	printf("bytes read: %d\n", err);
	printf("size: %d, address: %d, raw data: %s\n\n\n", size, address, buf);
	return 0;
}

pread returns 0 for all mtexts byte offset in the file (by pread docs it means EOF).
can you see any problem with the way I tried to get out data?

[rurban]

od --skip-bytes=275739 --read-bytes 193 my_file.dwg would be easier.
dwgread -v5 my_file.dwg would also be useful, as it converts your mtext value to utf-8

[IdoGuzi]

It's easier (thanks), but as I expected od gives:
od: cannot skip past end of combined input

[iisaxin]

Try changing your dwg version to r2000,
Too high version to work properly

[IdoGuzi]

Try changing your dwg version to r2000,
Too high version to work properly

I'm guessing you are talking about the object byte offset.
The thing is it looks like the file is parsed correctly, so I don't think it's a version problem.
After @rurban explained that the text isn't utf-8 (and it was a rookie mistake on my part) I'm extracting the text and all the other values just fine.
The only problem I see is with the byte offset, though this problem doesn't affect my work, I would like to help make sure there's not a bug here.