How to detect DF_DSM2 flag correctly?

[Michael]

Hi, everyone!

I defined DF_ flags in accordance with “twain.h” file:

#define DF_DSM2 0x10000000L
#define DF_APP2 0x20000000L
#define DF_DS2 0x40000000L

It means that all these flas reside in the high byte of 32bit TW_IDENTITY.SupportedGroups field.

Then I initialized TW_IDENTITY.SupportedGroups as “DF_APP2 | DG_CONTROL | DG_IMAGE” which in binary is “0010 0000 0000 0000 0000 0000 0000 0011”. That’s OK so far.

After executing of “DG_CONTROL, DAT_PARENT, MSG_OPENDSM” the TW_IDENTITY.SupportedGroups is “0010 0000 0000 0001 0000 0000 0000 0011”.

That is DSM set a bit but it doesn’t look like DF_DSM2. What’s wrong? Is something wrong with data alignment in my TW_IDENTITY structure? Or DSM allways
set bit in a middle of 32bit value?

I would appreciate any help.

[Michael]

Yes! It was C++ structure data alignment!

I used “#pragma pack(push, 2)” and found DF_DSM2 flag in its correct place.

But i’m stilll confused. Does DSM receive pointer to application’s TW_IDENTITY structure and then use fixed byte offset to find SupportedGroups field? Weird…

Anyway, “all-clear”! Thanks.

[Author]

Posts