I found an error in your doc

BaDanWang-3245 20

https://learn.microsoft.com/en-us/openspecs/office_file_formats/ms-ovba/c375cbf2-b9c6-4594-b064-6dfb830f1851

The size of Data in bytes MUST be CompressedChunk.CompressedChunkHeader.CompressedChunkSize (section 2.4.1.1.4) plus 3.

This should be plus 1.

Mike Bowen 1,276 Reputation points Microsoft Employee

2023-03-25T03:03:12.5633333+00:00

Hi @BaDanWang-3245,

Thanks for submitting your questions. One of our engineers will follow up with you soon.

Michael Bowen

Escalation Engineer - Microsoft Open Specifications
Tom Jebo 1,906 Reputation points Microsoft Employee

2023-03-26T06:27:28.5366667+00:00

Hi @BaDanWang-3245,

I will test this out and get back to you Monday at the latest.

Best regards,
Tom Jebo
Microsoft Open Specifications Support

BaDanWang-3245 20

according to your pseudocode in 2.4.1.3.2:

SET Header TO the CompressedChunkHeader (section 2.4.1.1.5) located at CompressedChunkStart
CALL Extract CompressedChunkSize (section 2.4.1.3.12) with Header returning Size
CALL Extract CompressedChunkFlag (section 2.4.1.3.15) with Header returning CompressedFlag
SET DecompressedChunkStart TO DecompressedCurrent
SET CompressedEnd TO the minimum of CompressedRecordEnd and (CompressedChunkStart PLUS Size)
SET CompressedCurrent TO CompressedChunkStart PLUS 2
IF CompressedFlag EQUALS 1 THEN
WHILE CompressedCurrent is LESS THAN CompressedEnd
CALL Decompressing a TokenSequence (section 2.4.1.3.4) with CompressedEnd
END WHILE
ELS

size of CompressedChunkData = CompressedChunkHeader.CompressedChunkSize + 3 - 2 = CompressedChunkHeader.CompressedChunkSize + 1

the pseudocode in doc is right

Tom Jebo 1,906 Reputation points Microsoft Employee

2023-03-28T00:01:50.3633333+00:00

Hi @BaDanWang-3245,

I reviewed the compression code for this again. I will need to write the decompression code to confirm this but I believe the code works both ways so that it can handle Office data in an interoperable manner. I can see that there appear to be "behaviors" in the operation of the pseudo code that might be confusing as well. Question: do you confirm that following the pseudo code in the document produces compressed data that upon uncompressing returns the original data? If so, is there any other reason you believe the document is incorrect?

It will take me a few days to verify the decompression unless you don't need me to do that.

Tom

BaDanWang-3245 20

@Tom Jebo the pseudo code is right, and i've written the code. I mean your document is wrong.

struct CompressedChunkHeader
{  
	uint16 CompressedChunkSize : 12;  //0~4095, 值为CompressedChunk大小-3
	uint16 CompressedChunkSignature : 3; // 必须为3
	uint16 CompressedChunkFlag : 1;      // 1为压缩, 0位未压缩
};

struct CompressedChunk
{
	CompressedChunkHeader cch;
	//uint8 CompressedChunkData[cch.size + 1];  // this should be plus 1,   not plus 3 as your doc(2.4.1.1.6)
};

cch.size plus 3 is the total CompressedChunk size.

and the CompressedChunkHeader is 2 bytes big.

so the CompressedChunkData is CompressedChunk size plus 3 minus 2 ---> CompressedChunkSize plus 1.

BaDanWang-3245 20 Reputation points

2023-03-28T01:36:03.4966667+00:00

the wrong doc is https://learn.microsoft.com/en-us/openspecs/office_file_formats/ms-ovba/c375cbf2-b9c6-4594-b064-6dfb830f1851

in this sentence:

"The size of Data in bytes MUST be CompressedChunk.CompressedChunkHeader.CompressedChunkSize (section 2.4.1.1.4) plus 3."

should change to:

"The size of Data in bytes MUST be CompressedChunk.CompressedChunkHeader.CompressedChunkSize (section 2.4.1.1.4) plus 1."

and maybe somewhere else needs to change.
Deleted

This comment has been deleted due to a violation of our Code of Conduct. The comment was manually reported or identified through automated detection before action was taken. Please refer to our Code of Conduct for more information.

Accepted answer

Tom Jebo 1,906 Reputation points Microsoft Employee

2023-03-28T18:23:06.54+00:00

Hi @BaDanWang-3245,

I was assuming you had implemented and that the results were correct but just wanted to make sure this description error was not, in your opinion, affecting the pseudocode or actual coding. I now agree that you are right. The description is not taking into account that the 2 byte header has been subtracted already.

The 3 bytes that it includes are the 2 byte header and then just 1 byte since the size was calculated with a difference in addresses which doesn't include the starting byte. So once you remove the size of the header for the chunk, then you still have to include the 1 byte increase to include the starting byte. So I now agree that that description needs to be changed as you mentioned to:

"The size of Data in bytes MUST be CompressedChunk.CompressedChunkHeader.CompressedChunkSize (section 2.4.1.1.4) plus 1."

I will file a bug to get the documentation updated.

Thank you for bringing this to our attention!

Best regards,
Tom Jebo
Microsoft Open Specifications Support
Please sign in to rate this answer.

1 person found this answer helpful.

0 comments No comments
Sign in to comment

I found an error in your doc

0 additional answers