[Samba] Code? or Test Pattern?

Kenji Ichinoseki ichinoseki at sei-networks.com
Thu Jun 23 19:17:29 MDT 2011

My name is Kenji Ichinoseki and I am in charge of 
a project at Sumitomo Electric Networks, Inc.

Please give me cooperation by all means about the affair in the account of a title.
Now, I included "Samba 3.0.37" in the Linux system currently developed (us).
And it is testing using "smbtorture(samba3.5.8 source4) ".

The point in question was found in our entry of this test.
It has a question to implementation of "smbtorture" of "MODE_INFORMATION" 
which is an entry of a "RAW-SFILEINFO" test.


-- it is written to the 1114th line in this URL as follows.

Request is made using the MODE_INFORMATION (1016) info level.  File is
referenced by NAME.
	Set file mode to 0.

	SMBTorture Output: None.

	Server should respond with NT_STATUS_OK
		MODE_INFORMATION query response should show that the file mode is set to 0.
		Same test as above except the file is referenced by name.

Please see "source4/torture/raw/setfileinfo.c."

In "CHECK_CALL_PATH" in the 409th line of a version 3.5.8, "2" is set up to "mode".

 	printf("test mode_information level\n");
 	sfinfo.mode_information.in.mode = 2;
 	CHECK_VALUE(MODE_INFORMATION, mode_information, mode, 2);

 	CHECK_VALUE(MODE_INFORMATION, mode_information, mode, 0);

 	sfinfo.mode_information.in.mode = 1;

This is the same also in version "3.5.9" and "4.0.0 alpha15."

I think that I should perform "sfinfo.mode_information.in.mode = 0;", 
before executing "CHECK_CALL_PATH", but is this understanding a mistake?

Or is it a mistake in a document?


Kenji Ichinoseki (ichinoseki at sei-networks.com)

More information about the samba mailing list