6 Win32 function stubs, 8 constants, 1 structure, and 1 helper function #535

mlfreeman2 · 2015-11-14T01:39:20Z

Advapi32
- Added LOGON_WITH_PROFILE constant
- Added LOGON_NETCREDENTIALS_ONLY constant
- Added CreateProcessWithLogonW(String, String, String, int, String, String, int, Pointer, String, STARTUPINFO, PROCESS_INFORMATION)
  - No Unit Test - I would have to make it create a user to be 100% able to run a process as another user and that would be a security issue I figure.
Crypt32
- Added CertAddEncodedCertificateToSystemStore(String, Pointer, DWORD)
  - No Unit Test - I doubt anyone would want the security risk of a unit test installing a root certificate.
GDI32
- Added SRCCOPY constant
- Added BitBlt(HDC, int, int, int, int, HDC, int, int, int)
  - No direct unit test - the test for GDI32Util.getScreenshot() seemed to cover it just fine.
- Added GDI32Util.getScreenshot(HWND)
  - Added unit test as GDI32UtilTest.testGetScreenshot()
Shell32
- Added SHERB_NOCONFIRMATION constant
- Added SHERB_NOPROGRESSUI constant
- Added SHERB_NOSOUND constant
- Added SEE_MASK_NOCLOSEPROCESS constant
- Added SHEmptyRecycleBin(HANDLE, String, int)
  - No unit test - no idea how to tell if the recycle bin is empty afterwards
- Added ShellExecuteEx(SHELLEXECUTEINFO)
  - No unit test - there are a bunch of cases where the hProcess member in SHELLEXECUTEINFO isn't set - not sure how to control for that
ShellAPI
- Added SHELLEXECUTEINFO structure
User32
- Added GetDesktopWindow()
  - Added test as User32Test.testGetDesktopWindow()
WinGDI
- Added HGDI_ERROR
- Removed superfluous "public" and "public final" from WinGDI

Hopefully this is more manageable and more acceptable than my last way-too-large pull request

* Added LOGON_WITH_PROFILE constant * Added LOGON_NETCREDENTIALS_ONLY constant * Added CreateProcessWithLogonW(String, String, String, int, String, String, int, Pointer, String, STARTUPINFO, PROCESS_INFORMATION) * No Unit Test - I would have to make it create a user to be 100% able to run a process as another user and that would be a security issue I figure. * Crypt32 * Added CertAddEncodedCertificateToSystemStore(String, Pointer, DWORD) * No Unit Test - I doubt anyone would want the security risk of a unit test installing a root certificate. * GDI32 * Added SRCCOPY constant * Added BitBlt(HDC, int, int, int, int, HDC, int, int, int) * No direct unit test - the test for GDI32Util.getScreenshot() seemed to cover it just fine. * Added GDI32Util.getScreenshot(HWND) * Added unit test as GDI32UtilTest.testGetScreenshot() * Shell32 * Added SHERB_NOCONFIRMATION constant * Added SHERB_NOPROGRESSUI constant * Added SHERB_NOSOUND constant * Added SEE_MASK_NOCLOSEPROCESS constant * Added SHEmptyRecycleBin(HANDLE, String, int) * No unit test - no idea how to tell if the recycle bin is empty afterwards * Added ShellExecuteEx(SHELLEXECUTEINFO) * No unit test - there are a bunch of cases where the hProcess member in SHELLEXECUTEINFO isn't set - not sure how to control for that * ShellAPI * Added SHELLEXECUTEINFO structure * User32 * Added GetDesktopWindow() * Added test as User32Test.testGetDesktopWindow() * WinGDI * Added HGDI_ERROR * Removed superfluous "public" and "public final" from WinGDI

mlfreeman2 · 2015-11-14T01:40:15Z

I did try to incorporate all the feedback that @lgoldstein and @dblock gave me on the last (now closed) pull request.

…2 / Crypt32 functions

mlfreeman2 · 2015-11-14T18:27:42Z

Added negative unit tests for CreateProcessWithLogonW and CertAddEncodedCertificateToSystemStore

I figure those will be secure since they're meant to fail with specific windows error codes.

mlfreeman2 · 2015-11-14T20:40:30Z

Got creative and managed to figure out what I think might be reasonable negative tests for the remaining functions in this PR.
In theory every function added in this PR has some sort of test associated with it now.

lgoldstein · 2015-11-15T05:20:04Z

contrib/platform/src/com/sun/jna/platform/win32/Crypt32.java

+	 *         For extended error information, call GetLastError.
+	 * @see MSDN {@link http://msdn.microsoft.com/en-us/library/bb736347(v=vs.85).aspx }
+	 */
+	boolean CertAddEncodedCertificateToSystemStore(String szCertStoreName, Pointer pbCertEncoded, DWORD cbCertEncoded);


You can usually safely use int instead of DWORD here (as well as for return values). I agree that DWORD and int are not entirely equivalent, but for most purposes the sign is not important - e.g., error codes or sizes (as in this case).

lgoldstein · 2015-11-15T15:01:50Z

contrib/platform/test/com/sun/jna/platform/win32/Advapi32Test.java

+
+    	// should fail with "the user name or password is incorrect" (error 1326)
+    	assertEquals("GetLastError() should have returned ERROR_LOGON_FAILURE because the username was bogus.", Native.getLastError(), W32Errors.ERROR_LOGON_FAILURE);
+    }


Reverse the order of the arguments to assertEquals - the expected comes first (in this case W32Errors.ERROR_LOGON_FAILURE should be first and getLastError() second).

lgoldstein · 2015-11-15T15:07:59Z

contrib/platform/test/com/sun/jna/platform/win32/Shell32Test.java

+	}
+
+	public void testShellExecuteEx() {
+		File file = new File(System.getProperty("java.io.tmpdir") + System.nanoTime() + ".txt");


See previous comment regarding backslash

…ring instead of WString

lgoldstein · 2015-11-16T05:40:18Z

Seems in order - just add a line in the CHANGES.md file under the Next release / Features section using the same format as the others, and we will be able to merge your pull request.

P.S. You can use the git commit --amend --no-edit option:

Modify CHANGES.md
git add CHANGES.md
git commit --amend --no-edit
git push -v --progress -f ...your-fork... ...branch-name:branch-name...

…ring instead of WString

dblock · 2015-11-16T12:07:13Z

I merged via db86ae6.

We can iterate on top for anything else that's left. Please try to make small, individual PRs. Thanks for contributing! Thanks for a thorough CR @lgoldstein.

dblock · 2015-11-16T12:07:56Z

@mlfreeman2 You have been working on JNA master on your fork, you should read http://code.dblock.org/2015/08/31/getting-out-of-your-first-git-mess.html.

…ava-native-access#535) Motivation: We need to ensure we also release the memory in the case of not be able to allocate a CRYPTO_BUFFER. Modifications: Move release calls before the early returns Result: No more possible leak

mlfreeman2 added 2 commits November 13, 2015 21:52

forgot to remove this system.out.println

7baceb6

added constants to WinCrypt and added negative unit tests for Advapi3…

4cf5188

…2 / Crypt32 functions

got creative - managed to come up with tests for remaining functions

1965478

lgoldstein reviewed Nov 15, 2015
View reviewed changes

addressed feedback from @lgoldstein

d8876fc

lgoldstein reviewed Nov 15, 2015
View reviewed changes

mlfreeman2 added 2 commits November 15, 2015 10:38

addressing latest round of feedback

8da5c17

applied fix suggested by @twall to make SHELLEXECUTEINFO work with St…

2833aca

…ring instead of WString

mlfreeman2 added 2 commits November 16, 2015 06:21

applied fix suggested by @twall to make SHELLEXECUTEINFO work with St…

7750103

…ring instead of WString

Merge branch 'master' of https://github.com/mlfreeman2/jna

29712fb

dblock closed this Nov 16, 2015

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

6 Win32 function stubs, 8 constants, 1 structure, and 1 helper function #535

6 Win32 function stubs, 8 constants, 1 structure, and 1 helper function #535

mlfreeman2 commented Nov 14, 2015

mlfreeman2 commented Nov 14, 2015

mlfreeman2 commented Nov 14, 2015

mlfreeman2 commented Nov 14, 2015

lgoldstein Nov 15, 2015

lgoldstein Nov 15, 2015

lgoldstein Nov 15, 2015

lgoldstein commented Nov 16, 2015

dblock commented Nov 16, 2015

dblock commented Nov 16, 2015

6 Win32 function stubs, 8 constants, 1 structure, and 1 helper function #535

6 Win32 function stubs, 8 constants, 1 structure, and 1 helper function #535

Conversation

mlfreeman2 commented Nov 14, 2015

mlfreeman2 commented Nov 14, 2015

mlfreeman2 commented Nov 14, 2015

mlfreeman2 commented Nov 14, 2015

lgoldstein Nov 15, 2015

Choose a reason for hiding this comment

lgoldstein Nov 15, 2015

Choose a reason for hiding this comment

lgoldstein Nov 15, 2015

Choose a reason for hiding this comment

lgoldstein commented Nov 16, 2015

dblock commented Nov 16, 2015

dblock commented Nov 16, 2015