On 03/22/2010 01:01 PM, Nikolay Sivov wrote:
Implement some IShellFolderView methods over IFolderView
Hi Nikolay,
+ /* ::MoveIcons */ + hr = IShellFolderView_MoveIcons(folderview, NULL); + ok(hr == E_NOTIMPL, "got (0x%08x)\n", hr); +
This one makes Win98 and win95 (both with shell32.dll version 4.72.x) crash.
The same is true for the one from patch "Add IShellFolderView::SetRedraw implementation":
+ /* ::SetRedraw without list created */ + hr = IShellFolderView_SetRedraw(folderview, TRUE); + ok(hr == S_OK, "got (0x%08x)\n", hr); +
(Changing TRUE to FALSE doesn't make this test crash!).
There is nothing in those tests yet that indicates version 5.0 (if that's what makes these tests succeed) or higher. It's easy enough to skip on Win9x but that also means no tests or WinMe or don't we care in this case?
On 3/22/2010 22:09, Paul Vriens wrote:
On 03/22/2010 01:01 PM, Nikolay Sivov wrote:
Implement some IShellFolderView methods over IFolderView
Hi Nikolay,
- /* ::MoveIcons */
- hr = IShellFolderView_MoveIcons(folderview, NULL);
- ok(hr == E_NOTIMPL, "got (0x%08x)\n", hr);
This one makes Win98 and win95 (both with shell32.dll version 4.72.x) crash.
I prefer to keep it for all platforms. We need stub IDataObject in tests for it. I'll try with a stub tomorrow maybe, if you could be faster please go ahead.
The same is true for the one from patch "Add IShellFolderView::SetRedraw implementation":
- /* ::SetRedraw without list created */
- hr = IShellFolderView_SetRedraw(folderview, TRUE);
- ok(hr == S_OK, "got (0x%08x)\n", hr);
(Changing TRUE to FALSE doesn't make this test crash!).
It's strange. Did you try it separately from others? If it's true there's no problem to change in to FALSE. It means nothing for this test.
There is nothing in those tests yet that indicates version 5.0 (if that's what makes these tests succeed) or higher. It's easy enough to skip on Win9x but that also means no tests or WinMe or don't we care in this case?
It's better to run them all.
On 03/22/2010 09:22 PM, Nikolay Sivov wrote:
On 3/22/2010 22:09, Paul Vriens wrote:
On 03/22/2010 01:01 PM, Nikolay Sivov wrote:
Implement some IShellFolderView methods over IFolderView
Hi Nikolay,
- /* ::MoveIcons */
- hr = IShellFolderView_MoveIcons(folderview, NULL);
- ok(hr == E_NOTIMPL, "got (0x%08x)\n", hr);
This one makes Win98 and win95 (both with shell32.dll version 4.72.x) crash.
I prefer to keep it for all platforms. We need stub IDataObject in tests for it. I'll try with a stub tomorrow maybe, if you could be faster please go ahead.
I'm not sure what you mean here. This MoveIcons call crashes on Win95 and Win98 so we must skip this somehow.
The same is true for the one from patch "Add IShellFolderView::SetRedraw implementation":
- /* ::SetRedraw without list created */
- hr = IShellFolderView_SetRedraw(folderview, TRUE);
- ok(hr == S_OK, "got (0x%08x)\n", hr);
(Changing TRUE to FALSE doesn't make this test crash!).
It's strange. Did you try it separately from others? If it's true there's no problem to change in to FALSE. It means nothing for this test.
I did try just the SetRedraw call and it crashes with 'TRUE' on Win95 and Win98.
There is nothing in those tests yet that indicates version 5.0 (if that's what makes these tests succeed) or higher. It's easy enough to skip on Win9x but that also means no tests or WinMe or don't we care in this case?
It's better to run them all.
All is not possible (see above) as some tests crash.
On 03/23/2010 09:43 AM, Paul Vriens wrote:
On 03/22/2010 09:22 PM, Nikolay Sivov wrote:
On 3/22/2010 22:09, Paul Vriens wrote:
On 03/22/2010 01:01 PM, Nikolay Sivov wrote:
Implement some IShellFolderView methods over IFolderView
Hi Nikolay,
- /* ::MoveIcons */
- hr = IShellFolderView_MoveIcons(folderview, NULL);
- ok(hr == E_NOTIMPL, "got (0x%08x)\n", hr);
This one makes Win98 and win95 (both with shell32.dll version 4.72.x) crash.
I prefer to keep it for all platforms. We need stub IDataObject in tests for it. I'll try with a stub tomorrow maybe, if you could be faster please go ahead.
I'm not sure what you mean here. This MoveIcons call crashes on Win95 and Win98 so we must skip this somehow.
The same is true for the one from patch "Add IShellFolderView::SetRedraw implementation":
- /* ::SetRedraw without list created */
- hr = IShellFolderView_SetRedraw(folderview, TRUE);
- ok(hr == S_OK, "got (0x%08x)\n", hr);
(Changing TRUE to FALSE doesn't make this test crash!).
It's strange. Did you try it separately from others? If it's true there's no problem to change in to FALSE. It means nothing for this test.
I did try just the SetRedraw call and it crashes with 'TRUE' on Win95 and Win98.
Weird, I just checked with your new patchset and now SetRedraw doesn't fail/crash anymore. Got rid off your patchset and SetRedraw still succeeds. No idea what went wrong here, probably my stupidity.
MoveIcons still crashes though so we need to work around that one.
On 3/23/2010 11:43, Paul Vriens wrote:
On 03/22/2010 09:22 PM, Nikolay Sivov wrote:
On 3/22/2010 22:09, Paul Vriens wrote:
On 03/22/2010 01:01 PM, Nikolay Sivov wrote:
Implement some IShellFolderView methods over IFolderView
Hi Nikolay,
- /* ::MoveIcons */
- hr = IShellFolderView_MoveIcons(folderview, NULL);
- ok(hr == E_NOTIMPL, "got (0x%08x)\n", hr);
This one makes Win98 and win95 (both with shell32.dll version 4.72.x) crash.
I prefer to keep it for all platforms. We need stub IDataObject in tests for it. I'll try with a stub tomorrow maybe, if you could be faster please go ahead.
I'm not sure what you mean here. This MoveIcons call crashes on Win95 and Win98 so we must skip this somehow.
I mean add a stub IDataObject to our test file and pass not null pointer to MoveIcons. I'll try that for next patch set.
-
Nikolay Sivov -
Paul Vriens