- rework the kmtest loader application to follow everything else's coding style and to allow user-mode test parts
svn path=/branches/GSoC_2011/KMTestSuite/; revision=52519
- add skip() functionality (with more useful syntax than Wine's)
- add format attributes for GCC
- do not show the full path of the source file for failed tests
- general cleanup
svn path=/branches/GSoC_2011/KMTestSuite/; revision=52489
- add START_TEST macro in case test functions ever have to be passed an argument
- add a few ok_* convenience macros
- remove unnecessary C library function renaming macros
svn path=/branches/GSoC_2011/KMTestSuite/; revision=52371
- add the proper testing framework functions for winetest-like feel
- remove thereby obsolete Log* functions
- update example test to show usage of the new macros
svn path=/branches/GSoC_2011/KMTestSuite/; revision=52359
- reference streamout.c and provide a wctomb stub to have an any-IRQL vsnprintf implementation for the test framework
svn path=/branches/GSoC_2011/KMTestSuite/; revision=52330
- use a shared memory buffer for storing test results to provide seamless communication between all test parts
- Now user mode code can easily add messages to the buffer
svn path=/branches/GSoC_2011/KMTestSuite/; revision=52216
- add a minimal example test so there's something to see when running kmtest
- remove _DLL and __USE_CRTIMP in accordance with r52213
svn path=/branches/GSoC_2011/KMTestSuite/; revision=52215
- add simple loader application that can list/run tests from a driver via DeviceIoControl, and prints output to the console
svn path=/branches/GSoC_2011/KMTestSuite/; revision=52157
- Launch a new listen IRP right after the current one completes instead of waiting until the current pending connection is accepted
- Reduces the chance of AFD missing two connection requests that are issued very close together and increases network performance by allowing our connection queue to do its job
svn path=/trunk/; revision=52013
- Implement sending GUID_TARGET_DEVICE_QUERY_REMOVE and GUID_TARGET_DEVICE_REMOVE_COMPLETE notifications
- Fix sending EventCategoryTargetDeviceChange notifications
- Remove some incorrect checks (IRP_MN_START_DEVICE CAN be called for a device that is already "started")
- Check the final status of the IRP_MN_START_DEVICE request sent after resource rebalancing and send IRP_MN_REMOVE_DEVICE if it fails
svn path=/trunk/; revision=52003
- Simplify remove lock tracking block list handling.
- Correct check for unlimited locking time being allowed.
- Eliminate use-after-free after removing a tracking block.
svn path=/trunk/; revision=52002
- dll/win32 and dll/shellext converted
- currently, only shell32 is left, due to incoming change to new version, and kernel32 - need to research how it should be done.
Tested on rbuild/cmake
svn path=/trunk/; revision=52000
- Fix a reference leak which prevented driver objects passed to IoRegisterPlugPlayNotification from being able to unload after the notification was unregistered
- Fix a non-paged pool leak
svn path=/trunk/; revision=51997
- Add stubs for CM_Set_Class_Registry_PropertyA/W.
- Add support for missing properties to CM_Set_DevNode_Registry_Property_ExA/W.
svn path=/trunk/; revision=51996
* Improve reactos.cab creation a bit: create it using a custom command instead of a custom target.. marking it as output this way ensures a proper cleanup with make clean and co.
svn path=/trunk/; revision=51987
- Get screen saver running. Need more work on WinLogon register issues and problems.
- Miscellaneous changes and cleanups.
svn path=/trunk/; revision=51971
- Load other driver sections from txtsetup.sif
- Enable a debug print
[TXTSETUP]
- Add PnP entries for storage devices
[PCIIDE/PCIIDEX]
- Add to bootcd
- Both usetup's PnP manager and the user-mode PnP manager now recognize PCI IDE channels
svn path=/trunk/; revision=51969
- PCIIDE should be installed for each PCI IDE controller and UNIATA should be installed for each IDE channel
- Remove an extra comma in the compatible adapters entries
svn path=/trunk/; revision=51968
- Don't report raw device support to allow the IDE channel driver (uniata) to take over control of the enumerated devices
- Ideally the PnP manager should deal with a driver collision like this but it doesn't currently
svn path=/trunk/; revision=51967
- Add a special case in IopLoadServiceModule for 1st stage setup when we have no service entries so we can still load drivers
- Fixes an old regression with 1st stage PnP which now can load drivers again
svn path=/trunk/; revision=51966
- Determine whether the device is really installed by reading the ClassGUID key instead of looking at the device flags
- Add a debug print for device installation (shouldn't be too spammy except and will be helpful for hotplugging)
[USETUP]
- Reenable the 1st stage PnP manager
svn path=/trunk/; revision=51963
- Implement a simple property sheet provider for serial ports. The property sheet does not show yet because the dialog resource is not found because a call to FindResourceW fails.
svn path=/trunk/; revision=51960
- SetupDiGetClassDevPropertySheetsW: Improve the way property sheet pages are added to a property sheet. This change enables us to retrieve the number of property sheet pages provided by a property sheet provider without leaking property sheets.
svn path=/trunk/; revision=51959
- Disable the 1st stage user-mode PnP manager because it is creating incomplete registry keys which is confusing the real user-mode PnP manager and preventing devices installed in 1st stage from appearing in device manager
- Fixes displaying PCI bus controllers in device manager
- I need to discuss with Herve what should be done with this code
svn path=/trunk/; revision=51947
