- Use the new function SampSetObjectAttributeString to write strings to the registry. Handle empty strings properly.
- Fix SampGetObjectAttributeString to handle empty strings properly as well.
svn path=/trunk/; revision=59964
* Implement an undocumented flag for the rebar control. This flag is used by browseui and is needed to show the explorer window properly. Should be sent Upstream. Brought to you by Giannis Adamopoulos (Smiley).
* Update the sync diff.
svn path=/trunk/; revision=59962
* Stretch/Skew dialog fully implemented (skewing not yet functional), please check translations
* pre Windows Vista MSPaint palette selectable via menu entry
* Attributes dialog supports setting width and height in inches and cm also
* drawing functions improved
* some variables grouped into structs
* Changed SendMessage(... WM_PAINT ...) into InvalidateRect(...), this is the way it should be and it fixes half of the repainting bugs
svn path=/trunk/; revision=59954
Remove operator LPWSTR() in favor of operator LPCWSTR() const (likely to fix its export ;-))
Add missing operators implementations as inline
svn path=/trunk/; revision=59947
- Improve SamRegCloseKey and fix all calls to it accordingly. Also replace all calls to NtClose by calls to SampRegCloseKey.
- Add a resource to protect the Samr functions from concurrent access.
svn path=/trunk/; revision=59945
Partplement framedyn.dll
So far, only CHString class is implemented in it
Regarding CHString class implementation:
It has been implemented matching Windows behavior and bugs
It might not be really nice at some points but this ensures
compatibility with applications relying on it.
As a side note: this might explain why this API was abandonned after w2k3...
Regarding the DLL itself:
Its implementation into ReactOS allows systeminfo.exe application from w2k3 to
load and start initialization into ReactOS. It stops working later on with some
IDispatch error.
The DLL cannot be used to build any application against framedyn.dll with our "default" BE (read with G++).
Only MSVC mangled name are exported so far (as on Windows ;-)) thanks to aliasing. This is why there are two spec files.
This makes G++ unable to link with it though. This is the reason why there are no framedyn_apitests so far.
Furthermore, the DLL only have exports for x86 and none for x64 for the moment.
Any help is welcome for having exports for x64 target and to be able to have aliasing for G++ (from its mangled names
to MSVC ones).
svn path=/trunk/; revision=59940
Add two headers: chstring.h and provexce.h
They are used for framedyn.dll stuff, for instance
So far, the CHString class is not fully populated. It is only
populated with exported stuff from framedyn.dll
svn path=/trunk/; revision=59939
* Don't overwrite the MZ header when there are no long section names.
* Correct an error including 4 more bytes than needed in the coff strings.
* Brought to you by Arty.
svn path=/trunk/; revision=59938
- Convert some language resource files into UTF-8 without BOM (aka. ANSI as UTF-8 with Notepad++).
- Improve some French translation.
- Fix Czech translation, thanks Radek :)
(Addendum to r59924 and r59925)
svn path=/trunk/; revision=59937
[NTDLL-KERNEL32-BASESRV-WINSRV]: Specify by hand (hackz for MSVC builds) that the dll will be a "Console DLL" (ie. not a native one nor a Gui DLL).
[BASESRV]: Explicitely set the DLL entry point to DllMain (unusued).
[WINSRV]: Explicitely set the DLL entry point to DllMain (unusued) AND REMOVE usage of msvcrt (that might pose problems when testing with Windows msvcrt etc...) and use libcntpr instead.
Thanks to Amine for having helped me in solving a problem with delayed imports.
svn path=/trunk/; revision=59933
Implement RtlDefaultNpAcl, needed for giving a default security to named pipes. Makes Win2k3 kernel32 happy when one creates named pipes for SCM and so on.
Getting the SID of the caller is done as described in http://www.osronline.com/article.cfm?article=50 in "Extracting the SID from the Token".
svn path=/trunk/; revision=59928
- Improve the stub for CAddressEditBox::SetOwner to make the rest of browseui happy for now
- In CShellBrowser::Initialize browse to the pidl that is passed as parameter instead of always browsing to the desktop folder
- Fix a bug in atl class CContainedWindowT::DefWindowProc that was calling the window proc with an invalid hwnd.
svn path=/trunk/; revision=59927
- FsRtlCopyWrite: Fix typos in local var names and comments.
- FsRtlCopyWrite: Fix copypasta in a call to FastIoCheckIfPossible: wait flag shouldn't be always TRUE in this case (it's determined by Wait parameter).
- FsRtlCopyRead: Fix updating of CurrentByteOffset.
- FsRtlAcquireFileForCcFlushEx / FsRtlReleaseFileForCcFlush: Handle the case when FastIO callback may fail and FsRtl still needs to act as if FastIO callback would be missing at all.
- Implement FsRtlAcquireFileForModWriteEx / FsRtlReleaseFileForModWrite thanks to Rajeev Nagar's book.
- Now all necessary FsRtl locking/unlocking mechanisms exist.
svn path=/trunk/; revision=59926
* Remove some Wine specific text and introduce a tiny update to the German translation. By zehnvor.
CORE-7221 #resolve #comment Committed in r59924. Danke.
svn path=/trunk/; revision=59924
- Halfplement CDesktopBrowser::BrowseObject. Now when the user tries to open a folder from the desktop we will try to create a new explorer window.
- ps: this affects only explorer_new and the rest of the shell is still not working well enough to actually show the new window
svn path=/trunk/; revision=59918
