No need to check for whether 'pBuffer' is NULL to call NetApiBufferFree, because:
- 'pBuffer' was used above this call without any particular checks;
- 'pBuffer' can be only null when NetUserEnum fails with an error different from NERR_Success or ERROR_MORE_DATA. But this case was actually checked for before using 'pBuffer'.
By Victor Martinez aka. Mr.Coverity ^^ :)
CID 1363629
CORE-11598 #resolve
svn path=/trunk/; revision=71929
Initialize utf8_ch_len to zero before using it (at each turn of the for-loop). As I couldn't find this code in Wine, I couldn't see whether they already fixed it or not. Caught by Victor.
CORE-11596
svn path=/trunk/; revision=71928
- Fix leaking 'pSetupData' in case of failure. CID 1363604
- Since we know that in the remaining of the code, 'pSetupData' is a valid pointer, we can free it at the very end without rechecking whether it was NULL or not.
CORE-11591 #resolve
svn path=/trunk/; revision=71927
Don't overrun FillGrid() when accessing the last elements. Rewrite the code using 'for' loops, in the same style as what is done elsewhere in the code. CID 1363552
CORE-11597 #resolve
svn path=/trunk/; revision=71924
- BiOpenKey is being feed with the wrong ElementHandle, it should be feed with ElementsHandle one. CID 1363670 . By Victor Martinez Calvo. CORE-11600 #resolve
[BOOTMGR]
- Fix an Assign vs Compare issue. CID 1363558 . By Victor Martinez Calvo. CORE-11592 #resolve
svn path=/trunk/; revision=71923
- Fix regression in API:RealGetWindowClass test due to wine sync/port.
- ReactOS use of DIALOG_get_info is an enhanced fork of the same wine function.
svn path=/trunk/; revision=71915
- Use StringCbCopyW
[PROGMAN]
- Use StringCbCopyW: CID #1363712.
- Don't read registry values in a registry key if we failed to open it. CID #514350.
svn path=/trunk/; revision=71912
- Don't hardcode buffer string size in GetPartTypeStringFromPartitionType calls.
- Possibly check whether the first character of the PartTypeString string is NULL before printing a generic partition information in case the partition type is unknown. We might check instead for the STRING_FORMATUNKNOWN string...
- Don't check for PartTypeString being NULL since it's not a pointer. CID #1363481 and CID #1363494.
svn path=/trunk/; revision=71911
- Cleanup error codes for Classes and Menus.
- Update (Sync/Port) User32:Dialog to Wine Staging 1.9.11, see CORE-11368.
- Now pass exactly the same tests as windows, except for the two ToDos lines 1596 & 1617. The other test failures are on wine for not being compatible!
svn path=/trunk/; revision=71904
CcMapData 4th arg isn't a boolean but a flag. Fix it accordingly
CcPinRead 4th arg isn't a boolean but a flag. Fix it accordingly
svn path=/trunk/; revision=71890
Call CcMapData(), CcPinRead() inside a SEH block
This allows returning more significant error codes on failure
[NTOSKRNL]
Enable Thomas code for raising exceptions
CORE-9848
svn path=/trunk/; revision=71888
- Raise exceptions on failure in CcMapData if RAISE_FROM_CC_MAP_DATA is defined. This is how the function should behave, and should be enabled once all our file system drivers correctly handle this.
CORE-9848
svn path=/trunk/; revision=71883
- Fix the last parameter of a StringCchCopyEx call; fix a misspelling.
- Do not exclusively use GetFileAttributesEx to retrieve file attributes (for the file properties dialog). Indeed, it happens that this API (as well as GetFileAttributes), and equivalently, NtQueryFullAttributesFile (as well as NtQueryAttributesFile), can fail on *locked system* files (on Win2k/Win2k3/Win7, and on ROS), such as C:\pagefile.sys . See for example http://stackoverflow.com/questions/16772931/getfileattributes-on-locked-system-file and my comment in CORE-10757 . Therefore to retrieve file attributes also for these files we do a trick: we call FindFirstFile on the full path to said file. It happens that FindFirstFile can work on locked system files, and thus we are able to retrieve their attributes (+ size and file dates) as well!
See the code for more details.
CORE-10757 #resolve
svn path=/trunk/; revision=71880
