David Quintana

* The work has been merged to trunk, this branch is now obsolete.
  1. … 27751 more files in changeset.
* Get rid of the old branch where I did my shell work. It was merged 3 years ago.
  1. … 19504 more files in changeset.
[RAPPS]

* Fix linux build? maybe? please?

Merge RAPPS improvements branch into trunk.

Contains the work done by sanchaez (Alexander Shaposhnikov) during GSoC 2017:

* Better information in the application details view.

* Bulk installation of multiple applications at once.

* Command-line automation of installation (single and bulk).

(for more details, see the related blog posts)

TRANSLATORS: Many resources have been changed, and they are in need of review!

    • -127
    • +143
    /trunk/reactos/base/applications/rapps/settingsdlg.cpp
  1. … 44 more files in changeset.
* Fix svn merge-info confusion regarding this file.
* Sync with trunk HEAD (r75820).
  1. … 2700 more files in changeset.
Given that VS2010 is the newest one that has a chance to work in ReactOS, that may take a while.

Given that VS2010 is the newest one that has a chance to work in ReactOS, that may take a while.

Yep seems like I was thinking of something else. Doing a quick search through our CMakeLists.txt files, it seems like we do this in quite a few places, so it's OK.

Yep seems like I was thinking of something else. Doing a quick search through our CMakeLists.txt files, it seems like we do this in quite a few places, so it's OK.

don't we have some "add_include_files" for this?

don't we have some "add_include_files" for this?

In that case it would be best to have some prefix, like LicenseType_OpenSource and so on, rather than "polluting" the global namespace with names such as "Min" or "None".

In that case it would be best to have some prefix, like LicenseType_OpenSource and so on, rather than "polluting" the global namespace with names such as "Min" or "None".

This looks a lot like a C# enum. It doesn't take into account that C++ enums are not namespaced, which means it puts "None", "OpenSource" etc into the global namespace. There is "enum class" in C++...

This looks a lot like a C# enum. It doesn't take into account that C++ enums are not namespaced, which means it puts "None", "OpenSource" etc into the global namespace. There is "enum class" in C++11 which is namespaced, but I'm not sure if VS2010 supports that. Regardless of style choice, this would have to be fixed one way or another.

[0.4.6]

- Update explorer banner

[EXPLORER]

Update start menu banner.

[FREELDR]

Fix compilation on ARM.

[BOOTLIB]

[BOOTMGR]

Disabled x86-specific code when building ARM.

[LIBSUPP]

Use a hack to replace inline asm which is not supported by MSVC ARM.

[DBGHELP]

Fix ARM register access from context struct.

[MSVCRT]

Fix typo in spec file.

[NTOSKRNL]

[WIN32SS]

Add asm dependency to *sys modules.

[CPPRT]

Add one of the missing constructor aliases. There's more, but my brain is fried at this point.

[BTRFS]

Define-away an include for x86-specific intrinsics.

This is the first batch of fixes for building reactos using the MSVC ARM toolchain. A lot more work is needed to get a full build including rostests and rosapps.

    • -0
    • +3
    /trunk/reactos/sdk/lib/delayimp/delayimp.c
    • -1
    • +5
    /trunk/reactos/boot/environ/CMakeLists.txt
    • -0
    • +3
    /trunk/reactos/sdk/lib/cpprt/arm/cpprt.s
    • -0
    • +5
    /trunk/reactos/dll/ntdll/rtl/libsupp.c
    • -0
    • +1
    /trunk/reactos/ntoskrnl/CMakeLists.txt
    • -0
    • +15
    /trunk/reactos/boot/environ/lib/platform/time.c
[CABMAN]

Fix pointer math ('unsigned long' will truncate the pointer).

[MKISOFS/SCHILY TOOLS]

Fix the invalid assumption that a 'long' variable can hold pointers. Will be sent upstream.

With this the whole of reactos and rostests builds on x64. Rosapps still has issues.

    • -2
    • +2
    /trunk/reactos/sdk/tools/cabman/mszip.cxx
[APITESTS]

MSVCRT_CRT_APITEST: Remove getmainargs tests from X64 build.

DELAYIMP: Fix callback prototype.

CComHeapPtr: Fix implementation of IMallocSpy to match the actual interface.

[WINETESTS]

ntdll: Disable 'generated' tests when not i386 platform, since they were generated for x86 and not x64 or ARM.

Fixes x64 compilation for tests.

    • -0
    • +2
    /trunk/rostests/apitests/crt/testlist.c
    • -1
    • +1
    /trunk/rostests/apitests/sdk/delayimp.cpp
[BOOTLIB]

Stub out some arch-specific functions (fixes x64 build).

[USBPORT]

Actually fix x64 build. The old code was not working at all, but the bootlib errors cancelled the build earlier than I thought.

The X64 compilation now succeeds, but cabman crashes trying to generate the bootcd.

    • -0
    • +27
    /trunk/reactos/boot/environ/lib/mm/stub/mm.c
    • -1
    • +7
    /trunk/reactos/drivers/usb/usbport/usbport.h
[EXPLORER]

[RSHELL]

[NTOBJSHEX]

[BROWSEUI]

[SHELL32]

Get rid of the unused "placement new" operators.

    • -5
    • +0
    /trunk/reactos/dll/win32/shell32/shell32.cpp
    • -5
    • +0
    /trunk/reactos/base/shell/explorer/explorer.cpp
    • -5
    • +0
    /trunk/reactos/dll/win32/browseui/browseui.cpp
    • -5
    • +0
    /trunk/reactos/base/shell/rshell/misc.cpp
[USBPORT]

Oops, brainfart (Fixes 32bit build).

[MSPAINT]

Correct callback return type (Fixes X64 build).

[USETUP]

Correct definition of the struct (fixes X64 build).

[USBPORT]

Improve struct size checks so they are X64-compatible (fixes X64 build).

The build as a whole still fails due to bootlib, but implementing those is beyond my expertise.

    • -2
    • +7
    /trunk/reactos/drivers/usb/usbport/usbport.h
Nothing wrong with this, just wanted to note that we have a CToolbar class in rosctls.h, which could be used to reduce the amount of code and improve readability.

Nothing wrong with this, just wanted to note that we have a CToolbar class in rosctls.h, which could be used to reduce the amount of code and improve readability.

[NTOBJSHEX]

Fix registration.

[NTOBJSHEX]

Make nt object symlinks work more like shortcuts. Fixes the addressbar not quite knowing what to show when you click on a symlink.

[NTOBJSHEX]

Make GCC happy.

[NTOBJSHEX]

* Added a common superclass for the ntobj and registry folders.

* Refactored the ntobj and registry folders to make use of this superclass.

* Removed the concept of a "pidl helper", since it works best to just have the methods be part of the folders.

* Cleaned up the ItemID comparison logic.

    • -95
    • +55
    /trunk/reactos/dll/shellext/ntobjshex/regfolder.h
    • -0
    • +676
    /trunk/reactos/dll/shellext/ntobjshex/foldercommon.h
    • -0
    • +10
    /trunk/reactos/dll/shellext/ntobjshex/foldercommon.cpp
    • -716
    • +198
    /trunk/reactos/dll/shellext/ntobjshex/ntobjfolder.cpp
[NTOBJSHEX]

* Fix gcc build.

* Remove magic.

* Address nitpicks.

[NTOBJSHEX]

Fixed registry value display for values that are too long to fit inside the shitemid.

Set the maximum for shitemid embedding back to a smaller number.

[NTOBJSHEX]

* Fixed NT object symbolic link target retrieval.

* Fixed NT path parsing (didn't consider paths sub sub-folders in them).

* Fixed BindToObject to properly resolve the target path for symlinks.

* Made the maximum embedded content length for registry keys a bit bigger.

[NTOBJSHEX]

Improve the GetDisplayNameOf logic from "badly broken" to "seems to work". I'm not convinced it's perfect, but it's better than before.