https://git.reactos.org/?p=reactos.git;a=commitdiff;h=7c81fb8f568c0dc875018…
commit 7c81fb8f568c0dc875018fa382d215c088eb839d
Author: Joachim Henze <Joachim.Henze(a)reactos.org>
AuthorDate: Wed May 6 23:44:50 2020 +0200
Commit: Joachim Henze <Joachim.Henze(a)reactos.org>
CommitDate: Wed May 6 23:45:37 2020 +0200
[FREELDR] Workaround regression CORE-16469
This prevents ReactOS asserting when 'My computer'
is opened, while it tries to send commands to floppy drive.
Many thanks to patches author Doug Lyons.
The regression was introduced by 0.4.13-dev-1081-g
eeff926ede1b2109bb0e129a960b74f3261cf2f1
patch was committed to 0.4.13rls and 0.4.14rls as well.
Today it was committed to master as well, as
the initially planned investigation for the root cause
did still not happen and we can not afford the time
to retest and workaround this over and over again.
Thank god that Oleg Dubinskiy was still around to retest
this again and confirmed it can still happen, because
unlike initially, personally I was not able to reproduce it
today anymore!
Since every release was work-arounded, we did never
expose the bug in any final release.
---
boot/freeldr/freeldr/arch/i386/pc/pcdisk.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/boot/freeldr/freeldr/arch/i386/pc/pcdisk.c b/boot/freeldr/freeldr/arch/i386/pc/pcdisk.c
index 37058bced70..550c424d15d 100644
--- a/boot/freeldr/freeldr/arch/i386/pc/pcdisk.c
+++ b/boot/freeldr/freeldr/arch/i386/pc/pcdisk.c
@@ -790,7 +790,7 @@ PcDiskReadLogicalSectors(
#if defined(__i386__) || defined(_M_AMD64)
VOID __cdecl DiskStopFloppyMotor(VOID)
{
- WRITE_PORT_UCHAR((PUCHAR)0x3F2, 0x0C); // DOR_FDC_ENABLE | DOR_DMA_IO_INTERFACE_ENABLE
+ WRITE_PORT_UCHAR((PUCHAR)0x3F2, 0); // DOR_FDC_ENABLE | DOR_DMA_IO_INTERFACE_ENABLE 0x0C // we changed 0x0C->0 to workaround CORE-16469
}
#endif // defined __i386__ || defined(_M_AMD64)
https://git.reactos.org/?p=reactos.git;a=commitdiff;h=68ed29b41f3865388684b…
commit 68ed29b41f3865388684b0f36e3a10123c68de78
Author: Alexander Rechitskiy <rechitskiy(a)reactos.org>
AuthorDate: Wed May 6 14:14:41 2020 +0300
Commit: GitHub <noreply(a)github.com>
CommitDate: Wed May 6 14:14:41 2020 +0300
[README] "Building ReactOS" belongs to "Building" part (#2741)
- Improved logic of the text
- Removed some line breaks to improve readability on some screen sizes
---
README.md | 11 ++++-------
1 file changed, 4 insertions(+), 7 deletions(-)
diff --git a/README.md b/README.md
index 8ef9eb7868a..e6acf3d7bfc 100644
--- a/README.md
+++ b/README.md
@@ -55,16 +55,15 @@ Up-to-date versions for Windows and for Unix/GNU-Linux are available from our do
Alternatively one can use Microsoft Visual C++ (MSVC) version 2015+. Building with MSVC is covered here: ["Visual Studio or Microsoft Visual C++"](https://reactos.org/wiki/CMake#Visual_Studio_or_Microsoft_Visual_C.2B.2B).
+See ["Building ReactOS"](https://reactos.org/wiki/Building_ReactOS) article for more details.
+
### Binaries
To build ReactOS you must run the `configure` script in the directory you want to have your build files. Choose `configure.cmd` or `configure.sh` depending on your system. Then run `ninja <modulename>` to build a module you want or just `ninja` to build all modules.
### Bootable images
-To build a bootable CD image run `ninja bootcd` from the
-build directory. This will create a CD image with a filename `bootcd.iso`.
-
-See ["Building ReactOS"](https://reactos.org/wiki/Building_ReactOS) for more details.
+To build a bootable CD image run `ninja bootcd` from the build directory. This will create a CD image with a filename `bootcd.iso`.
You can always download fresh binary builds of bootable images from the ["Daily builds"](https://reactos.org/getbuilds/) page.
@@ -74,9 +73,7 @@ By default, ReactOS currently can only be installed on a machine that has a FAT1
The partition on which ReactOS is to be installed (which may or may not be the bootable partition) must also be formatted as FAT16 or FAT32.
ReactOS Setup can format the partitions if needed.
-Starting 0.4.10, ReactOS can be installed using the BtrFS file system. But
-consider this as an experimental feature and thus regressions not triggered on
-FAT setup may be observed.
+Starting with 0.4.10, ReactOS can be installed using the BtrFS file system. But consider this as an experimental feature and thus regressions not triggered on FAT setup may be observed.
To install ReactOS from the bootable CD distribution, extract the archive contents. Then burn the CD image, boot from it, and follow the instructions.
https://git.reactos.org/?p=reactos.git;a=commitdiff;h=0461de33c5de3fee70a00…
commit 0461de33c5de3fee70a004b2f9c9e03ca70cfc67
Author: Kyle Katarn <contact(a)kcsoftwares.com>
AuthorDate: Tue May 5 10:44:45 2020 +0200
Commit: GitHub <noreply(a)github.com>
CommitDate: Tue May 5 11:44:45 2020 +0300
[REGEDIT] Fix HeapFree() on the wrong variable (#2736)
- When exporting registry keys (to .reg files) some variables from the heap are not free'd while the debug log indicates "HEAP: Trying to free an invalid address".
- This is due to the export_registry_key() function that calls
HeapFree() for reg_key_name. But this variable is an argument provided by the caller, which is always a statically defined array of WCHAR.
- Meanwhile reg_key_name_buf is never free'd and cause a memory leak each time the function gets called.
---
base/applications/regedit/regproc.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/base/applications/regedit/regproc.c b/base/applications/regedit/regproc.c
index c17221df5ae..41ac7c6be8b 100644
--- a/base/applications/regedit/regproc.c
+++ b/base/applications/regedit/regproc.c
@@ -1392,7 +1392,7 @@ BOOL export_registry_key(WCHAR *file_name, WCHAR *reg_key_name, DWORD format)
if (file) {
fclose(file);
}
- HeapFree(GetProcessHeap(), 0, reg_key_name);
+ HeapFree(GetProcessHeap(), 0, reg_key_name_buf);
HeapFree(GetProcessHeap(), 0, val_name_buf);
HeapFree(GetProcessHeap(), 0, val_buf);
HeapFree(GetProcessHeap(), 0, line_buf);
https://git.reactos.org/?p=reactos.git;a=commitdiff;h=5dfe3455db5d8c21a3261…
commit 5dfe3455db5d8c21a326178006ae8dab9c71fc4b
Author: Kyle Katarn <contact(a)kcsoftwares.com>
AuthorDate: Mon May 4 22:30:36 2020 +0200
Commit: GitHub <noreply(a)github.com>
CommitDate: Mon May 4 23:30:36 2020 +0300
[ADVAPI32] Fix undue debug print in nominal case (#2734)
In current implementation, when regedit opens HKCR root key, an error is logged
err:(dll/win32/advapi32/reg/hkcr.c:964) Returning 259.
This is not correct as the code 259 is ERROR_NO_MORE_ITEMS which is the nominal return value when end of enumeration is reached.
---
dll/win32/advapi32/reg/hkcr.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/dll/win32/advapi32/reg/hkcr.c b/dll/win32/advapi32/reg/hkcr.c
index 2756685d5fb..7d714f48388 100644
--- a/dll/win32/advapi32/reg/hkcr.c
+++ b/dll/win32/advapi32/reg/hkcr.c
@@ -961,7 +961,8 @@ EnumHKCRValue(
if (ErrorCode != ERROR_SUCCESS)
{
/* Most likely ERROR_NO_MORE_ITEMS */
- ERR("Returning %d.\n", ErrorCode);
+ if (ErrorCode != ERROR_NO_MORE_ITEMS)
+ ERR("Returning %d.\n", ErrorCode);
goto Exit;
}
FallbackValueName[FallbackValueNameLen] = L'\0';