Hermès,
why are you making such changes at all????
I do not see any benefit in this change apart from the obvious bug you
introduced.
Your change removes our ability to add debug code to the failure and
success code paths without making addition changes. The compiler will
most likely generate the same code for the old and the new code.
Result: No gain at all!
Eric
Am 13.05.2017 18:26, schrieb Thomas Faber:
> On 2017-05-13 18:13, hbelu...@svn.reactos.org wrote:
>> --- branches/setup_improvements/base/setup/usetup/filesup.c
>> [iso-8859-1] (original)
>> +++ branches/setup_improvements/base/setup/usetup/filesup.c
>> [iso-8859-1] Sat May 13 16:13:49 2017
>
>> @@ -524,14 +523,10 @@
>> &IoStatusBlock,
>> 0,
>> FILE_SYNCHRONOUS_IO_NONALERT);
>> - if (!NT_SUCCESS(Status))
>> - {
>> - return FALSE;
>> - }
>>
>> NtClose(FileHandle);
>>
>> - return TRUE;
>> + return NT_SUCCESS(Status);
>> }
>
> Not an improvement. You're now closing an uninitialized handle on
> failure.
>
>
> _______________________________________________
> Ros-dev mailing list
> Ros-dev@reactos.org
> http://www.reactos.org/mailman/listinfo/ros-dev
_______________________________________________
Ros-dev mailing list
Ros-dev@reactos.org
http://www.reactos.org/mailman/listinfo/ros-dev
