[PATCH] kernel32: Make GetFinalPathNameByHandle warn on mounted folder.

[Jeff Smith]

GetFinalPathNameByHandle attempts to translate a volume mount point to
an NT volume name using QueryDosDevice, but this can only succeed if the
mount point is a drive letter.  When GetVolumeNameForVolumeMountPoint
encounters a similiar situation, it emits a fixme message.

When attempting to translate a volume mount point to an NT volume name
with GetFinalPathNameByHandle, for a mount point that is not a drive
letter, emit a fixme message.

Signed-off-by: Jeff Smith <whydoubt(a)gmail.com>
---
 dlls/kernel32/file.c | 8 ++++++++
 1 file changed, 8 insertions(+)

diff --git a/dlls/kernel32/file.c b/dlls/kernel32/file.c
index 2b6a13fc0d..2d52b0a165 100644
--- a/dlls/kernel32/file.c
+++ b/dlls/kernel32/file.c
@@ -1442,6 +1442,14 @@ DWORD WINAPI GetFinalPathNameByHandleW(HANDLE file, LPWSTR path, DWORD charcount
     {
         WCHAR nt_prefix[MAX_PATH];
 
+        /* if length of drive_part is > 3 then it must be a mounted folder */
+        if (drive_part_len > 3)
+        {
+            FIXME("Mounted Folders are not yet supported\n");
+            SetLastError( ERROR_NOT_A_REPARSE_POINT );
+            return 0;
+        }
+
         /* QueryDosDeviceW sets error code on failure */
         drive_part[drive_part_len - 1] = 0;
         if (!QueryDosDeviceW( drive_part, nt_prefix, MAX_PATH ))
-- 
2.21.0

[Marvin]

Hi,

While running your changed tests, I think I found new failures.
Being a bot and all I'm not very good at pattern recognition, so I might be
wrong, but could you please double-check?

Full results can be found at:
https://testbot.winehq.org/JobDetails.pl?Key=56761

Your paranoid android.


=== debian10 (64 bit WoW report) ===

kernel32:
debugger.c:320: Test failed: GetThreadContext failed: 5

[Zhiyi Zhang]

Hi Jeff,

It would be better if you could add some tests for this.
For example, add a test that shows the failures on wine
and then next patch fix it and remove todo_wines.

And is this for winehq bug 47787?

Thanks,
Zhiyi

On 9/23/19 1:10 PM, Jeff Smith wrote:
> GetFinalPathNameByHandle attempts to translate a volume mount point to
> an NT volume name using QueryDosDevice, but this can only succeed if the
> mount point is a drive letter.  When GetVolumeNameForVolumeMountPoint
> encounters a similiar situation, it emits a fixme message.
>
> When attempting to translate a volume mount point to an NT volume name
> with GetFinalPathNameByHandle, for a mount point that is not a drive
> letter, emit a fixme message.
>
> Signed-off-by: Jeff Smith <whydoubt(a)gmail.com>
> ---
>  dlls/kernel32/file.c | 8 ++++++++
>  1 file changed, 8 insertions(+)
>
> diff --git a/dlls/kernel32/file.c b/dlls/kernel32/file.c
> index 2b6a13fc0d..2d52b0a165 100644
> --- a/dlls/kernel32/file.c
> +++ b/dlls/kernel32/file.c
> @@ -1442,6 +1442,14 @@ DWORD WINAPI GetFinalPathNameByHandleW(HANDLE file, LPWSTR path, DWORD charcount
>      {
>          WCHAR nt_prefix[MAX_PATH];
>
> +        /* if length of drive_part is > 3 then it must be a mounted folder */
> +        if (drive_part_len > 3)
> +        {
> +            FIXME("Mounted Folders are not yet supported\n");
> +            SetLastError( ERROR_NOT_A_REPARSE_POINT );
> +            return 0;
> +        }
> +
>          /* QueryDosDeviceW sets error code on failure */
>          drive_part[drive_part_len - 1] = 0;
>          if (!QueryDosDeviceW( drive_part, nt_prefix, MAX_PATH ))