[PATCH] crypt32/tests: Use GetWindowsDirectoryA instead of GetEnvironmentVariableA to avoid a test failure.

[Sven Baars]

Signed-off-by: Sven Baars 
---
 dlls/crypt32/tests/sip.c | 9 +++------
 1 file changed, 3 insertions(+), 6 deletions(-)

diff --git a/dlls/crypt32/tests/sip.c b/dlls/crypt32/tests/sip.c
index d8d6a87445..f0849219f5 100644
--- a/dlls/crypt32/tests/sip.c
+++ b/dlls/crypt32/tests/sip.c
@@ -145,8 +145,6 @@ static void test_SIPRetrieveSubjectGUID(void)
     BOOL ret;
     GUID subject;
     HANDLE file;
-    static const CHAR windir[] = "windir";
-    static const CHAR regeditExe[] = "regedit.exe";
     static const GUID nullSubject  = { 0x0, 0x0, 0x0, { 0x0,0x0,0x0,0x0,0x0,0x0,0x0,0x0 }};
     static const WCHAR deadbeef[]  = { 'c',':','\','d','e','a','d','b','e','e','f','.','d','b','f',0 };
     /* Couldn't find a name for this GUID, it's the one used for 95% of the files */
@@ -183,10 +181,9 @@ static void test_SIPRetrieveSubjectGUID(void)
      *
      * Use A-functions where possible as that should be available on all platforms
      */
-    ret = GetEnvironmentVariableA(windir, regeditPath, MAX_PATH);
-    ok (ret > 0, "expected GEVA(windir) to succeed, last error %d\n", GetLastError());
-    strcat(regeditPath, "\");
-    strcat(regeditPath, regeditExe);
+    ret = GetWindowsDirectoryA(regeditPath, MAX_PATH-12);
+    ok (ret > 0, "Expected GetWindowsDirectoryA to succeed, last error %d\n", GetLastError());
+    strcat(regeditPath, "\regedit.exe");
     MultiByteToWideChar(CP_ACP, 0, regeditPath, strlen(regeditPath)+1, regeditPathW,
                         ARRAY_SIZE(regeditPathW));

[Zebediah Figura]

On 11/8/19 3:54 PM, Sven Baars wrote:
> Signed-off-by: Sven Baars
>
> dlls/crypt32/tests/sip.c | 9 +++------
>   1 file changed, 3 insertions(+), 6 deletions(-)

As far as I see this only happens on Wine, and only on one machine. So 
while this code is simpler and probably more desirable anyway, "to avoid 
a test failure" is probably not a good reason for it.
> diff --git a/dlls/crypt32/tests/sip.c b/dlls/crypt32/tests/sip.c
> index d8d6a87445..f0849219f5 100644
> --- a/dlls/crypt32/tests/sip.c
> +++ b/dlls/crypt32/tests/sip.c
> @@ -145,8 +145,6 @@ static void test_SIPRetrieveSubjectGUID(void)
>       BOOL ret;
>       GUID subject;
>       HANDLE file;
> - static const CHAR windir[] = "windir";
> - static const CHAR regeditExe[] = "regedit.exe";
>  static const GUID nullSubject  = { 0x0, 0x0, 0x0, { 0x0,0x0,0x0,0x0,0x0,0x0,0x0,0x0 }};
>  static const WCHAR deadbeef[]  = { 'c',':','\','d','e','a','d','b','e','e','f','.','d','b','f',0 };
>  /* Couldn't find a name for this GUID, it's the one used for 95% of the files */
>
> @@ -183,10 +181,9 @@ static void test_SIPRetrieveSubjectGUID(void)
>        *
>        * Use A-functions where possible as that should be available on all platforms
>        */
> - ret = GetEnvironmentVariableA(windir, regeditPath, MAX_PATH);
> - ok (ret > 0, "expected GEVA(windir) to succeed, last error %d\n", GetLastError());
> - strcat(regeditPath, "\");
> - strcat(regeditPath, regeditExe);
> - ret = GetWindowsDirectoryA(regeditPath, MAX_PATH-12);
> - ok (ret > 0, "Expected GetWindowsDirectoryA to succeed, last error %d\n", GetLastError());
> - strcat(regeditPath, "\regedit.exe");
>  MultiByteToWideChar(CP_ACP, 0, regeditPath, strlen(regeditPath)+1, regeditPathW,
>                      ARRAY_SIZE(regeditPathW));

[Sven Baars]

On 08-11-19 22:59, Zebediah Figura wrote:
> On 11/8/19 3:54 PM, Sven Baars wrote:
> > Signed-off-by: Sven Baars
> >
> > dlls/crypt32/tests/sip.c | 9 +++------
> >   1 file changed, 3 insertions(+), 6 deletions(-)
>
> As far as I see this only happens on Wine, and only on one machine. So
> while this code is simpler and probably more desirable anyway, "to avoid
> a test failure" is probably not a good reason for it.

That's true. Maybe we should just wait for Francois to fix that machine
instead, since it has many unique failures. I was just clicking on
random test failures and not really thinking anymore since it's already
quite late... Sorry for the noise.

[Francois Gouget]

On Fri, 8 Nov 2019, Sven Baars wrote:
[...]
> That's true. Maybe we should just wait for Francois to fix that machine
> instead, since it has many unique failures. I was just clicking on
> random test failures and not really thinking anymore since it's already
> quite late... Sorry for the noise.

I'm not sure when I'll have time to dig into this so I disabled the 
WineTest runs on Wine on that VM to avoid further confusion.

https://bugs.winehq.org/show_bug.cgi?id=48092
https://bugs.winehq.org/show_bug.cgi?id=48093