[PATCH] advapi32/tests: Fix an uninitialized variable check in test_performance_keys().

[Francois Gouget]

Signed-off-by: Francois Gouget 
---
This should fix the failures on cw-gtx560-wow64, cw-rx460-wow64
and fg-deb64-wow64:
https://test.winehq.org/data/patterns.html#advapi32:registry
---
 dlls/advapi32/tests/registry.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/dlls/advapi32/tests/registry.c b/dlls/advapi32/tests/registry.c
index 5fac9aa429f..a6e48d7840f 100644
--- a/dlls/advapi32/tests/registry.c
+++ b/dlls/advapi32/tests/registry.c
@@ -3867,6 +3867,7 @@ static void test_performance_keys(void)
         ret = RegSetValueA(keys[i], "Global", REG_SZ, "dummy", 5);
         ok(ret == ERROR_INVALID_HANDLE, "got %u\n", ret);

+        key_count = 0x900ddeed;
         ret = RegQueryInfoKeyA(keys[i], NULL, NULL, NULL, &key_count, NULL,
                 NULL, &value_count, NULL, NULL, NULL, NULL);
         todo_wine ok(!ret, "got %u\n", ret);

[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=95678

Your paranoid android.

=== w10pro64_zh_CN (64 bit report) ===

advapi32:
registry.c:568: Test failed: data_count set to 9 instead of 7
registry.c:572: Test failed: data set to 'xxxxxxxxx' instead of 'foobar' or 'xxxxxxx'

[Francois Gouget]

On Mon, 16 Aug 2021, Marvin wrote:
[...]
> === w10pro64_zh_CN (64 bit report) ===
>
> advapi32:
> registry.c:568: Test failed: data_count set to 9 instead of 7
> registry.c:572: Test failed: data set to 'xxxxxxxxx' instead of 'foobar' or 'xxxxxxx'

This is one of the other random failures we get with advapi32:registry. 
The annoying part is that while here we get 9, other times it was either 
13 or 10 and whenever the value changes it looks like a new failure to 
the TestBot.

There are also several lines where it can happen so it combines to 
produce several different failure modes.

So if someone wants to look into it that will certainly avoid a bumch of 
false positives.

https://test.winehq.org/data/patterns.html#advapi32:registry