Re: #4 pdh: Implement and test PdhAddEnglishCounter{A, W} and PdhCollectQueryDataWithTime.

[Paul Vriens]

Hans Leidekker wrote:
> These APIs were introduced with Vista.
>
> -Hans
>
> Changelog
>   Implement and test PdhAddEnglishCounter{A, W} and PdhCollectQueryDataWithTime.

Hi Hans,

I was going through some test failures (already sent a patch for missing defines).

It appears that a lot of the used strings, like "System Up Time" is depending on 
the locale. Some Dutch tests show "Systeem ingeschakeld" instead.

Any idea how to fix that?

Another thing is that "pdh.dll" is not by default present on NT4 and below so 
the tests have to be changed to cope with that as well. Should I have a go at that?

Cheers,

Paul.

#4 pdh: Implement and test PdhAddEnglishCounter{A, W} and PdhCollectQueryDataWithTime.

[Hans Leidekker]

On Monday 30 July 2007 16:58:59 Paul Vriens wrote:
> It appears that a lot of the used strings, like "System Up Time" is depending on 
> the locale. Some Dutch tests show "Systeem ingeschakeld" instead.
>
> Any idea how to fix that?
>
> Another thing is that "pdh.dll" is not by default present on NT4 and below so 
> the tests have to be changed to cope with that as well. Should I have a go at that?

I have patches pending so I'll fix these issues in one go. In fact, I had fixed some
of the Vista failures already.

-Hans

#4 pdh: Implement and test PdhAddEnglishCounter{A, W} and PdhCollectQueryDataWithTime.

[Alexandre Julliard]

Paul Vriens  writes:
> Another thing is that "pdh.dll" is not by default present on NT4 and
> below so the tests have to be changed to cope with that as
> well. Should I have a go at that?

It may be a good idea to add generic support for that in
winetest. It's a bit silly to have to manually fix all the tests
that use new dlls.

#4 pdh: Implement and test PdhAddEnglishCounter{A, W} and PdhCollectQueryDataWithTime.

[Paul Vriens]

Alexandre Julliard wrote:
> Paul Vriens  writes:
> > Another thing is that "pdh.dll" is not by default present on NT4 and
> > below so the tests have to be changed to cope with that as
> > well. Should I have a go at that?
>
> It may be a good idea to add generic support for that in
> winetest. It's a bit silly to have to manually fix all the tests
> that use new dlls.

And how would you do that?

The first thing that winetest does is go through all the tests to retrieve all 
the subtests. If it finds a missing dll (or depending dll) we could/should catch 
that and not run the tests.

Is it as simple as that?

Cheers,

Paul.

#4 pdh: Implement and test PdhAddEnglishCounter{A, W} and PdhCollectQueryDataWithTime.

[Alexandre Julliard]

Paul Vriens  writes:
> And how would you do that?
>
> The first thing that winetest does is go through all the tests to
> retrieve all the subtests. If it finds a missing dll (or depending
> dll) we could/should catch that and not run the tests.
>
> Is it as simple as that?

Well, since the tests are named from the dll they are testing, I'd
suggest that before running the test app 'foo_test.exe' we try a
LoadLibrary("foo.dll"), and if that fails we report the whole test as
skipped.

#4 pdh: Implement and test PdhAddEnglishCounter{A, W} and PdhCollectQueryDataWithTime.

[Paul Vriens]

Alexandre Julliard wrote:
> Paul Vriens  writes:
> > And how would you do that?
> >
> > The first thing that winetest does is go through all the tests to
> > retrieve all the subtests. If it finds a missing dll (or depending
> > dll) we could/should catch that and not run the tests.
> >
> > Is it as simple as that?
>
> Well, since the tests are named from the dll they are testing, I'd
> suggest that before running the test app 'foo_test.exe' we try a
> LoadLibrary("foo.dll"), and if that fails we report the whole test as
> skipped.

I'll have a look.

How do yo want to report a missing test? In the normal report file? If that's 
the case we should also change the dissect/gather scripts:

'dissect' should cope with the new lines in the report files.
'gather' should make up some useful info in the generated html file. We only 
have information about subtests if one or more complete tests are run.

If this thing works out, should we change the old tests back or use this new 
approach only for new (and currently failing) dll's?

Cheers,

Paul.

#4 pdh: Implement and test PdhAddEnglishCounter{A, W} and PdhCollectQueryDataWithTime.

[Alexandre Julliard]

Paul Vriens  writes:
> I'll have a look.
>
> How do yo want to report a missing test? In the normal report file? If
> that's the case we should also change the dissect/gather scripts:

It has to be part of the report data, so yes the scripts probably need
to be updated.
> If this thing works out, should we change the old tests back or use
> this new approach only for new (and currently failing) dll's?

If it simplifies the code we can certainly change the old tests too.

#4 pdh: Implement and test PdhAddEnglishCounter{A, W} and PdhCollectQueryDataWithTime.

[Paul Vriens]

Alexandre Julliard wrote:
> Paul Vriens  writes:
> > I'll have a look.
> >
> > How do yo want to report a missing test? In the normal report file? If
> > that's the case we should also change the dissect/gather scripts:
>
> It has to be part of the report data, so yes the scripts probably need
> to be updated.
> > If this thing works out, should we change the old tests back or use
> > this new approach only for new (and currently failing) dll's?
>
> If it simplifies the code we can certainly change the old tests too.

Hi,

a patch somewhat like this?

diff --git a/programs/winetest/main.c b/programs/winetest/main.c
index f743ffc..3c4846d 100644
--- a/programs/winetest/main.c
+++ b/programs/winetest/main.c
@@ -433,6 +433,24 @@ extract_test_proc (HMODULE hModule, LPCTSTR lpszType,
                     LPTSTR lpszName, LONG_PTR lParam)
  {
      const char *tempdir = (const char *)lParam;
+    char dllname[MAX_PATH];
+    HMODULE dll;
+
+    /* Check if the main dll is present on this system */
+    strcpy(dllname, lpszName);
+    *strstr(dllname, "_TEST.EXE") = 0;
+
+    dll = LoadLibrary(dllname);
+    if (!dll) {
+        /* Here we add some information to the results file
+         * that indicates that the dll is not present
+         * on this platform.
+         */
+        report (R_WARNING, "dll : %s is not present", dllname);
+        return TRUE;
+    }
+    FreeLibrary(dll);
+
      get_subtests( tempdir, &wine_tests[nr_of_files], lpszName );
      nr_of_tests += wine_tests[nr_of_files].subtest_count;
      nr_of_files++;

Running this on win9x gives me 13 dll's that are not available (the R_WARNING is 
of course only for testing).

Cheers,

Paul.

#4 pdh: Implement and test PdhAddEnglishCounter{A, W} and PdhCollectQueryDataWithTime.

[Alexandre Julliard]

Paul Vriens  writes:
> a patch somewhat like this?

Yes; it may be better to do that in extract_test() since we get the
test name there, it avoids hardcoding _test.exe at two different
places. But the general principle is fine.

#4 pdh: Implement and test PdhAddEnglishCounter{A, W} and PdhCollectQueryDataWithTime.

[Paul Vriens]

Alexandre Julliard wrote:
> Paul Vriens  writes:
> > a patch somewhat like this?
>
> Yes; it may be better to do that in extract_test() since we get the
> test name there, it avoids hardcoding _test.exe at two different
> places. But the general principle is fine.

It doesn't make the code simpler though. When changing extract_test I also have 
to make sure that the tests are not run (--list) in get_subtests. I still have 
to get rid of the "_test.exe" in extract_test. I have to fiddle with 
test->exename and test->name etc...

Couldn't I just add a 'static const char testexec[] = "_test.exe" at the top?

Cheers,

Paul.

#4 pdh: Implement and test PdhAddEnglishCounter{A, W} and PdhCollectQueryDataWithTime.

[Paul Vriens]

Paul Vriens wrote:
> Alexandre Julliard wrote:
> > Paul Vriens  writes:
> > > a patch somewhat like this?
> >
> > Yes; it may be better to do that in extract_test() since we get the
> > test name there, it avoids hardcoding _test.exe at two different
> > places. But the general principle is fine.
>
> It doesn't make the code simpler though. When changing extract_test I 
> also have to make sure that the tests are not run (--list) in 
> get_subtests. I still have to get rid of the "_test.exe" in 
> extract_test. I have to fiddle with test->exename and test->name etc...
>
> Couldn't I just add a 'static const char testexec[] = "_test.exe" at the 
> top?
>
> Cheers,
>
> Paul.

Hi again,

I have got it somewhat working (needs more testing). The attached screenshot
shows the output from a win98 system (I only had d3d8 as missing in the report
file for testing purposes). When you hover over the test, you will see the text
as shown in the screenshot.

Cheers,

Paul.