Fwd: [PATCH v4] winetest: Add GUI to display results of individual tests.

[Isira Seneviratne]

---------- Forwarded message ---------
From: Isira Seneviratne 
Date: Mon, Feb 18, 2019 at 5:53 PM
Subject: Re: [PATCH v4] winetest: Add GUI to display results of individual
tests.
To: Zebediah Figura

On Mon, Feb 18, 2019 at 11:14 AM Isira Seneviratne 
wrote:
> On Mon, Feb 18, 2019 at 11:02 AM Zebediah Figura 
> wrote:
> > On 2/17/19 11:30 PM, Isira Seneviratne wrote:
> > > On Mon, Feb 18, 2019 at 10:50 AM Zebediah Figura <[e-mail]
> > > > wrote:
> > > On 2/17/19 11:17 PM, Isira Seneviratne wrote:
> > >  >
> > >  > On Mon, Feb 18, 2019 at 9:08 AM Zebediah Figura
> > > <[e-mail] <mailto:[e-mail]>
> > >  > <mailto:[e-mail] <mailto:[e-mail]>>>
> >
> > wrote:
> > >  >
> > >  >
> > >  >
> > >  >     On 2/13/19 2:57 AM, Isira Seneviratne wrote:
> > >  >      >
> > >  >      >
> > >  >      >
> > >  >
> > >  >     Again, I'm seeing multiple lines show up per test run,
> > >  >
> > >  > Do the tests normally print multiple sets of results? If so, then
> > > should
> > >  > any of them be ignored?
> > >  >
> > >  >
> > >
> > > No, only one set of results is printed per test. You can check this
> >
> > by
> > > running winetest normally with the -o option, and examining what
> > > data is
> > > printed.
> > >
> > >
> > > In that case, I think that something is probably wrong with the
> >
> > winetest
> > > output, since I'm reading data from the same file that winetest is
> > > writing data to, with these lines:
> > >
> > > SetFilePointer (out_file, -sizeof(buffer), NULL, FILE_CURRENT);
> > > ReadFile (out_file, buffer, sizeof(buffer), &read_count, NULL);
> >
> > Something is probably wrong with the way you're reading from it.
>
> Okay, I'll check my code. Thank you for the feedback.
>
> I'll get back to you in a few hours.

I advanced the file pointer by 1 after reading 1000 characters from it, and
that eliminated most of the duplicates, except for "advapi32:security". I
isolated the following two lines of interest from the output:

00d4:security: 23 tests executed (0 marked as todo, 0 failures), 0 skipped.
00cd:security: 2696 tests executed (175 marked as todo, 0 failures), 28
skipped.

[Gabriel Ivăncescu]

On 2/18/19 2:24 PM, Isira Seneviratne wrote:
> ---------- Forwarded message ---------
> From: *Isira Seneviratne* <[e-mail] 
> >
> Date: Mon, Feb 18, 2019 at 5:53 PM
> Subject: Re: [PATCH v4] winetest: Add GUI to display results of 
> individual tests.
> To: Zebediah Figura <[e-mail] >
>
> On Mon, Feb 18, 2019 at 11:14 AM Isira Seneviratne <[e-mail] 
> > wrote:
> On Mon, Feb 18, 2019 at 11:02 AM Zebediah Figura
> <[e-mail] <mailto:[e-mail]>> wrote:
>
>     On 2/17/19 11:30 PM, Isira Seneviratne wrote:
>      >
>      >
>      > On Mon, Feb 18, 2019 at 10:50 AM Zebediah Figura
>     <[e-mail] <mailto:[e-mail]>
>      > <mailto:[e-mail] <mailto:[e-mail]>>>
>     wrote:
>      >
>      >     On 2/17/19 11:17 PM, Isira Seneviratne wrote:
>      >      >
>      >      > On Mon, Feb 18, 2019 at 9:08 AM Zebediah Figura
>      >     <[e-mail] <mailto:[e-mail]>
>     <mailto:[e-mail] <mailto:[e-mail]>>
>      >      > <mailto:[e-mail]
>     <mailto:[e-mail]> <mailto:[e-mail]
>     <mailto:[e-mail]>>>> wrote:
>      >      >
>      >      >
>      >      >
>      >      >     On 2/13/19 2:57 AM, Isira Seneviratne wrote:
>      >      >      >
>      >      >      >
>      >      >      >
>      >      >
>      >      >     Again, I'm seeing multiple lines show up per test run,
>      >      >
>      >      > Do the tests normally print multiple sets of results?
>     If so, then
>      >     should
>      >      > any of them be ignored?
>      >      >
>      >      >
>      >
>      >     No, only one set of results is printed per test. You can
>     check this by
>      >     running winetest normally with the -o option, and
>     examining what
>      >     data is
>      >     printed.
>      >
>      >
>      > In that case, I think that something is probably wrong with
>     the winetest
>      > output, since I'm reading data from the same file that
>     winetest is
>      > writing data to, with these lines:
>      >
>      > SetFilePointer (out_file, -sizeof(buffer), NULL, FILE_CURRENT);
>      > ReadFile (out_file, buffer, sizeof(buffer), &read_count, NULL);
>      >
>      >
>
>     Something is probably wrong with the way you're reading from it.
>
>
> Okay, I'll check my code. Thank you for the feedback.
>
> I'll get back to you in a few hours.
>
>
> I advanced the file pointer by 1 after reading 1000 characters from it, 
> and that eliminated most of the duplicates, except for 
> "advapi32:security". I isolated the following two lines of interest from 
> the output:
>
> 00d4:security: 23 tests executed (0 marked as todo, 0 failures), 0 skipped.
> 00cd:security: 2696 tests executed (175 marked as todo, 0 failures), 28 
> skipped.

That's because it uses multiple threads (see 00d4 and 00cd at the 
leftmost column). Other tests also use multiple threads, but in those 
cases some other threads finish earlier so they are mixed with other 
traces or test failures and are not found at the end of the log.

Why don't you just scan the entire log and look for lines matching the 
above? 1000 sounds like a totally arbitrary number anyway. You shouldn't 
need to seek back into the file, unless I am missing something.