Re: [PATCH 2/3] shell32/tests: Remove now useless forced 0x0501 API versions.

[Francois Gouget]

On Sat, 14 Mar 2020, Serge Gautherie wrote:
> I did test on WTB:
> these failures happen even when removing the comment only, so not my patch!

The problem is that these failures contain a temporary path which is
unique to each run, as it's meant to. Just for your run we have five 
different values:

w1064v1809       wtEFCE.tmp
  w1064v1809_2scr  wtFA3E.tmp
  w1064v1809_ja    wt1690.tmp
  w1064v1809_zh_CN wt1A78.tmp
  w1064v1809       wtEDBB.tmp

This makes it impossible for the TestBot to know whether the latest 
failure matches one from a previous WineTest run so they are all 
reported as new.

To avoid that I sent a patch to omit te temporary directory from the 
failure messages.
> Shouldn't they be flagged as permanent failures?

Ideally the test author or anyone working on it would fix the 
failures. Just saying...

[PATCH 2/3] shell32/tests: Remove now useless forced 0x0501 API versions.

[Serge Gautherie]

On 15/03/2020 14:38, Francois Gouget wrote:
> On Sat, 14 Mar 2020, Serge Gautherie wrote:
> > I did test on WTB:
> > these failures happen even when removing the comment only, so not my patch!
>
> The problem is that these failures contain a temporary path which is
> unique to each run, as it's meant to. Just for your run we have five
> different values:
>
> w1064v1809       wtEFCE.tmp
>    w1064v1809_2scr  wtFA3E.tmp
>    w1064v1809_ja    wt1690.tmp
>    w1064v1809_zh_CN wt1A78.tmp
>    w1064v1809       wtEDBB.tmp
>
> This makes it impossible for the TestBot to know whether the latest
> failure matches one from a previous WineTest run so they are all
> reported as new.

Ah, right.
> To avoid that I sent a patch to omit te temporary directory from the
> failure messages.

Thanks.
> > Shouldn't they be flagged as permanent failures?
>
> Ideally the test author or anyone working on it would fix the
> failures. Just saying...

Agreed, though not me and now, sorry: too unrelated to what I am working on.