Re: USER32: Error checking for sysparams tests

[Paul Vriens]

On Sat, 2005-03-05 at 06:42, [e-mail] wrote:
> As discussed on wine-devel there are a handful of calls in the 
> SystemParametersInfo tests that don't check for returned errors and continue 
> on with testing a given parameter. This patch ensures that all initial calls 
> to SystemParametersInfoA in each parameter test have their returned errors 
> checked. To ease adding more error cases and reduce code duplication I have 
> consolidated error message handling into a single function.
>
> Changelog:
> - Consolidate error message handling
> - Add error checking to all initial SystemParametersInfoA calls in each
>
> parameter test to make sure that if a given parameter isn't supported or has 
> failed in some other way that we return and don't continue on with the test
> - Remove a couple of TABS and align registry keys/values at the top of the
>
> file

Hi Justin,

what about the attached one? Added/changed the following:

- Always set LastError before calling test_error_msg
- SPI_{GET,SET}ICONTITLEWRAP is supported on Win98, but was clouded by
an error that was set after a previous call failed.

Changelog:
Justin Chevrier

* Consolidate error message handling
* Add error checking to all initial SystemParametersInfoA calls in each 
parameter test to make sure that if a given parameter isn't supported or
has failed in some other way that we return and don't continue on with
the test
* SetLastError before calling test_error_msg
* Remove a couple of TABS and align registry keys/values at the top of
the file
* SPI_{GET,SET}ICONTITLEWRAP is supported on Win98

Cheers,

Paul.

USER32: Error checking for sysparams tests

On Saturday 05 March 2005 08:58, Paul Vriens wrote:
> On Sat, 2005-03-05 at 06:42, [e-mail] wrote:
> > As discussed on wine-devel there are a handful of calls in the
> > SystemParametersInfo tests that don't check for returned errors and
> > continue on with testing a given parameter. This patch ensures that all
> > initial calls to SystemParametersInfoA in each parameter test have their
> > returned errors checked. To ease adding more error cases and reduce code
> > duplication I have consolidated error message handling into a single
> > function.
> >
> > Changelog:
> > - Consolidate error message handling
> > - Add error checking to all initial SystemParametersInfoA calls in each
> >
> > parameter test to make sure that if a given parameter isn't supported or
> > has failed in some other way that we return and don't continue on with
> > the test * Remove a couple of TABS and align registry keys/values at the
> > top of the file
>
> Hi Justin,
>
> what about the attached one? Added/changed the following:
> - Always set LastError before calling test_error_msg
> - SPI_{GET,SET}ICONTITLEWRAP is supported on Win98, but was clouded by
>
> an error that was set after a previous call failed.
>
> Changelog:
> Justin Chevrier
> - Consolidate error message handling
> - Add error checking to all initial SystemParametersInfoA calls in each
>
> parameter test to make sure that if a given parameter isn't supported or
> has failed in some other way that we return and don't continue on with
> the test
> - SetLastError before calling test_error_msg
> - Remove a couple of TABS and align registry keys/values at the top of
>
> the file
> - SPI_{GET,SET}ICONTITLEWRAP is supported on Win98
>
> Cheers,
>
> Paul.

Looks good (outside of the spelling correction that I believe has already been 
applied). I can confirm that Win 98 isn't setting the extended error 
information on at least some calls (maybe all?) to SystemParametersInfoA. I 
can get WinXP/NT/2K to do the same thing by removing the rights of the user 
running the test to the keys in the registry that the parameter calls attempt 
to access ;). No extended error information there either.

Justin