Re: [PATCH 5/6]: msvcp110: Add tr2_sys__Lstat implementation and test.(resend)

[Piotr Caban]

On 07/03/15 06:00, YongHaoHu wrote:
> +/* ?_Lstat(a)sys@tr2(a)std@@YA?AW4file_type(a)123@PBDAAH(a)Z */
> +/* ?_Lstat(a)sys@tr2(a)std@@YA?AW4file_type(a)123@PEBDAEAH(a)Z */
> +file_type tr2_sys__Lstat(char const* path, int *err_code)
> +{
> +    DWORD attr;
> +    TRACE("(%s %p)\n", debugstr_a(path), err_code);
> +    if(!path) {
> +        *err_code = ERROR_INVALID_PARAMETER;
> +        return status_unknown;
> +    }
> +
> +    attr = GetFileAttributesA(path);
> +    if(attr == INVALID_FILE_ATTRIBUTES)
> +        return stat_set_error(err_code);
> +    if(attr & FILE_ATTRIBUTE_REPARSE_POINT)
> +        FIXME("reparse point is not supported yet\n");
> +
> +    *err_code = ERROR_SUCCESS;
> +    return (attr & FILE_ATTRIBUTE_DIRECTORY)?directory_file:regular_file;
> +}
> What's the reason for not calling tr2_sys__Stat function here?

Thanks,
Piotr

[PATCH 5/6]: msvcp110: Add tr2_sys__Lstat implementation and test.(resend)

[YongHao Hu]

Hi.

On 15/7/3 下午6:51, Piotr Caban wrote:
> On 07/03/15 06:00, YongHaoHu wrote:
> > +/* ?_Lstat(a)sys@tr2(a)std@@YA?AW4file_type(a)123@PBDAAH(a)Z */
> > +/* ?_Lstat(a)sys@tr2(a)std@@YA?AW4file_type(a)123@PEBDAEAH(a)Z */
> > +file_type tr2_sys__Lstat(char const* path, int *err_code)
> > +{
> > +    DWORD attr;
> > +    TRACE("(%s %p)\n", debugstr_a(path), err_code);
> > +    if(!path) {
> > +        *err_code = ERROR_INVALID_PARAMETER;
> > +        return status_unknown;
> > +    }
> > +
> > +    attr = GetFileAttributesA(path);
> > +    if(attr == INVALID_FILE_ATTRIBUTES)
> > +        return stat_set_error(err_code);
> > +    if(attr & FILE_ATTRIBUTE_REPARSE_POINT)
> > +        FIXME("reparse point is not supported yet\n");
> > +
> > +    *err_code = ERROR_SUCCESS;
> > +    return (attr & 
> > FILE_ATTRIBUTE_DIRECTORY)?directory_file:regular_file;
> > +}
> > What's the reason for not calling tr2_sys__Stat function here?

The handling of REPARSE_POINT in tr2_sys__Stat  is different from 
tr2_sys__Lstat, so we need to implement it again.
That's why I want to reuse my function stat_set_error.
Regarding ERROR_SUCCESS, I handle it wrong.
Thank you. ;)

[PATCH 5/6]: msvcp110: Add tr2_sys__Lstat implementation and test.(resend)

[Piotr Caban]

On 07/03/15 17:29, YongHao Hu wrote:
> > What's the reason for not calling tr2_sys__Stat function here?
>
> The handling of REPARSE_POINT in tr2_sys__Stat  is different from
> tr2_sys__Lstat, so we need to implement it again.
> That's why I want to reuse my function stat_set_error.
> Could you please describe the difference?

According to my testing reparse points have
FILE_ATTRIBUTE_DIRECTORY attribute if it's a directory. I don't see 
where _Lstat behaves differently than _Stat and why reparse points needs 
any special handling?

[PATCH 5/6]: msvcp110: Add tr2_sys__Lstat implementation and test.(resend)

[YongHao Hu]

On 15/7/3 下午11:50, Piotr Caban wrote:
> On 07/03/15 17:29, YongHao Hu wrote:
> > > What's the reason for not calling tr2_sys__Stat function here?
> >
> > The handling of REPARSE_POINT in tr2_sys__Stat  is different from
> > tr2_sys__Lstat, so we need to implement it again.
> > That's why I want to reuse my function stat_set_error.
> > Could you please describe the difference?
> > Sry, it seems that I made a mistake.
> > I will add more tests in the patch.
> > Thank you. ^^