On 9/19/06, David Golden <[EMAIL PROTECTED]> wrote:
Oleg V. Volkov wrote:
> Greetings.
>
> One day when I was tired from being (almost) unable to work with Unicode
> file names on Win32, I wrote a module with wrappers around Win32 API
> function that uses Win32::API to actually import them from .dlls. I did not
> use Win32API::File because it doesn't provide convient way to use Perl
> unicode strings as arguments, nor does it check for some errors that I feel
> should be critical, nor does it parse structures in output and, finally, it
> simply does not implement several function I needed, like FindFirstFileW,
> for example. So, right now my module have some *W functions implemented,
> namely: MoveFileW CopyFileW FindFirstFileW FindNextFileW FindClose. Glob
> function based on last three *W win32api_glob (which probably will be moved
> in future to sub-module ::Glob and made do be as close as posible to
> File::Glob functions), some helper functions that pack/unpack structures
> based on C-struct defenitions parsed verbatim from MSDN and I plan to
> implement other *W functions, especially those not covered by Win32API::File
> at all. Only important limitation is that I will only use *W functions,
> unless function is neutral, like FindClose.
>
> Right now I think Win32::FileW would be a good name for this module. Is
> there any problems with such name or suggestions of better one?
Oleg, thanks for tackling such a tricky subject. Would you consider
patching Win32API::File instead of releasing a new module? Or working
with the current maintainer, demerphq, on improvements? I think that
would be a better approach.
I agree. Id also like to see more detail about the perceived
weaknesses in Win32API::File. Also, id say that doing direct bindings
against the Win API is superior to using Win32::API, although im
probably a touch biased.
Yves
--
perl -Mre=debug -e "/just|another|perl|hacker/"
