Re: Patch: New version of "UTF-16 to 'Win32 locale' conversions"and filenames (replacing convertion tables with Win32 API calls)

[João Garcia <jgarcia at uk2 dot net>]

Hi Mohan,

Mohan Embar wrote:

> If you post a patch for 3.4 and a testcase, I will put out a binary
> on my website which incorporates this patch, even if the patch might
> not make it to the mainline. 
I'll do that (using only A functions -- less things to modify and less 
things to test).
Thanks!

> this will help a lot of MinGW people in
> the meantime.
I agree.

> I'd be interested in initially exploring a solution which invokes the
> W file functions on WinNT and the A file functions on Win9x.
I have a draft for something like that (I didn't finish debuging W calls 
sections -- I can recall a segmentation fault somewhere). As the patch 
is, it chooses between A and W functions "on-the-fly", based on platform 
detection. This leeds to code duplication...
I think that you would like to do the choice as a build option (or, at 
least, I might be happy with this option...). Or, probably, support only 
W functions (I don't blame you for that and I can always build my own 
binaries).

> Do you want to do this investigation?
I can do it (time permitting). For me, it would be just to adapt the old 
patch (and debug it).

> If not, would you have any objections
> to my looking into it (time permitting)?
>  
I have no objections if you (or anyone else) would like to finish this 
first.
But we would both loose time.
I think I can propose something before the end of  September.
Would you wish to do it yourself sooner?

> Thanks for working on this.
>  
You are welcome! But this is a small contribution... I am glad if it can 
be of any use at all!
From my part, many thanks to you for all your contributions to this 
project. And many thanks to all the people that make the gcc/gcj project 
work. It is a magnificent project and it is getting better and better 
all the time!

João