kernel32: Add UTF-7 support

[Alex Henrie]

Hi all,

Could I get some feedback on
http://source.winehq.org/patches/data/85898 ? It was marked "Rejected"
a week ago without explanation.

Most of the code was authored by Katayama Hirofumi, whose patches have
been accepted into Wine before. He gave permission to use his code
under the LGPL at http://bugs.winehq.org/show_bug.cgi?id=27388

What more needs to be done before this patch can be accepted?

-Alex

[Ben Klein]

On 8 May 2012 11:21, Alex Henrie  wrote:
> Hi all,
>
> Could I get some feedback on
> http://source.winehq.org/patches/data/85898 ? It was marked "Rejected"
> a week ago without explanation.

Without being a regular contributor, here are a few tips from a cursory glance:
1) Whitespace change on at least one line
2) Why the change from ERROR_INVALID_PARAMETER to ERROR_INVALID_FLAGS?
3) New function defs do not match format of surrounding defs in unicode.h
4) No new tests to confirm the behaviour is correct (or mention of
current tests succeeding)
> Most of the code was authored by Katayama Hirofumi, whose patches have
> been accepted into Wine before. He gave permission to use his code
> under the LGPL at http://bugs.winehq.org/show_bug.cgi?id=27388
>
> What more needs to be done before this patch can be accepted?
>
> -Alex

[Alexandre Julliard]

Alex Henrie  writes:
> Hi all,
>
> Could I get some feedback on
> http://source.winehq.org/patches/data/85898 ? It was marked "Rejected"
> a week ago without explanation.
>
> Most of the code was authored by Katayama Hirofumi, whose patches have
> been accepted into Wine before. He gave permission to use his code
> under the LGPL at http://bugs.winehq.org/show_bug.cgi?id=27388
>
> What more needs to be done before this patch can be accepted?

That code is hopelessly ugly, and broken in various ways. You are
probably better off starting from scratch.

[Alex Henrie]

2012/5/8 Alexandre Julliard :
> That code is hopelessly ugly, and broken in various ways. You are
> probably better off starting from scratch.

Okay. Could we use the POSIX iconv functions? Why doesn't Wine use
iconv for UTF-8 already?

-Alex

[Alexandre Julliard]

Alex Henrie  writes:
> 2012/5/8 Alexandre Julliard :
> > That code is hopelessly ugly, and broken in various ways. You are
> > probably better off starting from scratch.
>
> Okay. Could we use the POSIX iconv functions? Why doesn't Wine use
> iconv for UTF-8 already?

UTF-8 is trivial, using iconv would be more work than doing it by
hand. UTF-7 is slightly more complicated but not that hard.