Responder: Re: [PATCH 1/2] tools/make_unicode: Implement full Unicode character decomposition.

sdelreal at codeweavers.com sdelreal at codeweavers.com
Wed Apr 11 08:30:19 CDT 2018




---- Huw Davies escribió ----

> On Tue, Apr 10, 2018 at 04:25:54PM -0500, Sergio Gómez Del Real wrote:
> > Signed-off-by: Sergio Gómez Del Real <sdelreal at codeweavers.com>
> > ---
> >  libs/port/mbtowc.c |  14 +-
> >  tools/make_unicode | 423 +++++++++++++++++++++++++++++++++++++++++++++--------
> >  2 files changed, 370 insertions(+), 67 deletions(-)
> 
> What this patch is doing is to add a 'compat' flag to wine_decompose()
> and then always call that function with that flag set.  So essentially
> you've added a bunch of dead code (the compat not set case).
> 
> The patch needs to make some material difference to the Wine's
> implementation of the Win32 api, preferably demonstrated by a test; it
> currently does not do that.
> 
> Similar comments could be made about patch [2/2].
> 
> Huw.

Yes, it effectively replaces wine_decompose with wine_unicode_decompose_string adding, among other things, the flag that tells it to use canonical or compatibility decomposition.

In these patches, only the existing code in mbtowc.c use it; it already was using wine_unicode. Because it uses only canonical decomposition, there is no case in these patches for compatibility.

The thing is that the third patch I have, which implements NormalizeString (and make full use of the mentioned flag) wouldn't work in Wine without first running the make_unicode script. I could send the whole patchset with this in mind.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.winehq.org/pipermail/wine-devel/attachments/20180411/c8704cbe/attachment.html>


More information about the wine-devel mailing list