Re: UTF-8 Cygwin

[SUZUKI Hisao <suzuki611 at oki dot com>]

Dave Korn wrote:
> On 06 July 2006 07:28, SUZUKI Hisao wrote:
>
>
> > Sorry, but I cannot access to CVS server because of firewall.  So the
> > patch file was made from the cygwin-1.5.20-1-src.
>
>   Here it is, blindly applied to current CVS and then regenerated.  I've
> checked that it still builds and I've installed and started running with it.
> I'll report any quirks I find, but I'm not likely to be using the UTF-8
> features; I'll just look out for any possible breakage of existing stuff.

Thank you!

>   Just a couple of comments that I noticed straight away: there's lots of
> commented out blocks that should be removed if they aren't going to be used,
> and there's a worrying number of XXX tags that suggest some work remains to be
> done....?

The tags had meant so in the early stage of development
circa 1 March indeed.  During testing and debugging, they
have been remained and used as the marks where I touched
the source specifically.  Generally, I have replaced every
occurrence of ANSI-WIN32 API that operates on a file name
with macros in winsup.h

And a few tags means some work remains yet really.  They
are in miscfuncs.cc, where I put codes that replace ANSI-
WIN32 API generally:

u_mbstowcs:  combining form conversion needs work on
   combining grave, acute etc. especially.

GetCommandLineU:  a large fixed-size dynamic buffer is
   used.  It just works fine for most cases, and falls
   back gracefully for ASCII input.  It operates well
   in 99% cases.  For the rest cases, it needs work
   (in such cases, you will see wrong conversion of
   non-ASCII characters for now).

I'm sorry for confusion.

>     cheers,
>       DaveK

Regards,
-- SUZUKI Hisao