Re: [PATCH] Added Kazuhiro's new wchar functions to cygwin.din

>>> On Fri, 30 Aug 2002 17:13:58 +0400
>>> egor duda <deo <at> logos-m.ru> said:

> btowc, wctob,
> mbsinit, mbrlen,
> mbrtowc, mbstowcs, mbsrtowcs,
> wcrtomb, wcstombs, wcsrtombs

The current implementations of these functions don't conform the
standard at all. I will correct these implementations in the
near feature. I recommend the dll doesn't export these functions
and the functions added by me until then.
____
  | AIST      Kazuhiro Fujieda <fujieda <at> jaist.ac.jp>
  | HOKURIKU  Center for Information Science
o_/ 1990      Japan Advanced Institute of Science and Technology

Re: [PATCH] Added Kazuhiro's new wchar functions to cygwin.din

Kazuhiro Fujieda wrote:

> >>> On Fri, 30 Aug 2002 17:13:58 +0400
> >>> egor duda <deo <at> logos-m.ru> said:
>
> > btowc, wctob,
> > mbsinit, mbrlen,
> > mbrtowc, mbstowcs, mbsrtowcs,
> > wcrtomb, wcstombs, wcsrtombs
>
> The current implementations of these functions don't conform the
> standard at all. I will correct these implementations in the
> near feature. I recommend the dll doesn't export these functions
> and the functions added by me until then.

So, instead of Corinna fixing the cygwin.din to/not to export them,
could you submit the patch?

Earnie.

RE: Unicode filename patch

> Hi,
>
> I haven't been able to follow cygwin ml for a while (too busy at
> work ;( ),
> but did Christopher Faylor ever consider merging the unicode
> filename patch
> you created? I followed CVS for a while, but it didn't appear there (I
> haven't checked lately, though.)
Ask on cygwin-patches what Chris wants to do with it. It's best discussed
there.

BTW, the patch is not complete; certain functions don't work. It's main
purpose was to allow reading and writing files with Unicode filenames -
something that wasn't possible at all before.

Saying that, it did work quite well with UTF8 aware shell tools and a UTF8
terminal.

Chris

Re: [PATCH] Added Kazuhiro's new wchar functions to cygwin.din

>>> On Sun, 01 Sep 2002 08:50:39 -0400
>>> Earnie Boyd <earnie_boyd <at> yahoo.com> said:

> So, instead of Corinna fixing the cygwin.din to/not to export them,
> could you submit the patch?

Here it is.

2002-09-02  Kazuhiro Fujieda  <fujieda <at> jaist.ac.jp>

	* cygwin.din: Revert exporting new wchar functions.
	* include/cygwin/version.h: Revert bumping API minor number.

Index: cygwin.din
===================================================================
RCS file: /cvs/src/src/winsup/cygwin/cygwin.din,v
retrieving revision 1.62
retrieving revision 1.59
diff -u -p -r1.62 -r1.59
--- cygwin.din	30 Aug 2002 13:30:26 -0000	1.62
+++ cygwin.din	28 Aug 2002 10:50:26 -0000	1.59
 <at>  <at>  -1031,58 +1031,6  <at>  <at>  wcscmp
 _wcscmp = wcscmp
 wcslen
 _wcslen = wcslen
-wcscat
-_wcscat = wcscat
-wcschr

Re: Unicode filename patch

On Sun, Sep 01, 2002 at 02:09:52PM +0100, you [Chris January] wrote:            
> > Hi,                                                                         
> >                                                                             
> > I haven't been able to follow cygwin ml for a while (too busy at            
> > work ;( ),                                                                  
> > but did Christopher Faylor ever consider merging the unicode                
> > filename patch                                                              
> > you created? I followed CVS for a while, but it didn't appear there (I      
> > haven't checked lately, though.)                                            
> Ask on cygwin-patches what Chris wants to do with it. It's best discussed     
> there.                                                                        
>                                                                               
> BTW, the patch is not complete; certain functions don't work. It's main       
> purpose was to allow reading and writing files with Unicode filenames -       
> something that wasn't possible at all before.                                 
>                                                                               
> Saying that, it did work quite well with UTF8 aware shell tools and a UTF8    
> terminal.                                                                     

cgf, to give you some context (I hope you read this :) : I originally begged    
for unicode filename support a couple of months ago. Chris January very         
kindly went ahead and implemented a patch that enabled accessing unicode        
filenames through their utf8 name.                                              

See the discussion in thread starting at                                        
http://sources.redhat.com/ml/cygwin/2002-07/msg00006.html.                      

The executive summary is that when one has unicode filenames (russian,          
chinese, whatever) it is currently not possible to access them _at_all_ with    
cygwin. Chris's patch makes it possible to do so.                               

Re: [franck.leray <at> cheops.fr: tcsetattr timeout problem ?]

Hi!

Saturday, 31 August, 2002 Christopher Faylor cgf <at> redhat.com wrote:

CF> On Fri, Aug 30, 2002 at 10:12:36PM +0400, egor duda wrote:
>>ed> Actually, the patch is wrong :(. I'm looking into it and post a
>>ed> correct one asap.
>>
>>Forgot to include fhandler_tty.cc part.
>>
>>Please check it if possible and feel free to apply it, as i will be
>>away from computer for some time.

CF> I took a look at your patch but it didn't seem right to me.

CF> You were asking about circumventing the 'ready_for_read' code and,
CF> since fhandler_tty_slave::read has the ability to detect a signal,
CF> removal of fhandler_tty_slave::ready_for_read is the right way to
CF> go, AFAICT.

Absolutely. Silly me, i haven't noticed a FH_NOEINTR flag, which was
exactly what was needed. Thanks!

Egor.            mailto:deo <at> logos-m.ru ICQ 5165414 FidoNet 2:5020/496.19

Re: Readv/writev patch

On Fri, Aug 30, 2002 at 10:48:26PM -0400, Chris Faylor wrote:
> On Sat, Aug 31, 2002 at 03:35:48AM +0100, Conrad Scott wrote:
> >"Corinna Vinschen" <cygwin-patches <at> cygwin.com> wrote:
> >> Especially I'm reluctant to introduce your changes
> >> to the sendto and recvfrom implementation since I know there is
> >> a good reason to use the WinSock1 calls in the non-blocking case
> >> even though I don't recall why, right now.  Please skip that
> >> beautyifing patches and just add the readv/writev functionality.
> >
> >I went back to the mailing list archives to see if I could dig up the
> >problem here and it seems that the code to fallback to the winsock1
> >calls in the non-blocking case was introduced as a result of the
> >discussion in the thread starting at
> >http://cygwin.com/ml/cygwin/2001-08/msg00617.html.  Interestingly, the
> >test program that demonstrated the "problem" was itself bogus.  Like,
> >you don't set the non-blocking flag w/ the following code:
> >
> >  printf("Setting NONBLOCK\n");
> >  flags = fcntl (sock, F_GETFL, 0);
> >  flags &= O_NONBLOCK;
> 
> OUCH!

It wasn't quite that but actually the whole thread seem to be based
on a handful of misinterpretations.

Ok Conrad, send your patch right away.

Corinna

Re: [PATCH] Added Kazuhiro's new wchar functions to cygwin.din

On Fri, Aug 30, 2002 at 01:32:04PM -0400, Nicholas Wourms wrote:
> Corinna Vinschen wrote:
> >On Fri, Aug 30, 2002 at 05:13:58PM +0400, Egor Duda wrote:
> >
> >>It was a typo, sorry. Now, after double-checking, it should read
> >>
> >>btowc, wctob,
> >>mbsinit, mbrlen,
> >>mbrtowc, mbstowcs, mbsrtowcs,
> >>wcrtomb, wcstombs, wcsrtombs
> 
> Corinna,
> 
> You forgot to bump the API after you added these remaining 
> symbols to cygwin.din.

Not this time.  I didn't bump it since the additions where within a
few minutes.

Corinna

--

-- 
Corinna Vinschen                  Please, send mails regarding Cygwin to
Cygwin Developer                                mailto:cygwin <at> cygwin.com
Red Hat, Inc.

Re: [PATCH] Added Kazuhiro's new wchar functions to cygwin.din

Corinna Vinschen wrote:

>On Fri, Aug 30, 2002 at 01:32:04PM -0400, Nicholas Wourms wrote:
>
>>Corinna Vinschen wrote:
>>
>>>On Fri, Aug 30, 2002 at 05:13:58PM +0400, Egor Duda wrote:
>>>
>>>
>>>>It was a typo, sorry. Now, after double-checking, it should read
>>>>
>>>>btowc, wctob,
>>>>mbsinit, mbrlen,
>>>>mbrtowc, mbstowcs, mbsrtowcs,
>>>>wcrtomb, wcstombs, wcsrtombs
>>>>
>>Corinna,
>>
>>You forgot to bump the API after you added these remaining 
>>symbols to cygwin.din.
>>
>
>Not this time.  I didn't bump it since the additions where within a
>few minutes.
>
Well for some reason, the don't-forget-to-bump-the-api thingy (Egor's 
script) came into action while trying to compile, causing the build to 
fail unless I bumped the api.  Perhaps this has to do with the fact that 
I'm compiling from a branch(cygdaemon-branch) rather then HEAD?

very small passwd patch

I thought there was some mention of this already, but I guess
not. This adds a note about passwd not working with Win9x/ME.

ChangeLog:

2002-09-02  Joshua Daniel Franklin <joshuadfranklin <at> yahoo.com>

	* passwd.c (usage): Add note about Win9x/ME not working.

--- passwd.c-orig	2002-09-02 08:23:40.000000000 -0500
+++ passwd.c	2002-09-02 08:24:38.000000000 -0500
 <at>  <at>  -259,6 +259,8  <at>  <at>  usage (FILE * stream, int status)
   "Other options:\n"
   " -h, --help      output usage information and exit\n"
   " -v, --version   output version information and exit\n"
+  "\n"
+  "Note: passwd does not work on Win9x/ME systems.\n"
   "", prog_name, prog_name);
   exit (status);
 }