__small_sprintf

[Andrew Dalgleish]

> -----Original Message-----
> From:	Andrew Dalgleish 
> Sent:	Thursday, August 26, 1999 12:13
> To:	'Chris Faylor'
> Cc:	cygwin-developers@sourceware.cygnus.com
> Subject:	RE: __small_sprintf
> 
> 
> 
> 	-----Original Message-----
> 	From:	Chris Faylor [SMTP:cgf@cygnus.com]
> 	Sent:	Thursday, August 26, 1999 11:12
> 	To:	Andrew Dalgleish
> 	Cc:	cygwin-developers@sourceware.cygnus.com
> 	Subject:	Re: __small_sprintf
> 
> 	On Thu, Aug 26, 1999 at 10:53:58AM +1000, Andrew Dalgleish
> wrote:
> 	>> From the ChangeLog:
> 	>> 
> 	>> > Mon Aug 23 21:37:07 1999  Christopher Faylor
> <cgf@cygnus.com>
> 	>> >   * times.cc (timezone): Use __small_sprintf.
> 	>> 
> 	>> Did you test the change?
> 	>> The format string uses the '+'-modifier and __small_sprintf
> doesn't
> 	>> know
> 	>> this modifier. Therefor a '?' is inserted for the hour.
> 	>[Andrew Dalgleish]  
> 	>Not directly related but...
> 	>After building winsup-src-19990823.tar.gz,  uname now says my
> CPU is a
> 	>"i?ld".
> 	>This breaks configure in a big way. :-)
> 	>
> 	>(As expected, same result on 95, 98, and NT4)
> 
> 	Actually, this should only be an issue for Windows 95 and
> Windows 98.
> [Andrew Dalgleish]  
> My mistake.
> I do an automated build, install, build on all three.
> The 2nd build broke on all three, and I only checked why on two.
> 
> I'll have a close look at the NT build.
[Andrew Dalgleish]  
I'm not 100% sure, but I think a process running in a separate DOS box
loaded an earlier version of the DLL just as the 2nd build started.
Is there an easy way to test the DLL in isolation?