From: cgf AT bbc DOT com (Christopher Faylor) Subject: Re: Question about getopt in Cygwin32 6 Feb 1998 06:09:52 -0800 Message-ID: References: <199802060311 DOT TAA23947 AT rtl DOT cygnus DOT com> Reply-To: cygwin32-developers AT cygnus DOT com To: cygwin32-developers AT cygnus DOT com In article <199802060341 DOT WAA20238 AT subrogation DOT cygnus DOT com>, Ian Lance Taylor wrote: > From: Geoffrey Noer > Date: Thu, 5 Feb 1998 19:11:33 -0800 (PST) > > And after my latest change, they aren't. Should we include a version of > getopt in libcygwin.a? What about in cygwin.dll as an exported > function? > >I don't think we should include getopt in cygwin.dll. getopt >communicates using several variables--optind, opterr, etc.--and DLLs >do not provide convenient access to variables. > >If we did put getopt in cygwin.dll, we would have to fix up the getopt >header file to make the right happen by default. I assume that's why getopt was put in libcygwin.a rather than in the .dll to begin with. I was surprised to find that getopt() doesn't show up in any file in my include area. Am I missing something? If cygwin.dll is supposed to mimic a libc.so then I think getopt should be a part of it. There are several time related variables being exported already using the "normal" #define var (*__imp_var) method. I don't see any reason why getopt couldn't use the same technique -- much as I hate it. -- http://www.bbc.com/ cgf AT bbc DOT com "Strange how unreal VMS=>UNIX Solutions Boston Business Computing the real can be."