Mailing-List: contact cygwin-developers-help AT sourceware DOT cygnus DOT com; run by ezmlm List-Subscribe: List-Archive: List-Post: List-Help: , Sender: cygwin-developers-owner AT sources DOT redhat DOT com Delivered-To: mailing list cygwin-developers AT sources DOT redhat DOT com Message-ID: <009201c0b373$270a6ee0$0200a8c0@lifelesswks> From: "Robert Collins" To: Subject: setup revisit Date: Fri, 23 Mar 2001 19:27:43 +1100 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 5.50.4133.2400 X-MimeOLE: Produced By Microsoft MimeOLE V5.50.4133.2400 X-OriginalArrivalTime: 23 Mar 2001 08:22:16.0914 (UTC) FILETIME=[5F13A320:01C0B372] What's the general take on the following two options: 1) Setup.exe does everything. It's all in the one exe, there's no "bootstrapping" concept. 2) Setup.exe downloads a _very_ minimialistic cygwin environment, and then uses that to do the rest of the install. Such an environment might consist of cygwin1.dll, tar,rpm,rpmfind. 1: benefits: *less to go wrong, *it's all pretty all the way. downside: *it's _much_ harder for us to leverage unix tools (and remember that also impacts getting contributors). *Understanding symlinks is also harder. *Users can't upgrade userland tools from within cygwin. 2: benefits: * very easy to leverage unix tools. * much more flexability. * contributors can work with the native interface to the packaging format. This includes things like automatic menu driven config prompts.. Note that included menus should be data driven so we can probably make them GUI fairly easily. downside: * it won't be pretty when the real work occurs -- at least not unless we include gtk+/win32 and a gui rpm/rpmfind! * The "bootstrap" intial tarball will be around 1-1.5 Mb (cygwin1.dll is 600 Kb on it's own). I'm in favour of 2 - IMO we'll get more code reuse and faster turnaround time. (All the code for resumable downloads is already written for example). Rob