Return-Path: Delivered-To: apmail-httpd-dev-archive@httpd.apache.org Received: (qmail 26581 invoked by uid 500); 5 Dec 2001 21:34:04 -0000 Mailing-List: contact dev-help@httpd.apache.org; run by ezmlm Precedence: bulk Reply-To: dev@httpd.apache.org list-help: list-unsubscribe: list-post: Delivered-To: mailing list dev@httpd.apache.org Received: (qmail 26570 invoked from network); 5 Dec 2001 21:34:03 -0000 Message-ID: <05e001c17dd4$84814e50$93c0b0d0@v505> From: "William A. Rowe, Jr." To: References: <20011204180020.B13181@Lithium.MeepZor.Com> <02de01c17d1f$fd77a760$93c0b0d0@v505> <3C0E1C4C.B18FCBFD@Golux.Com> <050301c17dad$e1c98860$93c0b0d0@v505> <3C0E6FEB.E046A90@Golux.Com> Subject: Re: Source in .msi packages Date: Wed, 5 Dec 2001 15:27:08 -0600 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 5.50.4522.1200 X-MIMEOLE: Produced By Microsoft MimeOLE V5.50.4522.1200 X-OriginalArrivalTime: 05 Dec 2001 21:33:45.0269 (UTC) FILETIME=[84814E50:01C17DD4] X-Spam-Rating: daedalus.apache.org 1.6.2 0/1000/N From: "Rodent of Unusual Size" Sent: Wednesday, December 05, 2001 1:05 PM > "William A. Rowe, Jr." wrote: > > > > I meant to say [to be absolutely clear] that the decision > > was unilateral due to the two 'majority scenarios' below, > > the interest in conserving absolutely wasted collab donated > > bandwidth [not to mention our mirror friends], and the last > > bugaboo I discovered. > > Okey, did this change in our distribution model ever come up > for discussion on this list? That's the thread to which I > would like a pointer.. I haven't found one yet. No, it was my decision, not discussed, and the entire win32 installer is always up for discussion. I've rarely received extensive feedback on Win32 installer questions, and 0 help. Based on my active participation on c.i.w.ms-w. I didn't see this would be too great an issue. My #1 goal is to _diminish_ the discrepancies between unix and Win32, whenever possible. And the change to the 'new build model' (cvs testing, then tarball testing [for which win32 .zips are made available,] finally binaries testing) isn't a good fit for the old packaging schema. But yes (as you pointed out in the previous message), I've let the docs in several areas lapse, and perhaps over Christmas vacation I will have time to give these all a good vetting. [I think the last comprehensive review I did, for that matter, was last Christmas holiday :-] From: "Rodent of Unusual Size" Sent: Wednesday, December 05, 2001 1:31 PM > "William A. Rowe, Jr." wrote: > > > > So I started to look at how we distribute source in other > > binary packages. > > You mean other ASF packages, other open-source packages, or > specifically the httpd project packages? Other /dist/httpd/binaries/* distributions. This is a binary, those are binaries. > > While reviewing those changes to the 2.0 package, I looked > > at how we package both 1.3 and 2.0 on Unix. And I discovered > > that the inclusion of the src tree was relatively non-existant. > > The unix httpd binaries were just that (binary). > > Uh? In > (for instance), *all* of the .tar.gz files contain the source. > (You can't count how we've done the 2.0 packages because we haven't > released it yet; in 1.3b days, the main difference between a binary > and an alpha was that we built and included a binary 'httpd' for the > former -- but the source was still in the tarball.) If that's so, I need to review this entire issue once again!!! > > There is no way to continue to proliferate this. > > Explain that remark, please.. I don't think I understand > what you mean. If (1) other binaries did not come with source [I'll be checking this week] and if (2) Apache 2.0 can't package the sources+binaries due to the new directory structure (it appears they don't mix well), then there is no reason to continue the expectation of binaries including src/ tree on Win32. If, OTOH, we _do_ consistently include source in 1.3, _or_ we will be consistently including source in all 2.0 packages, then I need to revisit and return to creating a -with-src flavor of the win32 installer. But I think it's safe to conclude that someone seeking -with-src can follow directions to first install the .msi installer [which the .exe flavor does on it's own.] So there is no need for .msi + .exe flavors of a -with-src build. The net package would be a waste of bandwidth, and the .exe flavor was created soley for non-programmer end-users who just can't follow the instructions. Bill