From: sandmann AT clio DOT rice DOT edu (Charles Sandmann) Message-Id: <10206132142.AA16953@clio.rice.edu> Subject: Re: DJGPP 2.04 release To: djgpp-workers AT delorie DOT com Date: Thu, 13 Jun 2002 16:42:48 -0500 (CDT) In-Reply-To: <3D08B922.E328F8B8@phekda.freeserve.co.uk> from "Richard Dawe" at Jun 13, 2002 04:24:18 PM X-Mailer: ELM [version 2.5 PL2] Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Reply-To: djgpp-workers AT delorie DOT com Errors-To: nobody AT delorie DOT com X-Mailing-List: djgpp-workers AT delorie DOT com X-Unsubscribes-To: listserv AT delorie DOT com Precedence: bulk > > The main issue, as Charles points out, is to find a > > victim^H^H^H^H^Holunteer who will manage the release through the alpha > > and beta testing. Once we have that person in charge, we can simply > > produce the first alpha and start the release process. > [snip] > > What exactly is involved in managing the release? Mostly having time to do it. Creating distributions, getting the word out to get people to test it, consensus that it's a good thing to release. > How about this: > > * Review everything that's changed between 2.03 and 2.04 and formulate tests > for the new functionality and to check that nothing's broken. Probably overkill. Some new tests - ask for help. > * Track and chase down bugs. And ask for help ... check on progress ... > * Build lots of software using the release, to see if it breaks anything. > > * Build & package alphas, betas. > > Has this list omitted anything? A lot of it's just time to constantly keep everyone up to date, to let people know it's coming and what the key issues are. Sometimes you'll get some disagreement and just make a call on what you think is right. Planning for what would "break" badly in the future if it's not in the release (missing feature, utility, etc). What compiler version you build with. Etc. > Can you estimate how much time it takes roughly per week to be the "release > manager"? It goes in chunks. Ask Andrew on the builds/uploads/web pages (10-20 hours?). While in testing fairly small (few hours a week). V2.03 refresh was a trivial update, but I spent essentially all my free time around Thanskgiving and Christmas for 6 weeks getting it built, tested, re-released, done. About a month of weekends before that collecting and testing. > I was wondering if there are more/same number/less people involved in the > development of DJGPP now than there have been in the past (i.e.: for testing > purposes). > > I suppose there are more combinations to test now than there may have been: > DOS in its various "pure" incarnations (MS-DOS, FreeDOS, etc.), Windows '95s, > Windows '98s, Windows NT 4 without LFN TSR, Windows NT 4 with LFN TSR, Windows > 2000, Windows XP. I seem to remember you can use the LFN TSR with Win2k, > WinXP, although I'm not sure why anybody would want to do that. You ask for help, test what platforms you can. The LFN TSR on Win2K/XP fixes some bugs and provides features like big disk size support interrupts. > I'm thinking of, er, volunteering. 8) I won't be around > for August/some of September, but I'm nearing the end of my DJGPP and zippo > to-do lists, so I might have time for it over the 9-month period. That would be excellent. Plan to do a major test release a week or two before you are gone, then the testers collect results while you are away. Release schedules largely depend on how much you do and how hard you push; it could be much less than 9 months but everyone might be mad at you for a while :-) It could also be a small group effort with hand-offs at key times; for example 2.03 refresh would not have happened without Andrew's help and support (not to mention lots of private emails to the key "you-know-whos" around here).