Fri 19 Oct 2007
Getting ready for Leopard
Category : Technology/GettingReadyLeopard.txt
I've been getting lots of mail about the status of the projects, with Leopard coming in only seven days. And I don't have time to answer them all because I'm deep into this thing. I've got MailServe almost ready. Then there's DNS Enabler. And I may hold off on WebMon for a few more days after Leopard ships, to see if I've hit anything that I hadn't foreseen, with the first two apps.
So this is what I have up to now :
I'll probably be merging these two into just MailServe. Apple chose this latest developer release of 3 weeks ago to make a change as to how Postfix starts up. The launchd plist we used in Tiger doesn't work anymore. I managed to find a solution but, by having to do that, and having gone so far, I found that it opened up a way to put together a few things that I've been experimenting with over the last two years - like :
1) how to allow MailServe/Postfix Enabler to be configured from a non-admin account (a lot of requests for that), and
2) how to avoid using sudo so that it doesn't open up a security hole during the time MailServe/Postfix Enabler is running
3) how to configure Postfix and make changes to your system without touching a single one of the original system files, so that the system is left in its original pristine state if you choose to de-install any of my applications - also hope this will make it easier to get the mail server running for people who've messed up their original system files because I don't even need to look into these system files now
4) and how to give visual indications that the relevant mail-related ports are opened and working without getting people to do a telnet something or other on the terminal,
5) plus a few more that I forgot
And so I've put this lot in, only in the last three weeks, because somehow it has all clicked, and I'm furiously trying to get everything tested and ready for the new MailServe to be released also on the 26th.
I made a version that will work with the changes that Apple has wrought on Leopard. What I had planned to do was to release this version that I am already using on the 26th, so that people who need to get their DNS running on Leopard, using DNS Enabler, can have something to work on immediately, and then I'll put in (and release) the rest of the things I'm working on in the following weeks.
I expect a lot of support load the first few weeks Leopard ships and so I'm wary of changing too many things at once since I've had so few weeks for testing them.
MailServe already has a boatload of changes to take advantage of the things I can now do on Leopard but these things might break - if they don't, then all the new techniques that I'm applying on MailServe will also go into DNS Enabler.
But, then, I've still some days yet. so I might still be able to get these changes done and hope the gods smile on me without any mishaps. (If they break, please be patient. This is really living on the edge.)
I've got a WebMon that will now work with the new Apache2 on Leopard. But I think I've got one more major thing to fix. I may not be ready on the 26th with WebMon.
I plan to merge MailServe and Postfix Enabler into just MailServe and price it, like DNS Enabler and WebMon, at 15 USD. I won't be doing upgrades (because it costs significantly more to put in the backend administrative systems to handle all those upgrade permutations). I thought, if I charge upgrade-like pricing each time, it'll allow me to simplify my business and spend more time working on building these applications. I've got more planned - LDAP Enabler, Dovecot (if I get permission to bundle it), anti-spam stuff, iPhone! (I've got all the pieces to do a CRM-like system - contacts, accounting, email, address-book integration, and so on...)
PS : Luca is already Leopard-ready though I'm looking forward to improving its interface. Apple has bundled a nice new Xcode with Leopard. Whatever gives the developers more power will show up eventually in the quality of the applications. I think that's the one thing that won't be readily apparent to an end user - all those invisible improvements that Apple has made to the OS foundations with Leopard - it has made what was already a very productive platform even more conducive to the development of great apps. If we can get past the teething problems with Leopard in the first few weeks, I think we can all start to smile. This is a helluva of a good platform to build your business on, by far.