I was Cherno Bill's batman

[hirez]

Oh, this is just silly.

I've hacked on a recent version of Postfix to make it do IPv6 on HP-UX 11i. It's a bit of a lash-up of #ifdefs, but then the whole of that subsection's like that, and indeed makes some unwarranted assumptions about HP-UX which complicate things. I've tested it and it appears to deliver mail. I've built the thing on 11iv2 (11.23 on PA-RISC, Lord alone knows what it'll do on IPF) which supports IPv6 natively. I've checked that the patch applies cleanly to the latest development release of Postfix.

So why am I sweating over the (perceived) state of the code? I should just punt it off to TPTB and let them worry about it, right?

Right.

[Sigh]

Comments

[jarkman.livejournal.com]

Well, if you tidied it all up and made it lovely, how long would that take ? And who would care ?

[gaius-octavian.livejournal.com]

I love #ifdef. I call Perl scripts from my Makefiles to fake it in Java. And I love operator overloading too, but not as much as I love using negative subscripts in FORTRAN for self-modifying code!

[hirez.livejournal.com]

Lord knows. I think I'd start by refactoring the entire local address discovery code.

[jarkman.livejournal.com]

So, an extensive period of new brokenness would ensue, and it would be a fortnight before everything worked again ?

[hirez.livejournal.com]

Yes, indeed.

Anyway. Patches submitted with appropriate caveats. Mr Venema is far from being an idiot and won't accept anything obviously broken, so I get to discover how to do it right whatever happens.

(Tech validation is an obvious motivation, as is not wanting to look like a pillock in public.)

[hirez.livejournal.com]

Patches refused with good reasons. Better ones being tested now. I think I like this game.