apr-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Tollef Fog Heen <tfh...@err.no>
Subject Re: svn commit: r794485 - /apr/apr/trunk/configure.in
Date Sun, 19 Jul 2009 19:52:29 GMT
]] Bojan Smojver 

(no need to Cc me, I read the list, as my Mail-Followup-To says.)

| Sorry, still not convinced. If a particular build system sucks, be that
| whichever distribution, then the build system should be fixed. I do not
| see why everyone should be penalised at run time for someone else's
| bugs.

It's a trade-off of what you require of the machines running
applications vs what a supported platform is.  Maybe a buildd daemon has
to run a newer kernel in order to support its SCSI controller, it
doesn't have to be a wish to get all new and shiny syscalls available.

Not doing run-time detection means that everybody using Debian and
derived distros (including Ubuntu) will get the castrated version where
any and all features of new kernel versions are not used, so not doing
runtime detection also has a penalty for users.

| I don't think the above says anything about building APR on one platform
| and then running it on another. I could be wrong, of course. And, as
| always, if other folks think doing runtime detection is the way to go,
| then we'll do that.

I wouldn't call a vaguely different kernel version a different platform,
but that might just be splitting hairs.

Tollef Fog Heen
UNIX is user friendly, it's just picky about who its friends are

View raw message