Peter Hessler wrote: > Current and 4.0 are *very* different beasts. In this particular case, > there have been a number of changes to all of the perl ports, which > would show if the systems are out of sync (which they are). > OK. I suspected that, that I'd caused my own trouble, but I that I would pass it along, just in case I was wrong. I was used to how other systems worked, and was a bit surprised. > Downgrading is not supported, check out "Keeping Things in Sync" at > http://openbsd.org/faq/faq5.html > Thanks, I'll give that a read. I need to fixs this somehow.
Far as that goes (and maybe I shouldn't even talk about this on this particular list) but I tell you, I really think that having the default target doing the install is a mistake. That's not just me talking as a person who's just stubbed his toe, complaining at the world ... because I don't think I have even seen any other major project that does that ... it's really build behavior that's so unusual, I think I could be forgiven the error. If you wanted to keep the present target behavior for folks, you know that wouldn't be hard to manager, but for the tyro, it's rally like wiring your speaker system using 110V AC connectors (you know, I actually saw one bunch of amateur engineers, actually an acting group, who did that? I never saw it explode, though.) > > On 2007 Mar 20 (Tue) at 14:20:59 -0400 (-0400), chuckr wrote: > :I made a mistake in playing around with openbd, when I was playing about > :with the sources. I didn't realize that a "make" (using the default > :target) would actually, after doing the build, continue to do the full > :install, so I now have a openbsd system that's at what I guess would be > :called "current", right? > : > :Well, that's not really the problem, I think I know how to recover, and > :I'll probably do that shortly (check out the 4.0 release from my > :archive, and rebuild it again). What's happened is, it seems to have > :exposed what might be a major problem, so I thought I would tell what > :I'm seeing, and see what happens. If it' gets ignored or if I get a lot > :of abuse, it's a false alarm, and I can let it go .... > : > :What is it? Well, the indication I first saw was, when at the top level > :of the devel subdir, I did a clean, I was intending to visit all of the > :devel ports, but it seems that the first p5- port failed. OK, I wanted > :to see what was wrong with it, so I constructed a little script to visit > :other p5- ports, looking for one that WOULDN'T fail (all across > :/usr/ports). That just showed me that they ALL failed. OK, I looked at > :the error listsing, and it didn't ring any alarm bells for me, but it > :DID mention two make include files (bsd.ports.mk and modules.mk). I > :figured to goo explore the archive, so I looked at the log of those > :both, and didn't again see anything that meant anything to me, so just > :experimenting, I slid the version of modukes.mk back from 1.5 to 1.4, > :and that seemed to compledtely fix all of the p5 ports everywhere. > : > :So, maybe a couple things it could be, probably the most likely, I don't > :understand yet how you folks in openbsd manage the versioning of stuff > :between your src archive, and ports. I figure to drop this on someone > :who knows more. I hope this doesn't bother you folks too much, and I > :will let it die if it only looks like I was chasing my tail. > : > :Sorry if I'm long-winded. > : > > > -- > Lubarsky's Law of Cybernetic Entomology: > There's always one more bug. > >
