Subject: Re: bin/32573: [dM] fails gratuitously
To: None <,,,>
From: Elad Efrat <>
List: netbsd-bugs
Date: 01/20/2006 16:20:03
The following reply was made to PR bin/32573; it has been noted by GNATS.

From: Elad Efrat <>
To: der Mouse <mouse@Rodents.Montreal.QC.CA>
Subject: Re: bin/32573: [dM] fails gratuitously
Date: Fri, 20 Jan 2006 18:17:59 +0200

 der Mouse wrote:
 > I mentioned it because it was the first thing I found, going through
 > serially, looking for anything
 > kernel-build-ish.
 so you didn't read at least the TOC entirely...
 > If it is (and stays) necessary to read the crosscompiling chapter to
 > figure out how to use correctly when *not* crosscompiling,
 > then yes, the documentation needs fixing.
 > If it is necessary to read the webpage at all to figure out how to use
 > correctly, there needs to be a clear reference to the webpage
 > in BUILDING - and, preferably, in the comments at the top of
 > itself, though that's less important in the presence of BUILDING.
 i agree to both statements.
 > Right.  In this respect, the documentation accurately describes the
 > behaviour.  I just hold that the behaviour is broken.  Determining what
 > needs rebuilding and rebuilding just that is what make is all about;
 > pushing that task off onto a human (ie, deciding whether to use -u or
 > not) is a step backwards.
 sometimes a "clean" build is required, sometimes you can use update.
 i rather we keep the tools as they are and provide easy instructions
 on how to use them.
 > If that were the case, I'd be ranting about's mere presence.
 > But I've long since stopped expecting modern NetBSD to be pleasant to
 > use; I'm running 3.0 for work, where I do a number of things I wouldn't
 > put up with at home.
 this is exactly why i am in strong opposition to any code change you
 imply in this pr. you are the first person i know -- and i know quite a
 lot of people, who use the various bsd variants and/or linux -- to
 complain about in such a way that would suggest removing it
 > In this case, you are wrong about what I want; I want to
 > change, because I believe a that works as currently documented
 > would be significantly better than correct documentation of the way
 > currently works.
 i never said the documentation is broken. what i said is that i would
 agree to *improving* it; if that wasn't clear...
 > How does that mean I suddenly have to care about something in a chapter
 > on cross-compiling, when I know perfectly well I'm not cross-compiling?
 because that what the documentation tells you to do. you don't really
 expect to selectively follow the documentation and then rant about how
 broken it or the code is, do you?
 > My previous Unix experience, which is not used to having to do make's
 > job for it.
 you are not using make. you are using from BUILDING:
 tools/	``Reachover'' build structure for the host build tools.
 	This has a special method of determining out-of-date status.
 in that case, your unix experience *is* a detriment because you are too
 attached to one conservative way of doing things.
 Elad Efrat