On Thu, Jan 04, 2007 at 05:52:53PM -0500, Chuck Robey wrote: > This is my first post to a OpenBSD list, so please, if I make any > mistake, go on and correct me, I will take anything constructively > (like, am I hitting the right list, or the best list for my topic, or > violating any rules, such as having longish lines, which I am trying > hard to avoid using this Seamonkey editor).
Yes, this is the right list; yes, wrapping lines properly is appreciated, but you seem to have managed; and no, I don't see anything blatantly idiotic below. AFAIK, that's enough to avoid the flames. ;-) > OK, I have a fair amount of experience with FreeBSD kernels, but none > using OpenBSD. My platform is a teeny little Zaurus, and I am trying > to see if I could use some tools such as ccache to speed compilation. > I think that the best way for me to use ccache is to be able to > revector the CC and C++ compilers ... but I'm not certain, could I just > put something like > > "make CC=ccache build" > > as my main compilation command (after, of course, I do the dependencies) > and get the compioler revectored to my ccache tool? http://marc.theaimsgroup.com/?l=openbsd-misc&m=109043549403999&w=2 seems to suggest that putting the appropriate CC= line in /etc/mk.conf might work; this is not documented in mk.conf(5). http://marc.theaimsgroup.com/?l=openbsd-ports&m=116060629621783&w=2 suggests likewise, and points to another possible solution, provided you have quite a few Zauruses. However, note that this affects ports *only*. And frankly, I have no idea what your line above might do. It's not supported, that's for sure... > Other than that, the only thing I have done is to remotely nfs mount src > and obj directories onto a big server machine, and have that machine > handle my cvs completely remotely, but my real question is abount using > cache, or any other suggestion you could toss at me. Yes - there is one big 'why' in all this. ccache is really useful if you repeatedly rebuild the same thing, and cannot afford to actually rely on make/the makefiles doing the right thing. However, unless you are going to do some hefty development work on that Zaurus, which I really wouldn't recommend, you are unlikely to need to build much of anything on it. If you stick to the patch branch, you'd *never* have to compile anything more than a handful of files; and even if you follow -stable and periodically rebuild everything, you're not going to need to build that much. If, and only if, you wish to follow -current, and the snapshots do not suffice, using ccache might be a good idea. But only then - and even in that case, if you're going to be doing unsupported stuff, you *might* be able to get someone to explain you how to do cross-compilation (I wouldn't know, for OpenBSD; the FAQ suggests it might be possible in 5.11.13, but it is probably not a good idea, as is pointed out in the same place.) Joachim
