Mark J. Nelson writes: > I think I really fouled something up in my workspace state for those > builds, and the only thing I can think of is that I had an interrupted > build in a workspace that I cloned using tar.
Yikes ... ok. > I think what I'm willing to promise (jbeck and dmarker) is to not pollute > their findunref output with our putback. As long as we don't care about > successive sparc and i386 clobber builds for findunref in the same > workspace (who would DO such an icky thing?) we're pretty much there. I certainly don't do that. I use separate child workspaces to do my per-architecture builds, so I don't pollute my main workspace with generated trash, and so that if something goes wrong, a quick rm -rf (or perhaps zfs destroy) gets me back to normal. But there do seem to be people who (for whatever reason) like to do alternating builds in a single workspace, and even a few who are disappointed that they can't do both x86 and SPARC in the same workspace at the same time. I view them as mildly off-kilter (akin to desparately wanting to plumb "foo" as an IP interface name rather than "foo0," just because DLPI Style 1 appears to allow it), and I don't sweat it much. Any time someone runs into a problem like that, I can always say, "no, that never happens to me." ;-} > > % ws /path/to/ws > > % cd $SRC/tools > > % dmake install > > > > Right? > > Nope. It will just not include the scm-specific exception list entries. > But the tools that are built will be the same. Oh, right. I was expecting that to expand badly, but it doesn't. OK, then. > >> I don't think we want to hack the scripts for this. I think we want to > >> send a build machine flag day to install the right tools, and make sure it > >> mentions the which_scm failure. > > > > OK. > > I did fix both the nightly.sh and bldenv.sh scripts to get which_scm from > $(dirname $(whence $0)), and to fallback to $PATH (by way of invoking > which_scm with no leading path) if that file fails a -x test. Fair enough. > Tested bldenv unmodified on teamware and mercurial workspaces, hacked > nightly to make sure it found the right thing. > > Will push soon. Sounds good. -- James Carlson, Solaris Networking <james.d.carlson at sun.com> Sun Microsystems / 35 Network Drive 71.232W Vox +1 781 442 2084 MS UBUR02-212 / Burlington MA 01803-2757 42.496N Fax +1 781 442 1677
