[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]
Re: Update Problem
On 23 Apr 2010 at 13:05, Greg Troxel wrote:
> I had rar flagged as notouch and pretty sure those were
> specified when I used pkg_chk/pkg_rr
> anyway rar disapeared followed by anything requiring rar.
> it sounds like you are using "pkg_chk -u". Your call, but pkg_rr is not
> supposed to remove dependencies - just do 'make replace'. If you turn
> on DESTDIR, then it should be doing this via pkg_add -u after building a
> binary package. The intended behavior is that pkg_rr may fail, but it
> won't leave you system with (much) missing - zero missing packages with
> destdir, and one without destdir.
> If you send a patch to pkg_rr to have a variable akin to rebuild and
> automatic and to exclude packages with that variable, I'll look at it
> and very probably commit it. I haven't done this partially because I
> haven't figured out a good name. I lean to 'norebuild' because
> 'exclude' is only about pkg_rr, and this tag is intended to be
> After previous update there were 1324 packages installed
> pkg_rr knocked that back to 640 and there were lots of
> fix problem then continue
> I don't get this.
> That took total installed to 652 and after that it's
> now 2 days later without further interruption, 1130
> installed and update still working away on openoffice.
> again this sounds like pkg_chk -u, not pkg_rr.
I now think that's likely as I thought I had rar
deselected but mustn't have.
A few of the fix problems were due to "make clean"
failures with permission denied, ie make install
stage as su-root seemed to have left files with root
permissions in work directories. I should have kept
> I have now been doing 'pkg_rolling-replace -uv -k' and then solving
> problems afterwards. I find a lot of things get built and not too much
> broken, and that the broken things are usually more leaf-like than
> It's impressive of pkgsrc to be able to do an automatic from-source
> update of 1000 packages and have even 950 build ok at any random time.
Yes, I've found once the odd problems are resolved it
gets through reasonably fast especially when there are
Main Index |
Thread Index |