On 12/27/2018 11:54 PM, Ken Moffat via blfs-support wrote:
On Thu, Dec 27, 2018 at 10:38:09PM -0600, Bruce Dubbs via blfs-support wrote:
On 12/26/2018 05:05 PM, Paul Rogers via blfs-support wrote:
I thought I was being clear, but apparently not. I am a contrarian from your
suggestion. I subscribe to the theory and rebuild everything every time I
build a new LFS. It's just easier that way. Machine cycles are cheap,
debugging time is expensive.
And we do that as a part of the release process for every *stable* version
of BLFS, fixing issues as we go. Of course that doen't catch every possible
error. If you build packages in a different order or omit/include optional
dependencies, results will probably differ. It really not possible to build
everything using every combination of package order.
In addition, we have updated packages during the "package freeze",
even when there is no known need to do so (such as fixing
vulnerabilities), which means that the versions I have tested in the
pre-release testing might not be the same as those anybody else
tests.
Generally we don't do that. If a package has not yet been "tagged" then
it is OK since it will be tested as packages that depend on it are tested.
-- Bruce
--
http://lists.linuxfromscratch.org/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page
