[petsc-dev] confusing download instructions

[Jed Brown]

Barry Smith <bsmith at mcs.anl.gov> writes:

>      And how about generating a maint tarball each night? Then users
>      can get PETSc 3.4 with all current fixes directly without needing
>      git?

I don't believe that updating to a new petsc-3.4.x.tar.gz is anywhere
near as friendly as 'git pull && make'.

>       Philosphically it seems silly that we are putting fixes into
>       maint on a random regular basis but end users don't benefit
>       until someone manually creates a new petsc-3.4.*.tar.gz every
>       couple of weeks. Why doesn't each new fix to maint lead to an
>       automatic new petsc-3.4.*++ tar ball ready for people to use (or
>       at least each night generate a new one). Now people are
>       installing known buggy code for a bunch of days for no good
>       reason.

This is why I often recommend that people track 'maint'.  Rolling a new
petsc-3.4.x on each commit is needlessly noisy because users will
constantly be nagged that a newer version is available.  I prefer the
model where we make maintenance releases every month or so, such that
these tend to have sufficient value to justify installing.  I have no
objections to a nightly snapshot of 'maint', but I don't think we should
call them 3.4.x.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 835 bytes
Desc: not available
URL: <http://lists.mcs.anl.gov/pipermail/petsc-dev/attachments/20130628/d76d3313/attachment.sig>