Open MPI logo

Open MPI Development Mailing List Archives

  |   Home   |   Support   |   FAQ   |  

This web mail archive is frozen.

This page is part of a frozen web archive of this mailing list.

You can still navigate around this archive, but know that no new mails have been added to it since July of 2016.

Click here to be taken to the new web archives of this list; it includes all the mails that are in this frozen archive plus all new mails that have been sent to the list since it was migrated to the new archives.

From: Jeff Squyres (jsquyres_at_[hidden])
Date: 2005-09-06 09:41:22

On Sep 6, 2005, at 10:34 AM, Peter Kjellström wrote:

> What is the recommended way to update a local svn snap shot? (a tree
> that has
> been checked out, autogen'ed, configured and built) is it:
> $ svn up
> $ ./
> $ make distclean
> $ configure ...
> $ make all install

No need for "make distclean" -- the dependency analysis stuff should
take care of rebuilding what you need to rebuild.

Also, is not always necessary. is really only
necessary when something in configure changes (e.g., or
any *.m4 file). Unfortunately, we've been making a lot of changes to
the configure system recently, so autogen *has* been necessary a lot,
but in general, have a look at the files that go by when you "svn up"
and see if you see or any *.m4 file get updated. If no,
you should be able to simply "make all install" (note that parallel
builds are also supported -- e.g., "make -j 4 all; make -j 4 install".

FWIW, we try to make changes to configure during off-peak work hours
(e.g., US evenings, weekends, and holidays).

> Does this change if I vpath build?


> Is there a hard-cleanup-that-will-allways-work way?

distclean is a pretty good way to go if you want to start from
almost-scratch (starting from total scratch would be removing your SVN
checkout and getting a new one -- something that is rarely necessary).
Or if you do a VPATH build, just remove the entire build tree (or make
a new build tree).

> (and the opposite) Is there a minimum-recompile-regen-way that will
> usually
> work?

See above.

{+} Jeff Squyres
{+} The Open MPI Project