From: Karel Zak <kzak@redhat.com>
To: util-linux <util-linux@vger.kernel.org>
Subject: non-recursive build system
Date: Tue, 12 Jun 2012 12:03:40 +0200 [thread overview]
Message-ID: <20120612100340.GB23991@x2.net.home> (raw)
The original version was implemented by Diego Elio 'Flameeyes' Pettenò
two years ago:
http://thread.gmane.org/gmane.linux.utilities.util-linux-ng/3297
See also:
http://www.flameeyes.eu/autotools-mythbuster/automake/nonrecursive.html
My current version is available at github, 'non-recursive' branch:
https://github.com/karelzak/util-linux/commits/non-recursive
Notes:
* generate one Makefile for whole package
* make(1) has all information to run tasks in parallel
* numbers (make -j)
recursive | non-recursive
----------+--------------
2 cores: 14.5 sec | 13.2 sec
16 cores: 9.5 sec | 4.3 sec
(... more CPU cores means more fun;-)
* automake variables and rules are defined in subdir/Makemodule.am files
(create one huge Makefile.am is too crazy idea)
* top-level Makefile.am includes all Makemodule.am files
* $(top_srcdir) prefix for lib/ and include/ is unnecessary, just use
lib/bar.c and include/bar.h in makefiles
* lib/ has been converted to libcommon.la, it means that
foo_LDADD = libcommon.la
is enough to use arbitrary stuff from lib/.
* all rules are interpreted from top-level directory, it means that
cd subdir; make
or
make -C subdir
does not work.
* binaries are stored in top-level (build) directory, you don't
have to care about directories, just type:
$ make mount
$ ./mount --help
* all build targets (including test binaries) are visible for bash completion
scripts (e.g. /etc/bash_completion.d/make), so <tab> is your good friend
* disadvantages:
- full paths in makefiles (e.g. foo_SOURCE = subdir/foo.c)
- we can't use the same name for subdir/ and for a final binary
Necessary changes:
- remove partx/ directory, move partx.c to disk-utils/
- remove getopt/ directory, move getopt.c to misc-utils/
- remove hwclock/ directory, move hwclock.c to sys-utils/
- rename fdisk/ to fdisks/
- rename mount/ to mount-deprecated/ (we use new sys-utils/mount.c)
Note that consolidate the code and remove some subdirs (e.g. getopt/) is
probably a good idea independently on the build-system changes.
Comments & objections?
Karel
--
Karel Zak <kzak@redhat.com>
http://karelzak.blogspot.com
next reply other threads:[~2012-06-12 10:03 UTC|newest]
Thread overview: 5+ messages / expand[flat|nested] mbox.gz Atom feed top
2012-06-12 10:03 Karel Zak [this message]
2012-06-12 12:57 ` non-recursive build system Davidlohr Bueso
2012-06-12 15:18 ` Karel Zak
2012-06-12 19:26 ` Davidlohr Bueso
2012-06-27 9:30 ` Karel Zak
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20120612100340.GB23991@x2.net.home \
--to=kzak@redhat.com \
--cc=util-linux@vger.kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).