[cairo] Alternatives to autotools-based build systems

Alan W. Irwin irwin at beluga.phys.uvic.ca
Thu May 1 14:29:29 PDT 2008


On 2008-05-01 21:23+0200 Leon Woestenberg wrote:

> Hello,
>
> On Tue, Apr 29, 2008 at 9:39 PM, Behdad Esfahbod <behdad at behdad.org> wrote:
>>  All that sounds nice and good, except that autotools is the one I
>>  understand and have full control when using.  Our autotools setup is not
>>  scary for anyone who has checked it out.  To add a new cairo backend is
>
> autotools is scary for most of us, but so was the GNU toolchain when I
> first met it.
>
> However, I think autotools does do the job perfectly in a wild variety
> of scenarios, also considering the cross-compilation scenarios.

Just as a point of information, both cross-compilation support and Fortran
support are known to be poor for CMake 2.4.x, but I am told that CMake 2.6.0
(now in release candidate status) has dealt with both those issues.  So no
question, CMake is still not as mature as autotools in certain respects, but
it is addressing those issues, and it is substantially better than autotools
in other respects (which I and others have already indicated previously in
this thread).  In fact, we have actually had fewer complaints about our new
CMake-based build system than our previous "matured" autotools-based build
system.

>
> Kudos to Behdad for maintaining cairo's autotools setup.

Agreed.  From what he said it is not much effort for him and he is
relatively keen to continue so more power to him.

I was probably the second-ranked expert on autotools for PLplot, but in
contrast to Behdad, I got tired of supporting autotools for our rather
complicated build (core C library which depended on two smaller internal
libraries, many optional language bindings, many optional dynamically loaded
plugins with a wide variety of external library dependencies, and an
optional build of our documentation from docbook source).  So moving to
CMake (even though I did a substantial fraction of the work in developing
that new build system for PLplot) was a welcome relief for me.  Of course,
each project is different, and for cairo you have a keen autotools guy so it
makes sense to continue with autotools for a while at least.  But from my
good experience switching from autotools to CMake for PLplot and from the
experience of many others who have done a similar switch, I urge the cairo
developers to keep trying CMake and the other build-system alternatives because
I am sure you will find a build system that you will like much better than
autotools for the reasons that have been mentioned.

Alan
__________________________
Alan W. Irwin

Astronomical research affiliation with Department of Physics and Astronomy,
University of Victoria (astrowww.phys.uvic.ca).

Programming affiliations with the FreeEOS equation-of-state implementation
for stellar interiors (freeeos.sf.net); PLplot scientific plotting software
package (plplot.org); the libLASi project (unifont.org/lasi); the Loads of
Linux Links project (loll.sf.net); and the Linux Brochure Project
(lbproject.sf.net).
__________________________

Linux-powered Science
__________________________


More information about the cairo mailing list