[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
cmake version of RTEMS
- Date: Mon, 17 Dec 2007 17:32:51 +0100
- From: ralf.corsepius at rtems.org (Ralf Corsepius)
- Subject: cmake version of RTEMS
On Tue, 2007-12-18 at 00:18 +0800, Yu Chen wrote:
> 2007/12/17, Ralf Corsepius <ralf.corsepius at rtems.org>:
> > On Mon, 2007-12-17 at 20:59 +0800, Yu Chen wrote:
> > > hi,
> > > my team think cmake is better than autotools for rtems.
> > I could not disagree more.
> I am not very familar with cmake and autotools now, so our views
> aren't comprehensive.
I am very familiar with the autotools and am sufficiently familiar with
cmake to not liking and gradually start hating it.
IMNSHO, cmake is a foul promise, many people, esp. newcomers get trapped
into because it is simple to use for trivial cases.
> > > First we reduce retems only for pc386 (about 1MB src codes), then we
> > > chang it to be the cmake version. The performance testing results is
> > > shown below(in 1GB RAM, 2.0GHz Centrion CPU):
> > To me, cmake is nothing but a script-kiddy junk, whose proponents are
> > proud to have reinvented imake in modern clothes.
> You did a lot of great works and improved autotools build
> infrastructure on RTEMS I understand your choice.
That's a different aspect, but is not the core of my argumentation.
Yes, the auto*tools are not easy to get into for beginner and require a
steep learning curve, but they are flexible, portable and once a
configuration has matured very easy to use.
To cmake, the opposite applies. It is easy to get started with for
simple cases, but when things become complicated, they are facing their
limitations. That said, I think cmake currently is seeing way to much
exposure, this tool doesn't deserve.