new buildsystem -> first steps

Enrico Weigelt weigelt at metux.de
Tue Oct 4 15:22:54 PDT 2005


* Adam Jackson <ajax at nwnk.net> wrote:
> On Tuesday 04 October 2005 13:41, Enrico Weigelt wrote:
> > And now I'm going some steps further:
> > <snip>
> 
> a) What actual problem that people are having does this solve?

1. Consistent and reliable buils. Simplicity reduces the amount 
   possible points of erros. 
2. A base platform for automated optimizations (ie. choosing what 
   to link statically, when to strip, etc) and QA methods, which 
   are especially interesting for embedded systems.
   
> b) What is your timeframe for implementing this?

The first described steps: about one or two weeks. 

BTW: the toolchain abstraction (including my own libtool implementation)
is available at:

    cvs://anoncvs@nibiru.metux.de/home/cvs/repositories/unitool/

<snip>

> c) Given that probably _the_ motivation for selecting autotools as the new 
> Xorg build system was its familiarity to the existing broader OSS development 
> community, what does Xorg stand to gain by moving back to the state of using 
> a build system that no other OSS project uses?

#1: "moving forward" would describe better.
#2: I'm gonna adapt all packages in my repository in nearest future. 
    More than 100.

> d) Given that we absolutely will not be changing the basic build system 
> from its current form before 7.0, what do we gain by throwing away the 
> effort invested in the new build system for 7.1 or later?

You do not have to do very much. You just have to include a few files
and a pointer to my buildsystem.

I'll do the porting, at least for the packages of my interest.

> e) How do you justify presenting a new build system proposal for use 
> within X, when (as far as I know) there is no existing OSS project using 
> it, and thus there is no empirical evidence that it is suitable?  
> Admittedly, some project needs to be the first. Why should X be the 
> one to take that risk?

Because I need X get built properly.

> f) Why are you bringing up its design on an Xorg mailing list, given that 
> designing, implementing, and maintaining build system software is utterly 
> outside the scope of Xorg?  (Hence dropping imake in the first place.)

Simply because I have to start somewhere finding people who'll be
probably interested. 

In fact I've already got an own list for my project, but I'm still the
only one there, and talking to myself gets boring after a while ;-O


cu
-- 
---------------------------------------------------------------------
 Enrico Weigelt    ==   metux IT service
  phone:     +49 36207 519931         www:       http://www.metux.de/
  fax:       +49 36207 519932         email:     contact at metux.de
---------------------------------------------------------------------
  Realtime Forex/Stock Exchange trading powered by postgresSQL :))
                                            http://www.fxignal.net/
---------------------------------------------------------------------


More information about the xorg-modular mailing list