Makefile.in in Eridani release

[Robert Jongbloed]

Sheesh! I know what a Makefile.in is supposed to do.

We are now going to get into a deep philosophical argument over why the 
user MUST execute ./configure before compiling the system when they do 
not really need to. It is piece of tradition that has no basis in 
necessity and has bugged me for the 20 odd years I have been programming 
in Unix environments. In the PTLib and OPAL case, the Makefile can see 
./configure has not yet been run, executes it, then runs the 
sub-makefile that is generated. So, a user can just do one command, 
"make install", and it all happens in one go. This is bad?

And what about systems that use cmake, or some other make system, that 
does not have Makefile.in?

If your packaging system is so inflexible as to not be able to deal with 
a library without a Makefile.in then I suppose I will just have to add a 
dummy one. But it is, as I said before, stupid.

*Robert Jongbloed*
/OPAL/OpenH323/PTLib Architect and Co-founder./
Commercial support at http://www.voxlucida.com.au

On 16/09/2013 9:17 PM, Mark Purcell wrote:
> On Mon, 16 Sep 2013 09:34:12 Robert Jongbloed wrote:
>> So, what are you proposing? We include an empty file with that name?
> Robert,
>
> No not an empty file, rather a Makefile.in which can be used to regenerated
> the Makefile via ./configure. This is pretty standard practice using autoconf
> to support many different build systems:
> http://www.la-samhna.de/library/compile/configure.html
>
> Even more interesting is that is exactly the way ptlib/ opal used to function
> prior to Eridani:
>
> http://sources.debian.net/src/ptlib/2.10.10~dfsg-2/Makefile.in
> http://sources.debian.net/src/ptlib/2.12.5~dfsg-1/Makefile
>
> Mark

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.alioth.debian.org/pipermail/pkg-voip-maintainers/attachments/20130917/1a78714c/attachment.html>