pd-extended minimum needs Re: [PD-dev] Re: [GEM-dev] pdextended autobuild august 10th 2006

Hans-Christoph Steiner hans at eds.org
Fri Sep 15 00:39:01 CEST 2006


I am sick of build problems, so I am not going to do it.  But I'll  
answer questions where I can.

There is some documentation of the Pd-extended makefiles here:

http://puredata.org/docs/developer

.hc

On Sep 2, 2006, at 12:13 PM, B. Bogart wrote:

> Hello Thomas,
>
> Sorry I've taken so long to get back to you.
>
> To me option B, to make pd-extended makefiles that build flext w/ 
> out the
> flext build system, seems the best option.
>
> Of course I (personally) have little/no understanding of those  
> makefiles.
>
> Do you have examples of stand-alone makefiles to build the flext  
> static
> lib and a flext external? Perhaps based on my PSO code in CVS?
>
> When you have time.
>
> Hans, how does that sound to you?
>
> .b.
>
> Thomas Grill wrote:
>> Hi all,
>> i appreciate your engagement for the pd-extended build system but
>> because of large concert projects in the next weeks (or rather  
>> months)
>> i'm currently not able to help you with it.
>> I see three possibilities:
>> - you are using the strange flext build system which is fairly ok
>> documented by writing proper configuration files with your sane build
>> system
>> - you are not using the strange flext build system by building  
>> flext and
>> flext-based externals with your straightforward make files - i can  
>> help
>> you with the few relevant preprocessor symbols
>> - you are not going to include flext and flext-based externals at  
>> all in
>> the pd-extended system
>>
>> all the best,
>> Thomas
>>
>> Am 11.08.2006 um 02:51 schrieb Hans-Christoph Steiner:
>>
>>>
>>> On Aug 10, 2006, at 7:53 PM, B. Bogart wrote:
>>>
>>>> Hans-Christoph Steiner wrote:
>>>>>
>>>>>> also seems pool is missing. I guess I'll have to checkout
>>>>>> everything on
>>>>>> the laptop and try and build the whole monster myself and see  
>>>>>> whats
>>>>>> going on.
>>>>>
>>>>>
>>>>> No one's gotten that stuff working with the Pd-extended build  
>>>>> system.
>>>>> The flext build stuff is strange and it seems Gem is making  
>>>>> progress
>>>>> with the ./configure stuff.  Really, its just a matter of someone
>>>>> doing
>>>>> the work.  If either build system uses the GNU standard   
>>>>> 'DESTDIR' and
>>>>> 'prefix' variables correctly, then the install targets  should  
>>>>> be easy.
>>>>> Otherwise those will have to be implemented.
>>>>
>>>> Hi Thomas,
>>>>
>>>> So can (we/you) alter the flext buildsys so that things like the  
>>>> DESTDIR
>>>> and prefix variables (if they are set) provide reasonable  
>>>> defaults to
>>>> the conf files?
>>>>
>>>> Hans, are these all environment variables or variables within the
>>>> makefile? As in do we need flext to build like this:
>>>>
>>>> build.sh pd gcc
>>>> build.sh pd gcc install $DESTDIR $prefix
>>>>
>>>> or can we just run "build.sh pd gcc" with the environment vars set
>>>> properly?
>>>
>>> They are make variables, which are kind of interchangable with env
>>> vars.  Usually, you would use them like this:
>>>
>>> make DESTDIR=/tmp/ prefix=/usr install
>>>
>>> but you can also do this:
>>>
>>> export DESTDIR=/tmp/
>>> export prefix=/usr
>>> make install
>>>
>>>
>>>> I'll help were I can, It would be really nice to have flext play  
>>>> nice
>>>> with the pd-extended build system.
>>>
>>> It would be nice to have Gem and flext working this way.  What needs
>>> to happen is:
>>>
>>> a) there needs to be some command line call to build everything,
>>> whether its ./configure && make or whatever.
>>>
>>> b) the build system needs to use the standard GNU install variables,
>>> so that they can be overridden easily by the Pd-extended build  
>>> system.
>>>
>>> Here are the GNU standard variables for installation directories:
>>>
>>> http://theory.uwinnipeg.ca/gnu/make/make_117.html#SEC116
>>>
>>> For Pd, since its an odd layout, I had to had a few more, while  
>>> trying
>>> to keep in the spirit of the GNU standards (from
>>> packages/Makefile.buildlayout):
>>>
>>> ---------------
>>> ifeq ($(OS_NAME),darwin)
>>>   libpddir = $(DESTDIR)$(prefix)
>>>   libpdbindir = $(bindir)
>>>   else
>>>     ifeq ($(OS_NAME),windows)
>>>       libpddir = $(DESTDIR)$(prefix)
>>>       libpdbindir = $(bindir)
>>>     else
>>>       libpddir = $(libdir)/pd
>>>       libpdbindir = $(libpddir)/bin
>>>     endif
>>> endif
>>>
>>> objectsdir = $(libpddir)/extra
>>> pddocdir = $(libpddir)/doc
>>> helpdir = $(pddocdir)/5.reference
>>> ----------------
>>>
>>> For Pd-extended, there are three more:
>>> ----------------
>>> examplesdir = $(pddocdir)/examples
>>> manualsdir = $(pddocdir)/manuals
>>> readmesdir = $(manualsdir)/READMEs
>>> ----------------
>>>
>>> I hope that's clear enough, let me know if you have any specific
>>> questions...
>>>
>>>
>>> .hc
>>>
>>>
>>> _______________________________________________
>>> PD-dev mailing list
>>> PD-dev at iem.at
>>> http://lists.puredata.info/listinfo/pd-dev
>>>
>>>
>>
>> Thomas Grill
>> http://grrrr.org
>>
>>
>>



------------------------------------------------------------------------

If you are not part of the solution, you are part of the problem.






More information about the Pd-dev mailing list