Re: [Axiom-developer] Re: MacOS X development

[root]

Pierre,

> I had already tried to build axiom with 'linux' as a platform, and was
> able to correct the few problems you talk about in the past few
> messages of this thread. I also noticed the MACOSX target (as
> discussed in a previous thread in the mailing list) - which should be
> advertised by configure, by the way.

configure is just a shell script. If you look at the first four lines
they say:

# The sysname function uses uname -s to try to decode what kind of
# system to build. Currently the return value of uname is mapped as
#       Linux          --> linux
#       MINGW32_NT-5.1 --> windows

which is information generated by 'uname -s'

If you send me the output from that call I can fix the script.


> I'm now working with the axiom--MACOSX--1 branch, which seems to
> correct the libintl and SIGCHLD problems. Now my problem is in wct.c,
> with errors related to time.h. Quite weird because if I compile the
> same code somewhere else it just works...

Post the failure to address@hidden as see if anyone 
has a clue.

> My questions were more like : 
> - "how should the patches be made; i.e. in a gentle manner so that the
> MACOSX and main branches can be merged, or can they remain separate ?"
> If I understand well the past few messages, main and MACOSX have to be
> able to be merged at some point.

Patches are actually made by hand. I don't use a patching program.
The main point of a patch is that it needs to be documented. The
last merge I did, from axiom--windows, was a real horror show as
there was almost no documentation.

Basically I do a recursive diff, find every difference in every file
and look at the modification. If the mod makes sense (documentation
helps here) and the change seems reasonable I hand-apply the change
to the axiom--main branch, rebuild and test. If the change passes
testing then it remains. Rinse and repeat...

> - Is there a defined plans, or is there already some people working
> seriously on it, who have gone further in the process of building
> axiom in a Mac, but haven't commited it yet, so that I don't duplicate
> their work ?

Chuck Miller is the only other MAC person that I'm aware of.
He and I worked on his machine originally which is where the SIGCHLD
and libintl issues were first found. Not being a MAC person myself
I did not know how to install new software so the library issue was
a show-stopper for me.

As to defined plans.... my plan is to follow you and Chuck. When
you have a working version for the MAC and upload the changes
then I'll diff the axiom--MACOSX branch and merge it back into
the axiom--main branch.

Tim