kicad-developers team mailing list archive
-
kicad-developers team
-
Mailing list archive
-
Message #00618
Re: Building on Mac OS X and GNU autotools
-
To:
kicad-devel@xxxxxxxxxxxxxxx
-
From:
Dick Hollenbeck <dick@...>
-
Date:
Fri, 26 Oct 2007 13:47:00 -0500
-
In-reply-to:
<be4220fc0710261007se464b13w5756638f0a88515b@...>
-
User-agent:
Thunderbird 1.5.0.14pre (X11/20071023)
Hi Chris,
I also have 24 years of experience writing code and have been pretty
active in contributing free software to this project over the last 4
months or so.
I am not a fan of the gnu-auto tools. I find it too hard to maintain.
So since you seem to asking for opinions, my opinion is that you have
identified a solution before you have identified a problem, or problems.
There are alternatives to auto-tools, a series of makefiles (i.e. status
quo) being only one of them.
There is cmake, jam, etc. all as potential solutions to "the problem".
What is the problem you are trying to solve, "it won't build on the Mac"?
Then I would ask, please define "Mac", as this seems to be a moving
target lately. Maybe we should agree on a terminology to describe the
Mac platform that involves greater specificity.
In the end, you can contribute anything you want to this project I
suppose.
Again, my opinion is that I don't care for the gnu auto-tools. It is
based on a 15 year old idea that I think has been outdone by *simpler*,
newer alternatives. Also, the gnu auto-tools require a relatively
large infra-structure before they will even run, and as has been stated,
Windows is the most important platform for most users.
Another consideration is that most folks don't build the software, they
download and install binaries and run those. So your actual "user base"
size of this new auto-tools effort is considerably smaller than you
might think.
Certainly I do encourage you to improve the "Mac" support however, if
that is broken.
Dick
Hi,
I'm a new Kican user. I may be able to contribute some code but
first some questions:
1) What are the plans for use of GNU auto tools? Is this the way to
go? Is it being worked? If no one is doing this and people want it I
could add this.
I would implement it as a parallel build system where users could use
either the current makefile or do the ./configure thing. Later
remove the makefiles. Also the first implementation would just
duplicate the current makefiles and not add any functionality. Once
this is reliable it could be extended to actually do some checks.
2) I think the Mac port needs work. I tried building it under Mac OS
X and got a huge number of compile problems and warnings. It may be
that #1 above needs to be done first. I decided to try building on
Linux first. This is going better. My goal is the get Kicad working
on the Mac but I think I'll gain some experience with it on Linux
first. I have Ubuntu installed using VMware Fusion on my Mac
My day job is UNIX software development mostly on Solaris and Linux.
I'ce been at this for maybe 20+ years now. But I'm new to both kicad
and the Mac development environment
--
Best Regards,
Dick
*** If he runs, I will be supporting Tom Tancredo for President in 2008. ***
http://tancredo.house.gov/
http://www.teamtancredo.com/
http://tancredo4prez.blogspot.com/
Follow ups
References