John Cowan wrote:
Okay, I built the darcs head successfully on Cygwin with CMake,
removed the installed silex egg, and successfully installed
the easyffi egg. I left WITHOUT_LIBFFI in the default value, OFF
(as well as all other ccmake options).
It might be time to abandon Cygwin autotools builds.
I'm not surprised that the CMake builds are doing better on Windows.
But currently there's some bug on the Linux CMake build. I'm really not
a Linuxer and don't plan to be one. It could be a Linux-specific CMake
bug, or it could be that my reuse of pcre object files doesn't work as
well on Linux as on Windows. Either way, it's not an issue I plan to
chase down. Why is that...
Reality is, I've spent the past 3 years learning computer skills that
have enriched my mind and bolstered my confidence as a software
developer. I know I am capable of big things, having accomplished the
CMake build. But such efforts have left me broke and without
commercially valuable skills. Nobody's hiring me for computer work, and
it has been a poor season for signature gathering in Washington state
also. Consequently, I'm leaving the Seattle metro area and retreating
to the Carolinas, before I'm completely broke. My family is out there
and will be giving me some assistance. I'll probably be seeking some
kind of minimum wage work, whatever it takes to survive. The idea is
the Carolinas are cheaper than the Seattle metro area and I'll have some
help out there, unlike here. In my copious free time I plan to work on
things that have actual commercial value, i.e. that people will pay me
to do. Once I've done that, I may seek work in Research Triangle Park,
or elsewhere in the country. I may even return to Seattle, since all of
my stuff is in storage there.
Getting my career back online is going to be difficult. I don't see
Chicken or CMake as part of that effort; they are the problem. CMake, I
know well, and I use for my own work because I'm comfortable with it.
But nobody cares about that. Chicken, I do not know well, and I am not
comfortable with it. For me, it is still a R&D technology of uncertain
value to my personal projects. It may yet be of use to me someday, but
it is also clear to me that right now, it has no value at all. It
doesn't interface to 3D graphics and operating system capabilities any
better than anything else, and it doesn't make me money. I suspect I
could code for at least 6 months on my own projects, using only C++,
before I started feeling an actual pressing need for something more
"advanced." Maybe Scheme macros will be useful for something, but I've
yet to have a proven need. My poverty has gotten me beyond the point of
"gee this might be a good idea." It has to be a proven need.
So... what I'm willing to do for the CMake build anymore is very
limited. Probably only things that are dead easy. Chasing around weird
Linux bugs isn't, so I won't be doing them. The cost of whatever I do
for the CMake build has to be pretty close to zero. Anything that
resembles real work, I don't have the time for anymore.
Where does that leave you guys? Well, you've got a CMake build system
that works well on Windows, seems to work fine on MacOS X, and currently
has a bug on Linux. At some point you have to buy into the vision of 1
unified build system, instead of persisting with Autoconf. It's just
wasted effort. If anyone actually worked substantially on CMake other
than myself, would there be any Linux bugs? I seriously doubt it. I've
given you a build system that does 97% of the job. It's up to you to
make it 100%.
Cheers,
Brandon Van Every
_______________________________________________
Chicken-users mailing list
[email protected]
http://lists.nongnu.org/mailman/listinfo/chicken-users
