[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: just one of a million reasons why autoconf is a worthless piece of s
Re: just one of a million reasons why autoconf is a worthless piece of shit
Tue, 18 Mar 2008 10:11:39 -0600
On Tue, Mar 18, 2008 at 9:05 AM, Dirk <address@hidden> wrote:
> I know pretty well what I'm talking about. I was using autoconf for some
> months to configure my projects but then switched to my own solution
> because autoconf was just bloated overkill... overkill that forced me to
> spend time with it. I'm happy I stopped using it before I wasted time
> digging into M4.
"some months..." I hesitate to say this because I know what your
response will be, but here goes:
No one becomes good at a tool set that does as much as
autoconf/automake/libtool do without a few years of effort. I've
worked with the autotools (on and off) for about 5 years now. I feel
pretty comfortable with them at this point, and I'm not an m4 expert.
I know a little m4, but only because I'm curious - and it does help to
be a bit curious.
I'm sure you'll say now that a tool shouldn't take 5 years to learn
and that it just proves your point, but what you're not getting is
that autoconf really is as complex as C++ within it's own domain. No
one becomes a C++ expert in a few months, but the value proposition is
similar between the two comparisons - autotools vs hand-coded
configure/make, and C++ vs C.
The autotools are NOT a set of tools designed to make maintainers'
lives easier (although they do in the long run) - they ARE a set of
tools to make a project's USERS' lives easier. This has been said time
and again, but people who hate autotools just don't get it.
We don't use the autotools because we want to go home early from work
today - we use them because we want our users' experiences to be as
simple as possible - configure; make and it just works - often even on
systems never tested by the maintainer. The guy who gets to type
"autoreconf; configure; make" has a tougher job, but because he
understands the value he's getting, he put's up with the pain.
Yes, the learning curve is steep, but once you reach the top of the
curve, it's really not that bad - and look at the value you've gotten.