tech-userlevel archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Re: individual software releases for third parties

>> I don't see the point of avoiding autotools - {cross,} compiling on
>> a variety of architectures is precisely what it is designed to do.
>Eh? Autoconf determines compile-time choices by probing the build
>host.  How is that a design for cross-compilation? 

That's not accurate.  If you tell autoconf that you're cross-compiling
then it will use the target cross-compilation system to perform the
autoconf tests.  This obviously isn't possible with all tests and you
have to be careful about what tests you use if you want to support
cross-compiling, but I never found it to be a huge problem.

I don't want to get into a huge debate about Autotools, but I did the work
to convert nmh to Automake recently so I feel I have enough experience with
it to have a meaningful opinion (long ago mh it used it's own build system,
but when it transitioned to nmh it was converted to autoconf, but not
by me).  To me, the autotools have the following advantages:

- They are well know.  Like Julio, I grit my teeth every time I have to build
  a package that doesn't use autoconf because I have to spend time figuring
  out how the build system works.  I also grit my teeth when there are bugs
  in an autoconf build system, but I view those as a fault of the developer,
  not autoconf.
- They are documented, a welcome change from most open-source projects!
- When I use them I spend less time hacking on the build system (which I hate)
  and more time hacking on the software (which I find vastly preferrable).

I understand why people aren't a fan of Autotools, and I'm not
saying they're perfect by any stretch of the imagination.  If
something better comes along then I'll gladly take a look at it and
switch if it truely is better.  That hasn't happened yet.


Home | Main Index | Thread Index | Old Index