This is the mail archive of the mailing list for the GCC project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: committed: PR ada/14350 fix

And you'd think autoconf would be kind enough to generate an error...

It's okay, only a matter of style. The autoconf 2.5x style is what I hintd.

Anyway, I had a dream: one day, autoconf will be doomed and a real tool
properly designed and implemented will be used instead.

It will not be used, too much legacy alas. Just like those make replacements, jam or cons. The only build tool apart from make/autoconf/automake/libtool that is widely used is ant, but then, only for Java projects.

Whatever bad can be said of autoconf 2.5x, it also has many advantages, yet so many people are sticking with old, buggy and less portable autoconf/automake/libtools. Even gcc itself has started to move only recently, mostly thanks to Phil Edwards and Nathanael Nerode (in chronological order) -- and all the bad design decisions that were found are actually pretty unique to the gcc tree, so can only partly be blamed on the autoconf developers.

(I'm among those who think autoconf 2.5x should have been named autoconf 3)


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]