To quote the first page of the introduction "... and there are also many developers who hate the Autotools - with a passion". I think that I can safely say that I started out in that camp. In short, I find that autotools projects too oftern do not live up to the aim of being cross platform, and more philosophically, that they encourage non-standard C/C++ "standard" libraries and compilers. But I'm not writing a review of a book about Autotools, not the Autotools themselves.
The book is organized as 2 introductory chapters, then 5 chapters on autoconf/automaka/libtool. That's followed by 2 chapters covering an example of converting a non-trivial project to use Autotools. The last two chapters cover M4, the macro language Autotools is based on, and then a catalog of tips. This isn't a reference book, you won't find an exhaustive list of all the Autotools macros.
Even if you are not going to use autotools extensively, I found the introduction to "make" very clear. In fact, as good an introduction as anything I've read. Though I've read the O'Reilly books on "make" and "GNU make", I still learned a few things about "make". If, like me, you are just an occasional user of Autotools, then you'll find it useful if the "configure" script doesn't work out of the box. As an example, I tried to build CUnit just after reading this book. I followed the "readme.txt", but it didn't work. Running one extra command and adding some command line options sorted it out. If you're going to be using Autotools extensively, then you'll obviously want to read this book.
Calcote does go out of his way to recommend platform-independent practices, even going as far as to accommodate make tools other than GNU make. Example code is marked with numbered dots, which are then referred to in the text. I found that very readable and easy to follow. In general the tone of the book is quite calm and neutral (compared to some authors on GNU software who seem to get, well, a bit over-excited about their subject).
The only weakness I felt was in the chapter on libtool and dynamic linking. Mac OS X ought to have been mentioned (perhaps it was omitted because of its peculiar "app bundles"). I think that the difference between UNIX dynamic libraries ought to have been mentioned (all global data/functions exported by default, like static libraries) and Windows dynamic libraries (only data/functions decorated with __declspec(dllexport) are exported, for performance reasons).
In the end, did this book soften my views on Autotools? Yes, a bit. It may not match my ideals, but I can't imagine that Autotools is going to be replaced any time soon. So I'm better off knowing about how to use Autotools.