Confusing Sed One-Liner in Makefile Tutorial

Can anyone explain this sed one-liner in English (the more detail, the better)?

@sed 's/\($*\)\.o[ :]*/\1.o $@ : /g' < $*.d > $@; \
             rm -f $*.d; [ -s $@ ] || rm -f $@

It's part of this tutorial: http://mad-scientist.net/make/autodep.html

I have a non-constant set of source files and want to auto-generate my dependency tree, based on the contents (includes) spelled out in my开发者_如何学编程 source files.

I was following the tutorial pretty well up until that...

P.S. I have basic understanding of sed select/replace, but I'm confused by the matching string and all the layers of redirection.... I've also read through the makefile tutorial once so have basic knowledge of standard makefiles...

The sed pattern will be processed by make first, so if the rule that it applies to is trying to build foo.P then $@ will be translated to foo.P and $* to foo. This means that the actual sed command will be something like:

sed 's/\(foo\)\.o[ :]*/\1.o foo.P : /g' < foo.d > foo.P

\(foo\) matches foo exactly and sets the first replacement to what matches (i.e. foo) \. matches a literal dot and [ :]* matches any number of spaces and colons.

As you can see the \1 replacement is a bit redundant as the matched string is fixed. This would have worked just as well.

sed 's/foo\.o[ :]*/foo.o foo.P : /g' < foo.d > foo.P

which could have been made from:

sed 's/$*\.o[ :]*/$*.o $@ : /g' < $*.d > $@