>>>>> "Bob" == Bob Friesenhahn <[EMAIL PROTECTED]> writes:
Bob> Autoconf is both a framework, and a collection of macros.
Absolutely. Maybe there is a factor I should have stressed more:
time. We don't have time now to integrate all these macros, we first
have to finish the work needed by the base layer.
Bob> While the framework is quite valuable, by itself it does not
Bob> achieve the objective of improving the quality/portability of
Bob> free software across many platforms. Inclusion of a macro in
Bob> autoconf frees the free software developer to focus on his/her
Bob> program rather than portability of a shell script fragment across
Bob> many platforms (which may not even be available to the
Bob> developer).
I'm sorry, but I don't quite agree: your picture lacks the Autoconf
macro archive. Its existence makes a huge difference with the former
state of Autoconf.
Bob> The value of a macro has little to do with whether it is
Bob> "specialized" or not.
If your measure is the number of macros which depend upon one macro,
yes it does. Currently this is my main concern. The base layer of
Autoconf is aged, it was not designed for some new problems, and we
first have to clean it up so that everything downstream can use a
saner ground. Until this is done, any work on terminal macros is
wasting time, and in particular, integrating new macros now would slow
down the development of Autoconf.
Bob> There are many specialized macros which are of value to a large
Bob> population of software projects. I suggest that the selection
Bob> criteria should be based on how many existing projects/developers
Bob> find the macro to be of value.
Definitely! I insist on the fact I was not saying that there will
never be new specific macros in Autoconf, I was just saying that there
won't be many *in 2.15*. We are in a delicate balance between urgency
and care.
Akim
