Re: Improving autoscan: help!

[Pavel Roskin]

Hello, Akim!

> Running autoscan on (configured) CVS Automake gives:
>
> warning: missing AC_CHECK_HEADERS([malloc.h]) wanted by: ansi2knr.c:157
> warning: missing AC_CHECK_HEADERS([stdlib.h]) wanted by: ansi2knr.c:150
> warning: missing AC_CHECK_HEADERS([string.h]) wanted by: ansi2knr.c:128 
> ansi2knr.c:143
> warning: missing AC_CHECK_HEADERS([strings.h]) wanted by: ansi2knr.c:130 
> ansi2knr.c:138
> warning: missing AC_FUNC_MALLOC wanted by: ansi2knr.c:160 ansi2knr.c:163 
> ansi2knr.c:279 ansi2knr.c:518

I really want that we disable the new functionality of autoscan (i.e.
checking the existing configure.in) by default. A separate option
"--check" will need to be specified to enable it.

The checking functionality on autoscan should be marked as experimental
and should be left in this state until Autoconf-2.50 is released.
Otherwise our generation will not see Autoconf-2.50 at all.

The recent changes in Autoconf prompted by this new crying feauture of
autoscan have broken several tests in Automake, such as ansi3.test because
@LDFLAGS@ is not substituted anymore. (This will be described in a
separate message.)

I share the guilt for that breakage because I said Ok to that patch, but
it's now time to stop and switch to testing Autoconf and improving the
documentation.

It will be much more fun to release Autoconf-2.50 and stop supporting
backward compatibility with Autoconf-2.13 in hundreds projects, than to
make last-minute fixes and discover new bugs.

Releasing stable Autoconf is our moral obligation before the community -
years of bugfixing should crystallize to a new stable version.

I'm sorry, Akim, that I have to write it, but I feel that we are doing
something wrong.

Regards,
Pavel Roskin