[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Autoconf 2.50

On Wed, May 23, 2001 at 11:13:28PM +0100, Colin Watson wrote:
> Gordon Sadler <gbsadler1@lcisp.com> wrote:
> >Seeing autoconf 2.50 in unstable today prompts this message. My
> >understanding is 2.50 will not work out-of-the-box with 2.13
> >configure.in's.
> Two different experiences:
>   * man-db, a fairly elderly configure.in that hasn't been kept up to
>     date for a while. The only things that broke were an undocumented
>     hack (no surprise there) and some instances of underquoting where
>     the new autoconf is less tolerant.
>   * groff, which is well-maintained upstream, yet whose build system
>     breaks quite badly with the new autoconf. It seems that the new
>     AC_PROG_CXX_EXIT_DECLARATION macro (called by AC_PROG_CC and
>     AC_PROG_CXX) dumps text containing '#ifdef __cplusplus' into
>     confdefs.h, which is then pulled into DEFS, which gets substituted
>     into output makefiles - so the build system tries to put #ifdef\
>     __cplusplus on a compiler command line, which of course breaks.
> The only bugs for man-db belong to the package. Do the bugs for groff
> belong to the package or to autoconf?

About man-db, i guess from the changelog that you could have removed
autoconf from Build-Depends since 2.3.18-1
Same with groff, is there a reason to run autoconf from debian/rules?

Only upstream authors should decide whether or not to upgrade to
autoconf 2.50, i do not understand why so many packages run
auto{conf,make} when compiling (or even Build-Depends on these tools
without invoking them).


Reply to: