Re: AC_CONFIG_FILES/AC_CONFIG_COMMANDS lossage

[Thien-Thi Nguyen]

   From: Akim Demaille <address@hidden>
   Date: Wed, 20 Nov 2002 18:34:21 +0100

   [docs]

i suppose it could be inferred from the docs that in the case where a
.in file needs to be processed in the standard way as well as w/
subsequent actions, it is better to use AC_CONFIG_FILES w/ the second
arg than to use two AC_CONFIG_FILES and then AC_CONFIG_COMMANDS.  or is
it, really?  would the specified commands (2nd arg) replace the standard
processing?  this is not clear to me, even after reading the docs.

a realistic example is augmenting makefiles:

  AC_CONFIG_FILES([b/c/d/Makefile],[
    echo extrathing: extradep >> b/c/d/Makefile
  ])
  ## (apparently this approach is not recommended)
  ##- AC_CONFIG_COMMANDS([b/c/d/Makefile],[
  ##-   echo extrathing: extradep >> b/c/d/Makefile
  ##- ])

at this time i don't know w/o experimenting if the result of running
configure would be a proper makefile or just the single line.  why not
add a sentence describing the interaction between the 2nd arg and the
"standard .in processing"?

thi