OK,
In the latest round of modifications to configure.com I have pulled
all of subconfigure.com into configure.com which allows me to ask
the "Do you want to edit config.sh now?" question via a GOSUB myread.
The procedure now obviates the need for subconfigure.com, though if
you do remove yours be sure to remove mention of it from your MANIFEST.
file.
While it is capable of determining if you have cxx and whether to use
LINK or CXXLINK the initial tests for CXX are turned off with
C<$ IF 1 .EQ. 0> because at this time I still have not successfully
run a CXX compiled munchconfig.exe (although my test set up may have been
to blame for that problem).
Configure.com will now ask if you want to edit the config.sh before
running munchconfig provided that you have not supplied the "-e" switch,
which now finally has a purpose on VMS :-)
I have removed the "%Config-I-VMS" type strings and the WRITE_RESULT
symbol from SYS$OUTPUT and made the output look a lot more like the
unix equivalent (IMO anyway).
Items that remain to be done now include:
1) moving all the install location symbol assignments such as
arch{bin|lib}{exp|install} stuff up into the questioning
about such matters so that user provided answers are not
overridden.
2) add call cxx_demangler_cleanup to end of previous subconfigure.com
portion and after compilation of munchconfig.exe C<$ IF Using_Cxx>.
3) create a [.vms]makefile.in for support of GNU make, and spiff up
the testing of exit values from GNU make so that it is picked up
correctly.
BTW does the [.vms] directory still have any 5.\d strings hardcoded in
anywhere? One of the Porting/ docs (pumpking.pod?) mentions the need
to look after them but I think we have them all zapped.
An update to myconfig.com is probably overdue as well.
At any rate go to:
http://www.best.com/~pvhp/vms/cfgperl_6409/
and follow the instructions awaiting you in the README if you'd like
to test it out.
Peter Prymmer
