Convert::IBM390 -- functions for manipulating mainframe data
This distribution includes two versions of the code: Perl-only and Perl with XS and C. Ordinarily, you will probably want to install the latter (with XS), since the functions are written in C and run quite a bit faster. You, however, may prefer the Perl version -- if, for instance, you do not have access to the C compiler that was used to build Perl, or if you just prefer to have pure Perl code in your libraries.
If you want the Perl/XS version, begin thus:
perl Makefile.PL
If you want the Perl-only version, begin thus:
perl Makefile.PL noxs
The complete shtick:
perl Makefile.PL [noxs]
make
make test
make install
If your C compiler defaults to signed rather than unsigned char (many do), you may see some warnings about type mismatches, although I believe I have corrected all such problems. Anyhow, the mismatches are harmless. As long as the C code compiles correctly, the module should work.
If you have to build the module a second (third, fourth, ...) time, run "make clean" before running the above commands.
This module should work on most platforms, but if you have trouble with it (during "make test", for instance), you can compile the Perl/XS version with a debugging flag. This will produce a number of messages in stderr at run time.
The Perl-only version ignores this flag, but you can run it with the standard Perl debugger (-d) if you wish.
Note the following lines in Makefile.PL:
'DEFINE' => $p_define,
# 'DEFINE' => "$p_define -DDEBUG390", # if debugging is desired
To enable debugging, just comment out the first line (add a '' to the front) and uncomment the second line (delete the first ''). Then rebuild the module, starting with "make clean". E-mail me if you need help interpreting the debugging output.
Be sure to rebuild WITHOUT debugging before installing the module for general use.
Geoffrey Rommel
August 1999