At 06:49 PM 10/29/02 -0600, Craig A. Berry wrote: >At 10:28 AM -0600 10/29/02, Haining Yao wrote: > >>I've thought of just untar the file. But the thing was after untar, mms, >>mms test, without mms install, I put "use lib "[usr.perlmod]"" in the perl >>script, I had compilation problem saying that it couldn't locate date/calc >>module (for example, date/calc is the module I tried to use). I am not sure >>what's the problem. I can see the calc.pm module is under >>[usr.perlmod.date]. I guess it's about something with autoload or dynamic >>load? So I thought I have to install it locally, then hopefull I may not >>have compilation problem. I am not sure about this. > >The exact error message would be helpful since it tells you the >contents of @INC that it's actually using, not just what you think it >should be.
>From the error message, @INC was given and the [usr.perlmod] was one of the lib paths. I've tried and experienced same thing on Linux. If I didn't install module by using "PREFIX=localpath" and only untar, the same compilation error came out. But after I installed the modules locally in /tmp, and use lib /tmp/lib/site_perl/5.6.0/i386-linux, then no compilation problem. This made me think that I should also install the modules on VMS rather than just untar them. >Note that "[usr.perlmod]" is not an absolute path on VMS; you should >really specify a device or rooted logical name as well. If >everything is on the same disk, this probably isn't the immediate >cause of your problem. I also tried absolute path, disk$user:[usr.perlmod], same error message came out. >If you want verbose logging of what's happening while Perl loads >dynamic libraries, define a logical like so: > >$ define PERL_DL_DEBUG 5 > >(I'm not sure that a level higher than 2 or 3 means anything, but I >usually give it a big number just to make sure I see everything). > >If things get desperate, you can certainly install a complete copy of >Perl in your local directory structure where you have write access >and can install whatever modules you want to. It really shouldn't >come to that, though. > >-- >________________________________________ >Craig A. Berry >mailto:craigberry@;mac.com > >"... getting out of a sonnet is much more > difficult than getting in." > Brad Leithauser > >
