Hi, all My build of Math::Random has encountered some hitches, and so I'm throwing this out there for comment. I'm not strong on MacOS, so maybe someone will see something I should be trying. :-) Math-Random-0.6 consists of 5 source files, and a header. As of earlier this week I thought I'd built it for PPC, and that there was some linking problem with CFM only. Well, the problem specific to CFM went away when I compiled a new MathLib with the far model, and I figured I was set. The PPC build tested out under MPW, and I schlepped the entire dist over to Chris. Fortunately he uses the MacPerl app much more than I do, and he tested under that. Dynaloader barfs on dl_load_file() with a -2807 Code Fragment Manager error, which is "fragHadUnresolveds", or "Loaded fragment has unacceptable unresolved symbols". I verified this result on my own machine later on. This is just under the app; the same library works with the tool in MPW. So I figure that GetDiskFragment() in dl_mac.xs is failing out. I went and looked at Random.exp. My first question: why is only the bootstrap, "boot_Math__Random", being exported? (This is all happening in file "wrapper.c", which is SWIG, BTW.) I would have thought that where Random.pm is using, for example, function "genbet" from wrapper.c, which is mapped in the bootstrap as newXS("Math::Random::genbet", _wrap_genbet, file); that one would want to export _wrap_genbet so that Random.pm has access to code for "genbet" after the bootstrap. Or does the bootstrap mysteriously understand Codewarrior export conventions? :-) So I tried exporting all of the functions from wrapper.c that are called in Random.pm. Other than the fact that everything still happily links, it doesn't fix the problem one iota. And of course, each new library variant still works in MPW. Just not under the app. Which baffles me, too. Finally I tried exporting *all* symbols, as an experiment - same thing. So I'm wondering what gives. What would really please me is if someone with the default build setup described in MPPE could give this module a whirl, and see if _they_ can build it for MacPerl (the app). This is pretty arcane, so bear with me. Thanks. It _is_ related to getting stuff built for MacPerl, so I think it's relevant. I haven't run into this particular problem before, and I'm thinking it's maybe related to the fact that SWIG was used, although the code appears to be OK. Apologies to Kevin L. for sending a faulty binary over - you're probably scratching your head over error -2807 right now. I figure I'll have this thing worked out within a week - I've got the bit in my teeth. :-) Thanks very much for any suggestions. Arved ***** Want to unsubscribe from this list? ***** Send mail with body "unsubscribe" to mac-perl-request@iis.ee.ethz.ch