I assume the dirname "module.pmod" just was a coincidence; having a directory with that name would be pointless and will probably not work very well.
To answer your question: No, but this should be equivalent:
Path.pmod/To.pmod/Module.pmod/Class.so
and
Path.pmod/To.pmod/Module.pmod/___Class.so + Path.pmod/To.pmod/Module.pmod/Class.pike containing "inherit ___Class;":
/ Martin Stjernholm, Roxen IS
Previous text:
2003-11-11 19:42: Subject: Re: pike external module documentation
Based on the past statements, are the following two cases equivalent?
path.pmod/to.pmod/module.pmod/modulename.so
versus
path.pmod/to.pmod/___modulename.so
path.pmod/to.pmod/module.pmod/modulename.pike with the following contents:
inherit ___modulename;
If they were, it would certainly cut down on the amount of magic. especially if you can join a module.so with a module.pike (i think i remember someone saying this was possible.
Bill
On Tue, 11 Nov 2003, Per Hedbor () @ Pike (-) developers forum wrote:
Personally I find it to be irritating, if nothing else since it is now much harder to debug C-modules, you no longer get the errors from dlopen when they fail to load.
The reason is that .so-files now behave as .pike-files, not .pmod-files. This is somewhat more general, since its now possible to create C-level programs, not only modules, but the advantage is very small IMHO.
/ Per Hedbor ()
Previous text:
2003-11-11 16:16: Subject: Re: pike external module documentation
That's what I thought; though I hadn't had a chance to verify that. I guess the original question still stands (more for personal information than anything else, as it doesn't really matter to me) is what is the rationale? Is there a benefit, and if so, what is it?
Bill
Dynamic modules always get glue-files too now.
/ Per Hedbor ()
/ Brevbäraren
/ Brevbäraren