Re: ASDF: module and system level dependencies don't work right?
Gary King <[email protected]>
| Newsgroups | gmane.lisp.cclan.general |
|---|---|
| Message-ID | <[email protected]> |
> I'm afraid I don't quite understand what you mean by dependencies > flowing. Dependencies don't exist (as far as ASDF knows) unless > you've specified them with :depends-on, right? If you don't want your > modules to depend on each other, you just don't say that they do. So > I guess I've misunderstood you? Oh, I'm probably just being incoherent again. <smile> I agree that dependencies only exist if you specify them. Once specified, they flow. So I think we're on the same page there. The discussion below isn't quite right and I've got to get the kids to school (!) but I think there is grist here. There are (at least) two kinds of dependencies: 1. This "thing" won't compile correctly unless this other "thing" is already compiled/loaded 2. This "thing" won't run unless this other thing is already compiled/ loaded. An example of the first is a macro definition or a defvar/ defparameter. An example of the second is a regular function. Currently, ASDF conflates these two dependencies and treats them differently. The (:file X :depends-on Y) is an example of the first kind; it says "recompile X if anything in Y changes". The dependencies between modules are like the second kind; it says that "before you load X, load Y" but doesn't do any reasoning about dates. Perhaps all of this can be made more explicit. P.S. Does anyone know why I receive at least two and sometimes more messages for every message sent to this list and/or where to look to fix it!! Thanks. -- Gary Warren King metabang.com http://www.metabang.com/
smime.p7s
(application/pkcs7-signature, 2.3 KB) - not displayed