[EMAIL PROTECTED] (H�kon Alstadheim) writes: >> REQUIRE seriously only ever works halfway decently to load >> implementation-supplied stuff (if that is the documented way to do >> it), and even this usage is obviously implementation-dependent. > > Well, you have the clx thing again. What is implementation-supplied?
That which comes with the implementation, i.e. for CMUCL currently :clx, :clm, :gray-streams, :hemlock. >>> What was the reason for the -libraries change in 18d? >> >> The change was made in order to allow require to work as is to load >> the implementation-supplied libraries, i.e. CLX, CLM, gray-streams, >> etc. > > That might have been accomplished by changing the names of the fasl > files, but I guess the new names would be confusing. (i.e. clx.fasl in > the build-directory would not be the same as clx.fasl in the library). The major reason was that there were lots of scripts, etc. out there that depended on the *-library.* naming convention, so that in my (possibly wrong) estimation, more stuff would have broken this way than by changing the default pathname algorithm for require. People always had the chance to use require with a second argument, _and_ CMUCL also provides an implementation-dependent way to register modules with require. Another, even less defensible reason was that changing the implementation of require was easier than coordinating all distribution makers for CMUCL to change the naming/location of files in their distributions... Regs, Pierre. -- Pierre R. Mai <[EMAIL PROTECTED]> http://www.pmsf.de/pmai/ The most likely way for the world to be destroyed, most experts agree, is by accident. That's where we come in; we're computer professionals. We cause accidents. -- Nathaniel Borenstein
