On Tue, Apr 2, 2013 at 7:35 PM, Anton Vodonosov <[email protected]> wrote: > Suppose I want to configure ASDF output translations so that for all source > files under the ~/test/src/ directory the fasl files are placed into > ~/test/fasl/ directory. > > All other fasls are placed as usually, to the ~/.cache/common-lisp/ > > This is achieved by: > > (asdf:initialize-output-translations > `(:output-translations ("~/test/src" "~/test/fasl") > :inherit-configuration)) > > > That works. > > But what if I have source files in ~/test/fasl/ directory. Where their fasl > files should be placed, > near them in the ~/test/fals/ or in the default location > ~/.cache/common-lisp/ ? > The current ASDF implementation keeps such fasl files in ~/test/fasl/ > > My question, is it by design, or by accident? Because the cffi-grovel in the > latest cffi release relies on this behavior. > This is by design. The issue is to avoid double-translation, e.g. when some component creates a Lisp file from some other source, that is then compiled. You want to be able to use compile-op and load-op, except that the source-file will be the output of whichever create-lisp-op you're using, and therefore already translated. In general, being able to compose or not compose operations suggests that you shouldn't double-translate.
> I noticed this because test-grid-agent customizes output translations to place > fasls to a temporary directory in order to achieve fresh recompilation every > time. > And the way output translations are customized doesn't obey the expectation > of new cff-grovel. In the terms of the example above, any sources generated > and saved to > ~/test/fasl/ when compiled have their .fasl files in the default location - > ~/.cache/common-lisp/. > > So I am thinking, should I correct test-grid-agent output translations or > report it as a bug to CFFI. > I think you should correct test-grid-agent. Regards, —♯ƒ • François-René ÐVB Rideau •Reflection&Cybernethics• http://fare.tunes.org Why is there only one Monopolies Commission? — Lord Sutch _______________________________________________ asdf-devel mailing list [email protected] http://lists.common-lisp.net/cgi-bin/mailman/listinfo/asdf-devel
