Re: Save & load file
"Pascal J. Bourguignon" <[email protected]> Wed, 15 Jul 2015 22:58:26 +0200
| Newsgroups | gmane.lisp.slime.devel |
|---|---|
| Organization | Informatimago |
| Message-ID | <[email protected]> |
Stas Boukarev <[email protected]> writes: > Blake McBride <[email protected]> writes: > >> On Wed, Jul 15, 2015 at 6:48 AM, Attila Lendvai <[email protected]> wrote: >> >>> > I don't like to be forced to re-compile it in order to load it for the >>> > following reasons: >>> >>> ASDF solves most of these problems, including fasl file placement, >>> especially if you are willing to write a line or two to load your >>> codebase with some debugging extras (a (DECLAIM (OPTIMIZE DEBUG)), and >>> i also have some macros that react to variables, notably dribble level >>> logging stops being a no-op). >>> >>> among the things you listed the only thing that i don't know how to >>> solve in my ASDF/slime setup is losing track of what i've edited and >>> haven't given to the lisp for redefinition yet. what i do is i keep >>> track of it in my head, and whenever i suspect that things may be out >>> of sync, then i press 3 key combination to restart the lisp and >>> recompile/reload the project. >>> >>> hth, >>> >> >> So, it sounds like we can setup ASDF, learn a bunch of steps, add debugging >> code to our app, and just reset the whole world if we think we might have >> gotten confused, >> >> or, >> >> we can just use slime-save-and-load. >> >> Is that a fair statement? > compile-file has different semantics from load, so what you load may not > compile. Then why bother loading at all? C-M-x or C-c C-c is And load has different semantics from compile-file, sow what you compile may not load. Then why bother compiling at all? I too, use C-x C-l to load files from slime. And yesterday, I found something strage: when using C-x C-l apparently the functions I have in a (eval-when (:compile-toplevel :load-toplevel :execute) block weren't updated/redefined! I had to go (load "file.lisp") in the REPL… (I can't reproduce it now). > what you should use 98% of the time instead. Again, some implementations do a better job at debugging when you load than when you compile. > But however much I hate ASDF, if you have more than one file (or that > one file should be two files), you have really no other choice than to > use ASDF. When your modifications impact several other files, indeed, it might be a good idea to go asdf (or quickload). Perhaps we could have a few slime-asdf-load slime-asdf-load-source and slime-quickload commands, with a default current system determined semi-automatically? > And many implementations do have interpreters, not just compilers, in > 2015. -- __Pascal Bourguignon__ http://www.informatimago.com/ “The factory of the future will have only two employees, a man and a dog. The man will be there to feed the dog. The dog will be there to keep the man from touching the equipment.” -- Carl Bass CEO Autodesk