[Ecls-list] Changes with new ASDF version?

[Jean-Pierre Flori]

2012/12/15 Faré <fahree at gmail.com>:
> On Fri, Dec 14, 2012 at 3:20 PM, Juan Jose Garcia-Ripoll
> <juanjose.garciaripoll at gmail.com> wrote:
>> On Fri, Dec 14, 2012 at 3:12 PM, Jean-Pierre Flori <jpflori at gmail.com>
>> wrote:
>>>
>>> 2012/12/14 Juan Jose Garcia-Ripoll <juanjose.garciaripoll at gmail.com>:
>>> > ECL now relies on ASDF-bundle, an utility library which is part of ASDF
>>> > and
>>> > which allows creating executables and big FASL  [...]
>>> I saw that.
>>
>> No, you did not read carefully what I said: ECL now relies on
>> ***ASDF-BUNDLE***. ASDF-Bundle is a different library from the one we used
>> for building executables (ASDF-ECL), and it lives as a separate project,
>> http://common-lisp.net/gitweb?p=projects/asdf/asdf-bundle.git though it will
>> probably become part of ASDF soon.
>>
> Yes, asdf-bundle was merged into asdf as of 2.26.7.
> It is based on asdf-ecl, of which it should retain the entire
> functionality on ecl,
> but some changes were made to make it work on all implementations
> (except obsolete ones, and abcl which has its own abcl-jar).
>
>>> What is strange is that I changed nothing but now get maxima-system.fasb
>>> instead of maxima.fasb.
>>
>> You did not change anything, but ASDF has changed. The new library follows a
>> different naming policy and, since I do not develop it, I have little to
>> say.
>>
> It should actually be maxima.system.fasb.
> The reason I added the .system is that several implementations rely on
> a fixed pathname-type for what is loaded as fasls, so to be portable,
> I had to add a suffix to the pathname-name so as to avoid conflicts
> between
> "foo.fasl" as output of (compile-file "foo.lisp") and "foo.fasl" as
> output of the final linking stage
> -- and there are many systems named "foo" that have a "foo.lisp".
Thanks for the info, that makes a lot of sense.
>
> If that's an issue, I could add a special case to NOT add this suffix
> to the pathname-name on ECL and MKCL, since you already use a
> different pathname type. I don't have strong feelings on this one, but
> unless someone does, I prefer a uniform naming scheme to a lot of
> ad-hoc exceptions.
For us its not an issue, we'll just change the name of the file we are
looking for, so don't feel obliged to do anything.
Or if you feel really bored, maybe add an option to change the suffix?
Anyway, as I said the easiest solution is to just change the name of
the file we are looking for, so we don't need any work from you.
The info you provided is far from sufficient!

Thanks a lot!
>
>>> So are you telling me that our command is fine, and that the produced
>>> file is identical, except for its name because ASDF decided that the
>>> new name reflects better what is produced as the previous name did?
>>> I'll be fine very fine with that.
>>
>> Indeed. The name just reflects the protocol and sets of rules that the new
>> library uses for building executables and libraries. I believe the suffix
>> can be eliminated but I do not know how to do it right now.
>>
> Most people never even see the .fasb files:
> either they are hidden deep under your ~/.cache/common-lisp/
> or they are otherwise delivered as part of someone's binary distribution.
>
> —♯ƒ • François-René ÐVB Rideau •Reflection&Cybernethics• http://fare.tunes.org
> Any time you're asking the user to make a choice they don't care about,
> you have failed the user — Jeff Atwood



-- 
Jean-Pierre Flori