[Openmcl-devel] Naming anonymous functions

Anton Vodonosov avodonosov at yandex.ru
Sun Mar 2 14:04:55 UTC 2014


Sorry for annoyng the list with repeated proposals
of alexandria:named-lambda (I wanted
to ensure this approach is not rejected
by misunderstanding)

Now I see that Ron most likely doesn't have
an easy way to change the code which creates
the lambdas, hence creating named lambdas
is not possible.


Best regards,
- Anton

02.03.2014, 16:16, "Gary Byers" <gb at clozure.com>:
> On Sun, 2 Mar 2014, Anton Vodonosov wrote:
>
>>  CL-USER> (ccl::lfun-name (alexandria:named-lambda zu ()))
>>  ZU
>
> Ron already explained that he wasn't asking how to create functions
> with a specified name and was asking how to give names to functions
> that were anonymous when they were created.  Gail explained how to do
> what Ron asked about.  I assume that most interested parties get this
> by now.
>
> You're absolutely right (AFAICT) that ALEXANDRIA:NAMED-LAMBDA creates
> functions with a specified name.  If someone has asked how to do this,
> your answer would be both correct and useful.
>
> If someone asks how warm it is outside and I say "according to my
> watch, it's 4:14 AM", I've provided a correct answer to a question
> that no one asked.   We're all guilty of doing things like this
> from time to time; I don't know how to avoid that.
>
> If it's pointed out that we're answering a question that wasn't asked,
> we could either:
>
>    1) Try to determine the current temperature or admit that we don't
>       know that.
>    2) Continue to insist that the time was 4:14 AM, or that
>       ALEXANDRIA:NAMED-LAMBDA creates named functions.
>
> I'm sure that we're all guilty of (2) as well, but we can do
> something about this: the things that we can do include listening
> more carefully and thinking more clearly.   Please do those things
> as soon as possible so that this conversation can die an honorable
> death and we won't have to watch it lurching around violently like
> something from a bad zombie movie ...
>
> [For the terminally curiousL: many, many years ago, there were 3 types
> of functions in what eventually became CCL.  If I remember correctly:
>
>    CFUNs were implemented in C
>    SFUNs were interpreted functions
>    LFUNs were implemented by code produced by that newfangled compiler thing
>
> This is something like having 3 kinda of CONS cells.
>
> It took a long time for the other two categories to disappear, but
> LFUNs ("Lisp functions") survived.  This is relevant (to the extent
> that it is) in that some code that remains in CCL refers to functions
> as "lfuns", and there are still internal functions like
> #'CCL::LFUN-NAME.]



More information about the Openmcl-devel mailing list