Federico Beffa (2015-12-31 13:32 +0300) wrote:
> On Thu, Dec 31, 2015 at 11:17 AM, Alex Kost <alez...@gmail.com> wrote:
>> Federico Beffa (2015-12-30 20:33 +0300) wrote:
>>
>> [...]
>>> + (arguments
>>> + `(#:phases
>>> + (modify-phases %standard-phases
>>> + (add-before 'install 'configure
>>> + (lambda* _
>>> + (substitute* "slime.el"
>>> + (("inferior-lisp-program \"lisp\"")
>>> + "inferior-lisp-program \"sbcl\""))
>>> + #t))
>>
>> Why do you name this phase 'configure'? Actually I think this phase is
>> not needed, because 'inferior-lisp-program' should be set by a user to
>> whatever implementation (s)he prefers.
>
> It's called configure because I'm configuring a default CL implementation.
>
> The default value in slime is 'lisp'. I only used CLISP and SBCL and
> none of them provides a binary called 'lisp'. For this reason I tough
> it to be a good idea to provide a default name corresponding to an
> existing binary.
I don't agree, I think we shouldn't modify the upstream code just
because we think that some implementation is more likely to be used.
> Of course the user has to adapt to its own needs.
Yes, that's why an abstract "lisp" looks like a good default for me.
BTW instead of 'substitute*' it's better to use
'emacs-substitute-variables' for replacing variable values.
--
Alex
