freenode/#lisp - IRC Chatlog
Search
10:27:09
easye
I am not sure that this function is in the ASDF manual. I probably found it by inspecting the exported symbols of the ASDF package via M-x slime-apropos-package
10:27:47
myrkraverk
Just performed compiling #<MADEIRA-PORT "linedit" "ports" "ccl"> but failed to mark it done
10:30:08
easye
It seems to indicate that ASDF attempted the compilation of the "ccl.lisp" source unit, but failed in some way. If you are running under SBCL, it is probably difficult to compile the presumably CCL specific code in that file.
10:32:25
easye
I have no experience with linedit, so I cannot help you on that. You have the source: examine the ASDF system definitions. Look for assumptions about what feature conditionals are active under different implementations.
10:33:31
jackdaniel
myrkraverk: I've witnessed ASDF3's :if-feature being buggy, so if thre is :if-feature "ccl" you may try replacing it with good old #+ccl before (:file …)
10:41:37
antoszka
I'd be happy myself to know if somebody has investigated broken linedit on recent sbcl and has a solution :)
10:43:16
scymtym
iirc, linedit uses madeira-port to conditionally include certain files based on the lisp implementation (it probably predates :if-feature). madeira-port doesn't work right with recent ASDF versions. i seem to recall that somebody is already working in this, though
10:44:41
antoszka
that'd be cool, I really like using linedit as a quick handy alternative to firing up emacs + slime
11:21:27
myrkraverk
I'm actually a bit more likely to make a new frontend with cffi and libtecla than fix linedit; at least if fixing linedit is hard.
11:51:17
myrkraverk
jackdaniel: you mean the innards of it? I haven't dared to look at them, myself.
15:14:43
khrbt
myrkraverk: I have been sitting on this for a while. https://github.com/khirbat/linedit.git
17:41:07
stacksmith
Greetings. Need advice: at the end of my function I call a function that returns multiple values. I need to add an extra value at the end. Is there a way of doing that without m-v-b and a new values form?
17:42:50
Bike
i guess you could do (multiple-value-call #'values (call) other-value). but i would rate this as pretty weird
17:46:23
stacksmith
Since all this is happening at the end of the function, I thought there may be some magic way to append more...
19:51:22
fourier
also a question about asdf and quicklisp: if I have a system specified in defsystem-depends-on, and perform ql:quickload of my system, the system defined in defsystem-depends-on by some reason is not downloaded by ql
19:53:25
Bike
is (defstruct foo foos1 (foos2 foos1)) legal? i thought it was but now i think i was mistaken
20:04:00
sjl
> The symbols which name the slots must not be used by the implementation as the names for the lambda variables in the constructor function, since one or more of those symbols might have been proclaimed special or might be defined as the name of a constant variable.
20:05:03
sjl
If you're expecting foos2 to have an initform of foos1, and that that would evaluate to the keyword you passed for foos1, I think that bit of the spec explicitly forbids that
20:08:53
sjl
interestingly, sbcl and ccl DO use the symbols as the parameter names if you specify a BOA constructor
20:16:23
fourier
please recommend some simple documentation generation system from CL sources. I need something what can extract docstrings from functions and generate github markdown document. could write something myself but maybe something exists already
20:20:39
rme
I'm not trying to tell you what to do, but one option would be to write actual documentation, rather than fiddling around trying to generate it from docstrings and introspection.
20:21:22
Bike
the page on defstruct says the initforms are to be evaluated in the lxeical environment of the defstruct, which pretty much forbids that
20:21:51
sjl
Bike: yeah, I don't think that can work for the keyword constructors because they're supposed to use other symbols as the lambda arguments
20:26:46
sjl
> If no default value is supplied for an aux variable variable, the consequences are undefined if an attempt is later made to read the corresponding slot's value before a value is explicitly assigned. If such a slot has a :type option specified, this suppressed initialization does not imply a type mismatch situation; the declared type is only required to apply when the slot is finally assigned.
20:35:43
Bike
new question. if the clhs defines a function can an implementation have it be a generic function instead?