GNU bug report logs - #77048
30.1; Macro `cl-load-time-value` malfunctioning at compile-time and runtime

Previous Next

Full log

Message #14 received at 77048 <at> debbugs.gnu.org (full text, mbox):

From: Stefan Monnier <monnier <at> iro.umontreal.ca>
To: Bohong Huang <bohonghuang <at> qq.com>
Cc: 77048 <at> debbugs.gnu.org
Subject: Re: bug#77048: 30.1; Macro `cl-load-time-value` malfunctioning at
 compile-time and runtime
Date: Sun, 16 Mar 2025 17:03:56 -0400

> When using `cl-load-time-value` from `cl-lib`, I encountered an issue
> where the macro fails during both byte-compilation and runtime:
>
> `reference to free variable ‘--cl-load-time--0’`  

I tried

    ;; -*- lexical-binding:t -*-
    
    (require 'cl-lib)
    
    (defun my-foo ()
      (time-convert (time-subtract (current-time)
                                   (cl-load-time-value (current-time)))
                    'integer))

and I did get the above warning at compile-time, but it did work
correctly after loading the `.elc` file.
I.e. `M-: (my-foo)` gives me the number of seconds since the file was loaded.

My crystal ball tells me that perhaps you're using `cl-load-time-value`
"at toplevel" rather than within a function?  In that case, indeed,
I think the macro will fail because it relies on

    (defun cl--compiling-top-level-p ()
      "Nil if we're in the middle of byte-compiling something."
      ;; ¡¡Yuck!!
      (not (and (fboundp 'byte-compile-file-form-defmumble)
                ;; FIXME: These are lexical now!
                (boundp 'bytecomp-this-kind) (boundp 'bytecomp-that-one))))

which has stopped working with:

    commit ca1055060d5793e368c1a165c412944d6800c3a6
    Author: Stefan Monnier <monnier <at> iro.umontreal.ca>
    Date:   Wed Mar 16 16:08:39 2011 -0400
    
        Remove bytecomp- prefix, plus misc changes.

which made `(bytecomp-)this-kind` into a lexically scoped variable.
So, this is a long-standing bug (of which I've been somewhat aware, as
the FIXME above suggests).
The current implementation of `cl-load-time-value` is fairly
ugly/hackish, so I'd rather not add yet more hideous hacks to it.

> I find this macro is quite useful for performing compile-time
> computations within a function body without polluting global variable
> names and relying on `cl-macrolet`.

Hmm... I don't quite follow: `cl-load-time-value` does not perform
compile-time evaluation but load-time evaluation, as its name suggests.
Maybe you could/should use `eval-when-compile` instead?

Can you show some examples of how you use `cl-load-time-value`?
Maybe some detail may give me an idea how to go about making it work
without making it worse.


        Stefan

Previous Next

GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.