bug#14121: Fwd: Eager macroexpansion failure in Emacs 24.3

bug#14121: Fwd: Eager macroexpansion failure in Emacs 24.3

[Michael Olson]

[-- Attachment #1.1: Type: text/plain, Size: 864 bytes --]

---------- Forwarded message ----------
From: Michael Olson <mwolson@gnu.org>
Date: Mon, Apr 1, 2013 at 11:40 AM
Subject: Eager macroexpansion failure in Emacs 24.3
To: Emacs Development Discussions <emacs-devel@gnu.org>


emacs --no-init-file --load ~/test-case.el

In *Messages* buffer:

Eager macro-expansion failure: (error "(erc-response\\.contents parsed) is
not a valid place expression")

I can work around this by putting (require 'erc) and the function
definition in different 'when' blocks, but it's not ideal.

The erc-response struct is defined in erc-backend.el, which is loaded by
erc.el.

Putting (require 'erc) inside of a 'when' statement along with code that
relies on functions provided by a struct defined by a file it loads seems
to trigger the problem.

-- 
Michael Olson  |  http://mwolson.org/



-- 
Michael Olson  |  http://mwolson.org/

[-- Attachment #1.2: Type: text/html, Size: 1524 bytes --]

[-- Attachment #2: test-case.el --]
[-- Type: application/octet-stream, Size: 545 bytes --]


(when t
  (require 'erc)
  (defun my-erc-remove-trailing-whitespace (proc parsed)
    "Remove trailing whitespace from the current message.
Some IM clients use an OTR plug-in that sends some annoying
trailing space to the screen, so we want to clean that up."
    (let ((msg (erc-response.contents parsed)))
      (when (stringp msg)
        (setf (erc-response.contents parsed)
              (erc-replace-regexp-in-string "[[:space:]]+\\'" "" msg))
        nil)))
  (add-hook 'erc-server-PRIVMSG-functions 'my-erc-remove-trailing-whitespace))

bug#14121: Fwd: Eager macroexpansion failure in Emacs 24.3

[Stefan Monnier]

tags 14121 notabug
thanks

> emacs --no-init-file --load ~/test-case.el
[...]
> In *Messages* buffer:
[...]
> Eager macro-expansion failure: (error "(erc-response\\.contents parsed) is
> not a valid place expression")

Not a big surprise:

   emacs24 -Q --batch -f batch-byte-compile test-case.el 
   
   In end of data:
   test-case.el:14:1:Warning: the following functions are not known to be
       defined: erc-response.contents, setf, erc-replace-regexp-in-string
   Wrote /home/monnier/tmp/test-case.elc

Emacs-24.3 tries to expand macros more eagerly (i.e. more like the
byte-compiler), so files that fail to byte-compile properly are also
likely to fail that eager macro-expansion.

But do note that the above message is somewhat harmless: when eager
macro-expansion fails, Emacs falls back on macro-expanding lazily as it
has always done.  IOW other than emitting the above message your code
should still work fine.


        Stefan

bug#14121: Fwd: Eager macroexpansion failure in Emacs 24.3

[Eli Zaretskii]

> From: Stefan Monnier <monnier@iro.umontreal.ca>
> Date: Wed, 03 Apr 2013 14:41:34 -0400
> Cc: 14121@debbugs.gnu.org
> 
> Emacs-24.3 tries to expand macros more eagerly (i.e. more like the
> byte-compiler), so files that fail to byte-compile properly are also
> likely to fail that eager macro-expansion.
> 
> But do note that the above message is somewhat harmless: when eager
> macro-expansion fails, Emacs falls back on macro-expanding lazily as it
> has always done.

Perhaps the failure message should mention that.

bug#14121: Fwd: Eager macroexpansion failure in Emacs 24.3

[Michael Olson]

[-- Attachment #1: Type: text/plain, Size: 826 bytes --]

If the message really is harmless (and no change to the code is
recommended, which seems to be the case), then it really ought to be
removed completely.

On Wednesday, April 3, 2013, Eli Zaretskii wrote:

> > From: Stefan Monnier <monnier@iro.umontreal.ca <javascript:;>>
> > Date: Wed, 03 Apr 2013 14:41:34 -0400
> > Cc: 14121@debbugs.gnu.org <javascript:;>
> >
> > Emacs-24.3 tries to expand macros more eagerly (i.e. more like the
> > byte-compiler), so files that fail to byte-compile properly are also
> > likely to fail that eager macro-expansion.
> >
> > But do note that the above message is somewhat harmless: when eager
> > macro-expansion fails, Emacs falls back on macro-expanding lazily as it
> > has always done.
>
> Perhaps the failure message should mention that.
>


-- 
Michael Olson  |  http://mwolson.org/

[-- Attachment #2: Type: text/html, Size: 1275 bytes --]

bug#14121: Fwd: Eager macroexpansion failure in Emacs 24.3

[Stefan Monnier]

> If the message really is harmless (and no change to the code is
> recommended, which seems to be the case),

No, the message does mean "you should fix your code".  It's just that we
additionally try to handle it right, for backward compatibility's sake.


        Stefan

bug#14121: Fwd: Eager macroexpansion failure in Emacs 24.3

[Michael Olson]

[-- Attachment #1: Type: text/plain, Size: 535 bytes --]

What is the "fix" then?  Using a macro in the same 'when' block that the
macro is defined in doesn't seem wrong to me.


On Wed, Apr 3, 2013 at 5:51 PM, Stefan Monnier <monnier@iro.umontreal.ca>wrote:

> > If the message really is harmless (and no change to the code is
> > recommended, which seems to be the case),
>
> No, the message does mean "you should fix your code".  It's just that we
> additionally try to handle it right, for backward compatibility's sake.
>
>
>         Stefan
>



-- 
Michael Olson  |  http://mwolson.org/

[-- Attachment #2: Type: text/html, Size: 1082 bytes --]

bug#14121: Fwd: Eager macroexpansion failure in Emacs 24.3

[Stefan Monnier]

> What is the "fix" then?

E.g. move the `require' to the top-level.

> Using a macro in the same 'when' block that the
> macro is defined in doesn't seem wrong to me.

It is, because macro expansion can take place (long) before the code is
executed, i.e. long before the macro is defined.


        Stefan