[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: master 0a57dfcff8d 2/2: Ensure that specbind arg is always bare symb
|
From: |
Alan Mackenzie |
|
Subject: |
Re: master 0a57dfcff8d 2/2: Ensure that specbind arg is always bare symbol, and drop check |
|
Date: |
Sat, 20 Apr 2024 12:34:30 +0000 |
Hello, Mattias.
On Sat, Apr 20, 2024 at 09:29:24 +0200, Mattias Engdegård wrote:
> 20 apr. 2024 kl. 03.48 skrev Stefan Monnier <monnier@iro.umontreal.ca>:
> > Maybe I'm missing some context, but IIRC we had concluded that if sympos
> > reach actual evaluation it's a sign of a bug upstream.
> I certainly wouldn't mind using CHECK_BARE_SYMBOL_P but since the old
> code tolerated sympos as variable names in interpreted variable
> bindings, I just kept the semantics unchanged. This change was just a
> cycle shave.
How does that address Stefan's question? It seems to me it doesn't at
all.
Have you changed _any_ of the semantics, here?
It seems to me that when symbols_with_pos_enabled is non-nil, SWPs are
intended to be used as symbols, so provided the replacement is done only
in that case (which seems to be so), then the code will carry on working
according to its design.
--
Alan Mackenzie (Nuremberg, Germany).