Re: Excessive use of `eassert`

[Alan Mackenzie]

Hello, Paul.

On Mon, Jan 22, 2024 at 23:51:50 -0800, Paul Eggert wrote:
> On 2024-01-22 06:37, Alan Mackenzie wrote:
> > In a regular build, symbols with
> > position are handled differently depending on the value of
> > symbols_with_pos_enabled.  Or at least they were and they must be.  Thus
> > in XBARE_SYMBOL, one MUST test s_w_p_e, regardless of whether the build
> > is a regular one or a debugging one.

> None of my recent patches changed XBARE_SYMBOL. If XBARE_SYMBOL should 
> test something it isn't currently testing, then that is a bug that has 
> been present for some time.

> However, I don't see a bug there. XBARE_SYMBOL should be used only on 
> bare symbols, regardless of whether symbols_with_pos_enabled is set.

> Perhaps you meant to write XSYMBOL instead of XBARE_SYMBOL?

Yes I did.  Apologies.

> If so, I'm still not following, and more explanation would be helpful.
> What call to XSYMBOL formerly would have an eassert failure when
> debugging, but now won't have that failure?

It's not about the debug build, but a normal build, where easserts don't
play a role.

The scenario is when symbols_with_pos_enabled is false, and XSYMBOL is
invalidly given a symbol with position as argument.  Before your
patches, this would lead to an exception.  After your patches, this no
longer happens.  This is what I'm asking you to fix.

-- 
Alan Mackenzie (Nuremberg, Germany).