Re: [PATCH] [gdb/build] Fix adding -DNDEBUG to python flags of release build

[Tom de Vries <tdevries@suse.de>]

On 11/30/23 20:26, Tom Tromey wrote:
>>>>>> "Tom" == Tom de Vries <tdevries@suse.de> writes:
> 
> Tom> [ This was not obvious to me, but apparently evaluating an empty or undefined
> Tom> variable in this context is similar to using ':' or 'true', so the line is
> Tom> evaluated as:
> 
> TIL.
> 
> Tom> [ Unfortunately, the move might introduce issues similar to the one we're
> Tom> fixing, and I'm not sure how to check for this.  Shellcheck doesn't detect
> Tom> this type of problem.  FWIW, I did run shellcheck (using arguments -xa, in the
> Tom> src/gdb directory to make sure ../bfd/development.sh is taken into account)
> Tom> before and after and observed that the number of lines/words/chars in the
> Tom> shellcheck output is identical. ]
> 
> I don't think there's a good way to figure this out.  It's one of the
> problems with autoconf.
> 

There are cases where variables are either empty/undefined, or "yes", 
say variable GCC.

But this is one of the cases we know that the variable needs to be 
defined, so we can do something like:
...
+empty_var ()
+{
+    echo "EMPTY VAR, EXITING"
+    exit 1
+}
...

And then rewrite all references to use ${development:-empty_var}.

Then we can rewrite the GCC cases to be defined as either yes or no, and 
treat likewise.

Sounds like a major task though, and also cumbersome to maintain without 
any automation that warns you if you've added an unprotected variable 
reference.

Thanks,
- Tom

> If we do discover an issue, we may have to split the macro in two; or
> perhaps move more code into it.
> 
> Anyway I think this is ok.
> Approved-By: Tom Tromey <tom@tromey.com>
> 
> 
> Tom