Re: [elpa] externals/elisp-benchmarks fccb7bbaad 1/2: * Handle 'compilation-safety'

Re: [elpa] externals/elisp-benchmarks fccb7bbaad 1/2: * Handle 'compilation-safety'

[Stefan Monnier]

> +(unless (boundp 'compilation-safety)
> +  (defvar compilation-safety))

That makes no sense: (defvar compilation-safety) has no runtime effect,
it does not define the variable, it just tells the compiler that the
variable exists elsewhere and is dynamically scoped.  Furthermore, it
normally affects only the "current scope", so the effect of this
`defvar` should arguably end when we exit the `unless` (because of how
it's implemented it currently doesn't end when we exit the `unless`).


        Stefan

Re: [elpa] externals/elisp-benchmarks fccb7bbaad 1/2: * Handle 'compilation-safety'

[Andrea Corallo]

Stefan Monnier <monnier@iro.umontreal.ca> writes:

>> +(unless (boundp 'compilation-safety)
>> +  (defvar compilation-safety))
>
> That makes no sense: (defvar compilation-safety) has no runtime effect,
> it does not define the variable, it just tells the compiler that the
> variable exists elsewhere and is dynamically scoped.

Yes that was my intent, but I've got my test wrong and was sidetracked.
Anyway does 0eb498f77b looks better?

Thanks

  Andrea

Re: [elpa] externals/elisp-benchmarks fccb7bbaad 1/2: * Handle 'compilation-safety'

[Stefan Monnier]

>> That makes no sense: (defvar compilation-safety) has no runtime effect,
>> it does not define the variable, it just tells the compiler that the
>> variable exists elsewhere and is dynamically scoped.
> Yes that was my intent, but I've got my test wrong and was sidetracked.
> Anyway does 0eb498f77b looks better?

Yup, thanks!  🙂


        Stefan