[PATCH] Add inside-Emacs mode to GUI pinentry programs

[Neal H. Walfield]

At Fri, 12 Jun 2015 13:42:55 +0200,
Werner Koch wrote:
> On Fri, 12 Jun 2015 12:23, neal at walfield.org said:
> 
> > Is any change actually required on the pinentry side?  If gpg-agent
> > doesn't pass the INSIDE_EMACS environment variable through, then emacs
> > support is automatically disabled.
> 
> - I'd like to have this explicitly allowed.  Technically suppressing
>   INSIDE_EMACS is possible but it is a more code.  It also goes nicely
>   with the allow-external-cache option.

I disagree.  I see INSIDE_EMACS as a policy decision at the gpg-agent
level.  Passing (or not passing) the INSIDE_EMACS variable is similar
to an explicit delegation.  But, I could see how one would argue that
this is a pinentry policy.  In that case, I think the configuration
option should be in pinentry.conf and not gpg-agent.conf.

> - In case gpg-agent is started within Emacs, that envvar will be passed
>   anyway.  The special treatmeant of these envars affects is only fo gpg
>   and gpgsm to override the envvars known by gpg-agent if they want
>   different ones.  The best examples here are the DISPLAY or LC_* envars
>   which are often different than those use at the time gpg-agent was
>   started.

I understand better.  I thought they were to sanitize the environment.

Thanks,

:) Neal