bug#72827: [PATCH] Adds NSSpeechRecognitionUsageDescription to Info.plist.in

[Alvaro Ramirez <alvaro@xenodium.com>]

Eli Zaretskii <eliz@gnu.org> writes:

>> Cc: copyright-clerk@fsf.org, 72827@debbugs.gnu.org
>> Date: Thu, 29 Aug 2024 08:57:25 +0800
>> From:  Po Lu via "Bug reports for GNU Emacs,
>>  the Swiss army knife of text editors" <bug-gnu-emacs@gnu.org>
>> 
>> Alvaro Ramirez <alvaro@xenodium.com> writes:
>> 
>> > Here's a tiny patch to add 
>> > NSSpeechRecognitionUsageDescription to
>> > Info.plist.in.
>> >
>> > Without NSSpeechRecognitionUsageDescription, we cannot launch
>> > processes that use macOS's speech recognition APIs (including 
>> > built-in
>> > shells).
>> >
>> > Here's an eshell example:
>> >
>> > $ macosrec --speech-to-text --input /tmp/audio.mp3
>> >  abort trap: 6
>> >
>> > With NSSpeechRecognitionUsageDescription in place, macOS 
>> > displays a
>> > dialog asking users to grant permission and allows the speech 
>> > APIs to
>> > be used from the process.
>> >
>> > Added change to Emacs 31 NEWS. If there's any chance the 
>> > permission
>> > can make it to 30, I'd be happy to rework the patch as 
>> > needed.
>> >
>> > Cheers,
>> 
>> What speech recognition services exist on Free (GNU/Linux) 
>> systems, and
>> do they require like intervention on our part to function in 
>> Emacs?
>
> Can you explain why this question is relevant to the patch?
>
> AFAIU, the patch does not invoke any services nor adds any 
> features to
> Emacs, it just avoids interfering with existing system features. 
> IOW,
> it makes Emacs behave according to the platform requirements and
> conventions.

Correct. No feature added. Purely host OS permission 
integration. If macOS determines the user must be asked for this 
permission to run a process, the patch enables the permission 
request to surface to the user. Without the patch, the process 
just fails (terminated by the OS).