[BUG] gitk assumes availability of `osascript` on macos

[BUG] gitk assumes availability of `osascript` on macos

[Robert Scott]

Hi,

Tiny one here. On macos, gitk (tested 2.50.1) attempts to call `osascript` so 
it can focus/raise the UI. But it fails hard if `osascript` isn't present in 
the $PATH.

There are a number of reasons `osascript` might not be available in a 
particular environment, e.g. if launched from inside a "pure" nix shell, but 
seeing as it's just a quality-of-life feature it would be nice if failures to 
call `osascript` could just be ignored.

Thanks,


robert.

Re: [BUG] gitk assumes availability of `osascript` on macos

[Eric Sunshine]

On Sat, Sep 20, 2025 at 8:07 AM Robert Scott <lists@humanleg.org.uk> wrote:
> Tiny one here. On macos, gitk (tested 2.50.1) attempts to call `osascript` so
> it can focus/raise the UI. But it fails hard if `osascript` isn't present in
> the $PATH.
>
> There are a number of reasons `osascript` might not be available in a
> particular environment, e.g. if launched from inside a "pure" nix shell, but
> seeing as it's just a quality-of-life feature it would be nice if failures to
> call `osascript` could just be ignored.

A patch[*] was posted to the mailing list years ago which, I believe,
should fix this problem for you. Unfortunately, the patch was never
picked up due to the then-maintainer of Gitk being missing-in-action.
Perhaps the current Gitk maintainer (Cc:'d) will pick up the patch.

[*]: https://lore.kernel.org/git/20180724065120.7664-1-sunshine@sunshineco.com/

Re: [BUG] gitk assumes availability of `osascript` on macos

[Johannes Sixt]

Am 20.09.25 um 21:52 schrieb Eric Sunshine:
> On Sat, Sep 20, 2025 at 8:07 AM Robert Scott <lists@humanleg.org.uk> wrote:
>> Tiny one here. On macos, gitk (tested 2.50.1) attempts to call `osascript` so
>> it can focus/raise the UI. But it fails hard if `osascript` isn't present in
>> the $PATH.
>>
>> There are a number of reasons `osascript` might not be available in a
>> particular environment, e.g. if launched from inside a "pure" nix shell, but
>> seeing as it's just a quality-of-life feature it would be nice if failures to
>> call `osascript` could just be ignored.
> 
> A patch[*] was posted to the mailing list years ago which, I believe,
> should fix this problem for you. Unfortunately, the patch was never
> picked up due to the then-maintainer of Gitk being missing-in-action.
> Perhaps the current Gitk maintainer (Cc:'d) will pick up the patch.
> 
> [*]: https://lore.kernel.org/git/20180724065120.7664-1-sunshine@sunshineco.com/

I've picked up the patch:
https://github.com/j6t/gitk/compare/master...es/ignore-osascript-failure

Please test and suggest improvements. Should the links in the commit
message be redirected to lore.kernel.org?

-- Hannes

Re: [BUG] gitk assumes availability of `osascript` on macos

[Eric Sunshine]

On Mon, Sep 22, 2025 at 2:37 AM Johannes Sixt <j6t@kdbg.org> wrote:
> Am 20.09.25 um 21:52 schrieb Eric Sunshine:
> > On Sat, Sep 20, 2025 at 8:07 AM Robert Scott <lists@humanleg.org.uk> wrote:
> >> Tiny one here. On macos, gitk (tested 2.50.1) attempts to call `osascript` so
> >> it can focus/raise the UI. But it fails hard if `osascript` isn't present in
> >> the $PATH.
> >>
> >> There are a number of reasons `osascript` might not be available in a
> >> particular environment, e.g. if launched from inside a "pure" nix shell, but
> >> seeing as it's just a quality-of-life feature it would be nice if failures to
> >> call `osascript` could just be ignored.
> >
> > A patch[*] was posted to the mailing list years ago which, I believe,
> > should fix this problem for you. Unfortunately, the patch was never
> > picked up due to the then-maintainer of Gitk being missing-in-action.
> > Perhaps the current Gitk maintainer (Cc:'d) will pick up the patch.
> >
> > [*]: https://lore.kernel.org/git/20180724065120.7664-1-sunshine@sunshineco.com/
>
> I've picked up the patch:
> https://github.com/j6t/gitk/compare/master...es/ignore-osascript-failure

Thank you. Looks good to me.

> Please test and suggest improvements. Should the links in the commit
> message be redirected to lore.kernel.org?

Updating the links would not be a bad idea. Here are the updated links
if you want to amend locally:

[1]: https://lore.kernel.org/git/D295145E-7596-4409-9681-D8ADBB9EBB0C@me.com/
[2]: https://lore.kernel.org/git/CABNJ2G+h3zh+=wLA0KHjUn8TsfhqUK1Kn-1_=6hnXVRJUPhuuA@mail.gmail.com/

Otherwise, I can resend.