openssh bugs - Oct 2023 - [Bug 3622] New: Please add an option to generate fish shell output

https://bugzilla.mindrot.org/show_bug.cgi?id=3622

            Bug ID: 3622
           Summary: Please add an option to generate fish shell output
           Product: Portable OpenSSH
           Version: 9.5p1
          Hardware: Other
                OS: Linux
            Status: NEW
          Severity: enhancement
          Priority: P5
         Component: ssh-agent
          Assignee: unassigned-bugs at mindrot.org
          Reporter: kemelzaidan at gmail.com

Please, add an argument (maybe -f since it's not in use) to generate
fish shell commands to the stdout just like there are options for C
shell and Bash.

Fish shell sets variable with the set command. -x for export and -g for
global scope, so you would need:

```sh
set -xg SSH_AUTH_SOCK /tmp/ssh-XXXXXXKjJOVJ/agent.199681
set-xg SSH_AGENT_PID 199682
echo Agent pid 199682
```
To unset the variables, use the -e flag like:

set -e SSH_AUTH_SOCK

You can take a look in the fish shell documentation here:
https://fishshell.com/docs/current/cmds/set.html

Thanks,

Kemel Zaidan

-- 
You are receiving this mail because:
You are watching the assignee of the bug.

https://bugzilla.mindrot.org/show_bug.cgi?id=3622

Damien Miller <djm at mindrot.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |djm at mindrot.org

--- Comment #1 from Damien Miller <djm at mindrot.org> ---
FYI fish appears to be almost as popular as tcsh

https://qa.debian.org/popcon.php?package=fish
https://qa.debian.org/popcon.php?package=tcsh

-- 
You are receiving this mail because:
You are watching someone on the CC list of the bug.
You are watching the assignee of the bug.

https://bugzilla.mindrot.org/show_bug.cgi?id=3622

--- Comment #2 from Kemel Zaidan <kemelzaidan at gmail.com> ---
just amending my report. You should use set -xU instead of -xg. -U
stands for universal, so every terminal has access to this variable.

So, I must understand that's an accepted feature request since there's
support for tcsh and that is as popular (or less) than fish... ;-)

-- 
You are receiving this mail because:
You are watching someone on the CC list of the bug.
You are watching the assignee of the bug.

https://bugzilla.mindrot.org/show_bug.cgi?id=3622

--- Comment #3 from Damien Miller <djm at mindrot.org> ---
AFAIK it's the reverse: tcsh is more popular than fish. Also there's
csh:

https://qa.debian.org/popcon.php?package=tcsh

which is another C-shell.

Re -U, the fish documentation says this:
> -U or --universal
> Sets a universal variable. The variable will be immediately available to
all > the user?s fish instances on the machine, and will be persist across
restarts > of the shell.
This would be a significant difference between how ssh-agent under fish
would work vs how it works with other shells. It seems superficially
convenient, but IMO it would be a bad idea as it would break things
like ssh forwarded agents.

(None of this is should be interpreted as an agreement to implement
fish support in ssh-agent, I'm just trying to bring some data to the
decision)

-- 
You are receiving this mail because:
You are watching the assignee of the bug.
You are watching someone on the CC list of the bug.

https://bugzilla.mindrot.org/show_bug.cgi?id=3622

Dmitry Belyavskiy <dbelyavs at redhat.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |dbelyavs at redhat.com

--- Comment #4 from Dmitry Belyavskiy <dbelyavs at redhat.com> ---
I'd say that if syntax for various shells is to be supported,  it's
worth having smth like

-f shellname

for better maintenance

-- 
You are receiving this mail because:
You are watching the assignee of the bug.
You are watching someone on the CC list of the bug.

https://bugzilla.mindrot.org/show_bug.cgi?id=3622

--- Comment #5 from Kemel Zaidan <kemelzaidan at gmail.com> ---
Sorry @Damien. I'm still learning about fish. It's looks you are write
and that I missed this details of persistence between restarts.

I believe my first interaction was right and global scope would be
enough for the intent, without any contraindication.

Thanks for the attention given.

-- 
You are receiving this mail because:
You are watching the assignee of the bug.
You are watching someone on the CC list of the bug.