GOOPS-less Shepherd

GOOPS-less Shepherd

[Ludovic Courtès]

Hello!

I’d like to release the Shepherd 0.10.0 in a few weeks at most, with the
hope that it’ll be the last stable series before 1.0, which would be
released in a few months.

As part of this, I’d like to clean up the API, which includes removing
the dependency on GOOPS.  The Shepherd had been using GOOPS from the
start but for no particular reason: there’s no inheritance and only a
couple of cases of method overloading.

I started that work, which mostly involves renaming things like
accessors following typical Scheme conventions:

  https://git.savannah.gnu.org/cgit/shepherd.git/log/?h=wip-goopsless

I imagine the following deprecation scenario:

  • 0.10.x would still support GOOPS, as in (make <service> …), but
    that would be deprecated in favor of a more Schemey (service …)
    form.  Under the hood it’s still GOOPS.

  • 1.0.x would no longer use GOOPS at all.  We could provide a ‘make’
    macro so that (make <service> …) would still kinda work.

What’s at stake, mostly, is the ability to reconfigure a long-running
shepherd instance.  Once 1.0.x is in Guix, ‘guix system reconfigure’
will most likely fail to upgrade services on a pre-0.10.x shepherd:
users will have to reboot.

Thoughts?

Ludo’.

Re: GOOPS-less Shepherd

[Maxim Cournoyer]

Hi,

Ludovic Courtès <ludo@gnu.org> writes:

> Hello!
>
> I’d like to release the Shepherd 0.10.0 in a few weeks at most, with the
> hope that it’ll be the last stable series before 1.0, which would be
> released in a few months.
>
> As part of this, I’d like to clean up the API, which includes removing
> the dependency on GOOPS.  The Shepherd had been using GOOPS from the
> start but for no particular reason: there’s no inheritance and only a
> couple of cases of method overloading.
>
> I started that work, which mostly involves renaming things like
> accessors following typical Scheme conventions:
>
>   https://git.savannah.gnu.org/cgit/shepherd.git/log/?h=wip-goopsless
>
> I imagine the following deprecation scenario:
>
>   • 0.10.x would still support GOOPS, as in (make <service> …), but
>     that would be deprecated in favor of a more Schemey (service …)
>     form.  Under the hood it’s still GOOPS.
>
>   • 1.0.x would no longer use GOOPS at all.  We could provide a ‘make’
>     macro so that (make <service> …) would still kinda work.
>
> What’s at stake, mostly, is the ability to reconfigure a long-running
> shepherd instance.  Once 1.0.x is in Guix, ‘guix system reconfigure’
> will most likely fail to upgrade services on a pre-0.10.x shepherd:
> users will have to reboot.
>
> Thoughts?

Sounds reasonable to me.  Thank you for working on it!

-- 
Thanks,
Maxim

Re: GOOPS-less Shepherd

[Bodertz]

I don't have strong feelings either way, and the change won't really
affect me too much, but what benefit is there in breaking things?  From
what I understand from your message, users' configs will stop working in
a few months when 1.0.x releases (or with the macro would "kinda work"),
which is at least a short-term con, so what's the long-term benefit of
this change?  Is GOOPS so bad a thing to require?

Re: GOOPS-less Shepherd

[Ivan Sokolov]

Bodertz <bodertz@gmail.com> writes:

> I don't have strong feelings either way, and the change won't really
> affect me too much, but what benefit is there in breaking things?  From
> what I understand from your message, users' configs will stop working in
> a few months when 1.0.x releases (or with the macro would "kinda work"),
> which is at least a short-term con, so what's the long-term benefit of
> this change?  Is GOOPS so bad a thing to require?

If I understand correctly, that will allow Shepherd to run on GNU Mes,
a Scheme interpreter written for Guix bootstraping.

Re: GOOPS-less Shepherd

[Bodertz]

Thanks, makes sense.

Re: GOOPS-less Shepherd

[jbranso]

April 8, 2023 8:54 PM, "Ivan Sokolov" <ivan-p-sokolov@ya.ru> wrote:

> Bodertz <bodertz@gmail.com> writes:
> 
>> I don't have strong feelings either way, and the change won't really
>> affect me too much, but what benefit is there in breaking things? From
>> what I understand from your message, users' configs will stop working in
>> a few months when 1.0.x releases (or with the macro would "kinda work"),
>> which is at least a short-term con, so what's the long-term benefit of
>> this change? Is GOOPS so bad a thing to require?
> 
> If I understand correctly, that will allow Shepherd to run on GNU Mes,
> a Scheme interpreter written for Guix bootstraping.

So by removing the Shepherd's dependency on GNU Mes, the Guix project
benefits by simplying its bootstrapping process?

Re: GOOPS-less Shepherd

[Ivan Sokolov]

jbranso@dismail.de writes:

> April 8, 2023 8:54 PM, "Ivan Sokolov" <ivan-p-sokolov@ya.ru> wrote:
>
>> Bodertz <bodertz@gmail.com> writes:
>>
>>> I don't have strong feelings either way, and the change won't really
>>> affect me too much, but what benefit is there in breaking things? From
>>> what I understand from your message, users' configs will stop working in
>>> a few months when 1.0.x releases (or with the macro would "kinda work"),
>>> which is at least a short-term con, so what's the long-term benefit of
>>> this change? Is GOOPS so bad a thing to require?
>>
>> If I understand correctly, that will allow Shepherd to run on GNU Mes,
>> a Scheme interpreter written for Guix bootstraping.
>
> So by removing the Shepherd's dependency on GNU Mes, the Guix project
> benefits by simplying its bootstrapping process?

Where did this come from?  No, Mes is not a Shepherd's dependency,
please read the quotes again.

Re: GOOPS-less Shepherd

[Liliana Marie Prikler]

Am Donnerstag, dem 06.04.2023 um 22:16 +0200 schrieb Ludovic Courtès:
> What’s at stake, mostly, is the ability to reconfigure a long-running
> shepherd instance.  Once 1.0.x is in Guix, ‘guix system reconfigure’
> will most likely fail to upgrade services on a pre-0.10.x shepherd:
> users will have to reboot.
This caveat however, will only be a one-time thing, right?  That is,
once running Shepherd 0.10.x or 1.0.x, you can reconfigure without
needing to reboot supposedly until the next API break happening at
1.90.x or 2.0?

Cheers

Re: GOOPS-less Shepherd

[Ludovic Courtès]

Hi,

Bodertz <bodertz@gmail.com> skribis:

> I don't have strong feelings either way, and the change won't really
> affect me too much, but what benefit is there in breaking things?  From
> what I understand from your message, users' configs will stop working in
> a few months when 1.0.x releases (or with the macro would "kinda work"),
> which is at least a short-term con, so what's the long-term benefit of
> this change?  Is GOOPS so bad a thing to require?

It’s not that GOOPS is “bad”.  There are two things: some overhead (not
too bad, but still a bit), and a number of more or less subjective issues.

The programming style GOOPS encourages is not something I enjoy; I think
it makes it harder to reason about the code (you have to think about
applicable methods, new methods can be added at run time, etc.)

Another argument is consistency with Guix, and consistency within the
Shepherd (right now the GOOPS bits feel out of place).

Overall, I think we’d rather “clean things up” before 1.0 so we don’t
have to introduce breakage later.

I hope this clarifies the situation!

Ludo’.

Re: GOOPS-less Shepherd

[Ludovic Courtès]

Hi,

Liliana Marie Prikler <liliana.prikler@gmail.com> skribis:

> Am Donnerstag, dem 06.04.2023 um 22:16 +0200 schrieb Ludovic Courtès:
>> What’s at stake, mostly, is the ability to reconfigure a long-running
>> shepherd instance.  Once 1.0.x is in Guix, ‘guix system reconfigure’
>> will most likely fail to upgrade services on a pre-0.10.x shepherd:
>> users will have to reboot.
> This caveat however, will only be a one-time thing, right?  That is,
> once running Shepherd 0.10.x or 1.0.x, you can reconfigure without
> needing to reboot supposedly until the next API break happening at
> 1.90.x or 2.0?

Yes, exactly.

Ludo’.

Re: GOOPS-less Shepherd

[Christine Lemmer-Webber]

Ludovic Courtès <ludo@gnu.org> writes:

> Hello!
>
> I’d like to release the Shepherd 0.10.0 in a few weeks at most, with the
> hope that it’ll be the last stable series before 1.0, which would be
> released in a few months.
>
> As part of this, I’d like to clean up the API, which includes removing
> the dependency on GOOPS.  The Shepherd had been using GOOPS from the
> start but for no particular reason: there’s no inheritance and only a
> couple of cases of method overloading.

Neat!

"Obviously" Shepherd should use Spritely Goblins, which also does not
use GOOPS ;D

> I started that work, which mostly involves renaming things like
> accessors following typical Scheme conventions:
>
>   https://git.savannah.gnu.org/cgit/shepherd.git/log/?h=wip-goopsless

404s for me.

> I imagine the following deprecation scenario:
>
>   • 0.10.x would still support GOOPS, as in (make <service> …), but
>     that would be deprecated in favor of a more Schemey (service …)
>     form.  Under the hood it’s still GOOPS.
>
>   • 1.0.x would no longer use GOOPS at all.  We could provide a ‘make’
>     macro so that (make <service> …) would still kinda work.
>
> What’s at stake, mostly, is the ability to reconfigure a long-running
> shepherd instance.  Once 1.0.x is in Guix, ‘guix system reconfigure’
> will most likely fail to upgrade services on a pre-0.10.x shepherd:
> users will have to reboot.
>
> Thoughts?
>
> Ludo’.

Re: GOOPS-less Shepherd

[Ludovic Courtès]

Hello!

Christine Lemmer-Webber <cwebber@dustycloud.org> skribis:

> "Obviously" Shepherd should use Spritely Goblins, which also does not
> use GOOPS ;D

Yup!  I was thinking that a starting point would be to implement a
“bridge” as part of the new built-in “service collection” (under
shepherd/service/*.scm).

>> I started that work, which mostly involves renaming things like
>> accessors following typical Scheme conventions:
>>
>>   https://git.savannah.gnu.org/cgit/shepherd.git/log/?h=wip-goopsless
>
> 404s for me.

It’s merged now, and in 0.10.0rc1!  :-)

Ludo’.