bug#38529: Make --pure the default for `guix environment'?

bug#38529: Make --pure the default for `guix environment'?

[Pierre Neidhardt]

[-- Attachment #1: Type: text/plain, Size: 229 bytes --]

--pure seems to be the more sensible behaviour.  "Impure" environments
can have unexpected behaviours, so it makes sense to only allow them
when the user explicitly asks for it.

-- 
Pierre Neidhardt
https://ambrevar.xyz/

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 487 bytes --]

bug#38529: Make --pure the default for `guix environment'?

[zimoun]

Hi Pierre,

I agree.

And also change the default which populates by the dependencies. Something like:

  guix environment foo --inputs-of bar

should spawn an environment containing foo and the dependencies bar.
Well, keeping the --ad-hoc option for compatibility.


What do you think?

All the best,
simon

bug#38529: Make --pure the default for `guix environment'?

[Thompson, David]

Hi,

I have long thought that --ad-hoc should be implied, as that is the
mode I use 99% of the time, but I disagree that --pure should be the
default. Most of the time I (and I suspect most other users) just want
to temporarily augment the current environment with a package or two
and I think that shouldn't require any special flags (neither --ad-hoc
nor --pure).  The current default behavior of making an environment
from package dependencies is because that's how nix-shell worked (or
at least how I thought it worked) and 'guix environment' was created
as a clone of that tool.

- Dave

bug#38529: Make --pure the default for `guix environment'?

[Brett Gilio]

"Thompson, David" <dthompson2@worcester.edu> writes:

> Hi,
>
> I have long thought that --ad-hoc should be implied, as that is the
> mode I use 99% of the time, but I disagree that --pure should be the
> default. Most of the time I (and I suspect most other users) just want
> to temporarily augment the current environment with a package or two
> and I think that shouldn't require any special flags (neither --ad-hoc
> nor --pure).  The current default behavior of making an environment
> from package dependencies is because that's how nix-shell worked (or
> at least how I thought it worked) and 'guix environment' was created
> as a clone of that tool.
>
> - Dave

I was waiting for somebody to say this. But, I am 100% in agreement with
Dave as far as the behavior of --pure. I really would like to see either
nothing change, or we make --ad-hoc the implied default.

My reasoning confers with Dave's logic.

-- 
Brett M. Gilio
https://git.sr.ht/~brettgilio/

bug#38529: Make --pure the default for `guix environment'?

[Ludovic Courtès]

Hello!

"Thompson, David" <dthompson2@worcester.edu> skribis:

> I have long thought that --ad-hoc should be implied, as that is the
> mode I use 99% of the time, but I disagree that --pure should be the
> default.

I very much agree with that.  I don’t think ‘--pure’ should be the
default, because there are valid use cases for that.

As for ‘--ad-hoc’: making it the default is technically easy.  The
difficulty is to come up with a nice transition/deprecation mechanism so
that we don’t break everyone’s script overnight.

Ideas on how to achieve it are welcome!

Thanks,
Ludo’.

bug#38529: Make --pure the default for `guix environment'?

[raingloom]

On Tue, 2019-12-10 at 18:16 +0100, Ludovic Courtès wrote:
> As for ‘--ad-hoc’: making it the default is technically easy.  The
> difficulty is to come up with a nice transition/deprecation mechanism
> so
> that we don’t break everyone’s script overnight.
> 
> Ideas on how to achieve it are welcome!
> 
> Thanks,
> Ludo’.
> 
> 

Why not make it fully explicit, without either being the default?
It would make script more readable too.

Then the deprecation is as easy as printing a warning when the
invocation relies on the default.

For example `guix environment hello` would become `guix environment --
inputs-of hello`. Not using `--inputs-of` would print something like
"warning: implicit --inputs-of is deprecated".

bug#38529: Make --pure the default for `guix environment'?

[Christopher Lemmer Webber]

Ludovic Courtès writes:

> Hello!
>
> "Thompson, David" <dthompson2@worcester.edu> skribis:
>
>> I have long thought that --ad-hoc should be implied, as that is the
>> mode I use 99% of the time, but I disagree that --pure should be the
>> default.
>
> I very much agree with that.  I don’t think ‘--pure’ should be the
> default, because there are valid use cases for that.
>
> As for ‘--ad-hoc’: making it the default is technically easy.  The
> difficulty is to come up with a nice transition/deprecation mechanism so
> that we don’t break everyone’s script overnight.
>
> Ideas on how to achieve it are welcome!
>
> Thanks,
> Ludo’.

It suddenly struck me today that there is an easy way to change the
default behavior while supporting the legacy behavior.

How about we have a new command, "guix env", what is --ad-hoc by
default?  Then "guix environment" sticks around as legacy for supporting
the old interface.  Therefore, "guix env" is doubly short and to the
point.  :)

bug#38529: Make --pure the default for `guix environment'?

[zimoun]

Hi,
On Tue, 03 Nov 2020 at 12:38, Christopher Lemmer Webber <cwebber@dustycloud.org> wrote:

>> Ideas on how to achieve it are welcome!

[..]

> It suddenly struck me today that there is an easy way to change the
> default behavior while supporting the legacy behavior.
>
> How about we have a new command, "guix env", what is --ad-hoc by
> default?  Then "guix environment" sticks around as legacy for supporting
> the old interface.  Therefore, "guix env" is doubly short and to the
> point.  :)

Rehashing all that, and because different POV on CLI pops up sometime,
one move should be to have “aliases” [1]; I am sure the idea already
came long time before but I do not find the thread.


My understanding of the situation:

 1- specific "guix foo" case: implement something like Alice proposed
 2- add a mechanism to have aliases.

Replace ’foo’ by ’environment’, ’graph’, ’search’, ’show’, etc.


Let's expand explanations about the #2.

Now with “guix repl“, the user can extend Guix by their own scripts.
Therefore, it could be nice:

  a) to have a location by default (say ~/.config/guix/scripts)
  b) to run them with "guix foo" instead of
  "guix repl -- ~/.config/guix/scripts/foo.scm".

And then, we are all be happy. ;-)

It eases the tests of new experimental command-line tools, one can locally
change the behaviour of “guix environment --ad-hoc“, one can experiment
with new output formats for “guix search“ instead of ’recutils’, etc..

And using channel as “home-manager” or GWL [2] show, we could even
easily exchange them.  Somehow.

Well, the philosophy of custom extensions.


WDYT?

All the best,
simon



1: <http://issues.guix.gnu.org/issue/43477#11>
2: related recent patch by Ricardo:
<https://lists.gnu.org/archive/html/guix-devel/2020-10/msg00505.html>

bug#38529: Make --pure the default for `guix environment'?

[Ludovic Courtès]

Hi!

Christopher Lemmer Webber <cwebber@dustycloud.org> skribis:

> It suddenly struck me today that there is an easy way to change the
> default behavior while supporting the legacy behavior.
>
> How about we have a new command, "guix env", what is --ad-hoc by
> default?  Then "guix environment" sticks around as legacy for supporting
> the old interface.  Therefore, "guix env" is doubly short and to the
> point.  :)

Last time we discussed it we reached (I think!) a consensus that the new
command would be called ‘guix shell’ and be ‘--ad-hoc’ by default.

I almost started working on it, and then got stuck wondering whether
‘--inputs-of’ should be implemented as a package transformation or not.

Ludo’.

bug#38529: Make --pure the default for `guix environment'?

[Taylan Kammer]

On 10.12.2019 18:16, Ludovic Courtès wrote:
> Hello!
> 
> "Thompson, David" <dthompson2@worcester.edu> skribis:
> 
>> I have long thought that --ad-hoc should be implied, as that is the
>> mode I use 99% of the time, but I disagree that --pure should be the
>> default.
> 
> I very much agree with that.  I don’t think ‘--pure’ should be the
> default, because there are valid use cases for that.
> 
> As for ‘--ad-hoc’: making it the default is technically easy.  The
> difficulty is to come up with a nice transition/deprecation mechanism so
> that we don’t break everyone’s script overnight.
> 
> Ideas on how to achieve it are welcome!
> 
> Thanks,
> Ludo’.
> 

I suppose the straightforward way of doing such a transition would be:

- In the next release, add --inputs-of and make the use of 'guix 
environment' with neither switch emit a warning saying you should use 
the new switch to explicitly enable it because the default will change.

- Later, change the default behavior and deprecate --ad-hoc as it 
becomes a no-op.

- (Optional) In some future release, remove --ad-hoc.


Taylan

bug#38529: Make --pure the default for `guix environment'?

[Christopher Lemmer Webber]

Taylan Kammer writes:

> On 10.12.2019 18:16, Ludovic Courtès wrote:
>> Hello!
>> "Thompson, David" <dthompson2@worcester.edu> skribis:
>> 
>>> I have long thought that --ad-hoc should be implied, as that is the
>>> mode I use 99% of the time, but I disagree that --pure should be the
>>> default.
>> I very much agree with that.  I don’t think ‘--pure’ should be the
>> default, because there are valid use cases for that.
>> As for ‘--ad-hoc’: making it the default is technically easy.  The
>> difficulty is to come up with a nice transition/deprecation mechanism so
>> that we don’t break everyone’s script overnight.
>> Ideas on how to achieve it are welcome!
>> Thanks,
>> Ludo’.
>> 
>
> I suppose the straightforward way of doing such a transition would be:
>
> - In the next release, add --inputs-of and make the use of 'guix
>   environment' with neither switch emit a warning saying you should
>  use the new switch to explicitly enable it because the default will
> change.
>
> - Later, change the default behavior and deprecate --ad-hoc as it
>   becomes a no-op.
>
> - (Optional) In some future release, remove --ad-hoc.
>
>
> Taylan

That sounds good to me.

And could we make "env" a shortcut for "environment" anyway?  My fingers
are tired. ;)

bug#38529: Make --ad-hoc the default for guix environment proposed deprecation mechanism

[Gábor Boskovits]

Hello guix,

Based on discussion on IRC the following plan for deprecation might work.
Comments are welcome:

1.
Make guix environment aware of an environment variable:
GUIX_ENVIRONMENT_DEPRECATED_AD_HOC
if this is defined, then fall back to the current behaviour.

Motivation: this would enable most script users to bypass the problem,
while fixing them, but it makes the users aware that they are using a
deprecated feature.
At the same time this should come with a news entry, an any other
announcements should be made that we usually do, so that support don't
get overloaded. Is should also be announced that two releases later
the code supporting this will be removed, so that we don't have to
maintain it, but allow enough time for adoption.

2. add a flag to guix environment, something like
--ignore-deprecated-ad-hoc, that makes guix environment ignore the
environment variable, and default to the new behaviour.

Motivation: so that scripts can be fixed individually by modifying the
guix environment call to the new version, and adding the flag, so that
it does not cause a problem in the trasitional period while the
environment variable is defined.

3. on the specified release remove the environment variable support
code, and make the flag a noop, and also deprecated.

4. later if needed after an adoption period we can remove the flag.


Best regards,
g_bor
-- 
OpenPGP Key Fingerprint: 7988:3B9F:7D6A:4DBF:3719:0367:2506:A96C:CF63:0B21

bug#38529: Make --ad-hoc the default for guix environment proposed deprecation mechanism

[zimoun]

Hi Gábor,

Thank you for summarizing the discussion on IRC that I missed.

Maybe I miss a point. Is the aim to conserve the "--ad-hoc" option
with a different effect? Or why do we want to conserve this option
name?
It appears to me simpler to give another name, for example
"--inputs-of". And it is more meaningful.


To be concrete, the different cases; (-) means current behavior and
(+) the new one:

1.
- guix environment foo
+ guix environment --inputs-of foo

2.
- guix environment --ad-hoc bar
+ guix environment bar


First, when "--ad-hoc" is used then it reports a warning: deprecated
option and falls in the current behavior.
When "--inputs-of" is used then it falls in the new behavior.
Therefore, no needs of the ugly "--ignore-deprecated-ad-hoc".


In other words, with the same future guix version,

 # Alice
 $ guix environment foo --ad-hoc bar
 Warning: deprecated... explanations...
               instead use:
                guix environment bar --inputs-of foo

 # Bob
 $ guix environment bar --inputs-of foo


Second, the previous "guix environment foo" (dependencies of foo) is
inconsistent with the new "guix environment bar" (only the package
bar). Therefore, let introduce the GUIX_ENVIRONMENT_DEPRECATED
variable to distinguish both, as you said.

 # Alice
 $ guix environment foo
 Warning: previous behavior requires GUIX_ENVIRONMENT_DEPRECATED=1
               turn off the warning: GUIX_ENVIRONMENT_NOWARNING=1

And Alice has now a new shell with the package foo. If she wants the
dependencies, she has two options:

$ GUIX_ENVIRONMENT=1 guix environment foo
or
$ guix environment --inputs-of foo


 # Bob
 $ guix environment bar
 Warning: previous behavior requires GUIX_ENVIRONMENT

And if Bob is annoyed by the warnings each time, he globally turns off
with the variable GUIX_ENVIRONMENT_NOWARNING=1.


Couple of months later -- after the period adoption -- we remove the
variables GUIX_ENVIRONMENT_NOWARNING and GUIX_ENVIRONMENT_DEPRECATED;
still keeping the warning with the "--ad-hoc" option. And then, after
we can remove the "--ad-hoc" option if required.


Maybe a miss a point. But the addition of the flag appears
"--too-long-to-type" to me ugly.


What do you think?

All the best,
simon

bug#38529: Make --ad-hoc the default for guix environment proposed deprecation mechanism

[Gábor Boskovits]

[-- Attachment #1: Type: text/plain, Size: 3380 bytes --]

Hello,

zimoun <zimon.toutoune@gmail.com> ezt írta (időpont: 2019. dec. 12., Csü
17:47):

> Hi Gábor,
>
> Thank you for summarizing the discussion on IRC that I missed.
>
> Maybe I miss a point. Is the aim to conserve the "--ad-hoc" option
> with a different effect? Or why do we want to conserve this option
> name?
> It appears to me simpler to give another name, for example
> "--inputs-of". And it is more meaningful.
>
Sorry for the confusion. Ad-hoc should be retained with the same effect, so
that we do not break existing scripts.
Renamin the option would be ok. It even makes sense to me.

>
>
> To be concrete, the different cases; (-) means current behavior and
> (+) the new one:
>
> 1.
> - guix environment foo
> + guix environment --inputs-of foo
>
> 2.
> - guix environment --ad-hoc bar
> + guix environment bar
>
>
> First, when "--ad-hoc" is used then it reports a warning: deprecated
> option and falls in the current behavior.
> When "--inputs-of" is used then it falls in the new behavior.
> Therefore, no needs of the ugly "--ignore-deprecated-ad-hoc".
>
That could be done. The problem is caused by uses of guix environment that
does not use any of these options. Those mean different things after the
change.

>
>
> In other words, with the same future guix version,
>
>  # Alice
>  $ guix environment foo --ad-hoc bar
>  Warning: deprecated... explanations...
>                instead use:
>                 guix environment bar --inputs-of foo
>
>  # Bob
>  $ guix environment bar --inputs-of foo
>
>
> Second, the previous "guix environment foo" (dependencies of foo) is
> inconsistent with the new "guix environment bar" (only the package
> bar). Therefore, let introduce the GUIX_ENVIRONMENT_DEPRECATED
> variable to distinguish both, as you said.
>
Ok.

>
>  # Alice
>  $ guix environment foo
>  Warning: previous behavior requires GUIX_ENVIRONMENT_DEPRECATED=1
>                turn off the warning: GUIX_ENVIRONMENT_NOWARNING=1
>
> And Alice has now a new shell with the package foo. If she wants the
> dependencies, she has two options:
>
> $ GUIX_ENVIRONMENT=1 guix environment foo
> or
> $ guix environment --inputs-of foo
>
>
>  # Bob
>  $ guix environment bar
>  Warning: previous behavior requires GUIX_ENVIRONMENT
>
> And if Bob is annoyed by the warnings each time, he globally turns off
> with the variable GUIX_ENVIRONMENT_NOWARNING=1.
>
>
> Couple of months later -- after the period adoption -- we remove the
> variables GUIX_ENVIRONMENT_NOWARNING and GUIX_ENVIRONMENT_DEPRECATED;
> still keeping the warning with the "--ad-hoc" option. And then, after
> we can remove the "--ad-hoc" option if required.
>
>
> Maybe a miss a point. But the addition of the flag appears
> "--too-long-to-type" to me ugly.
>
We could recommend simply to use something like:
GUIX_ENVIRONMENT_DEPRECATED=0 guix environment ...
Instead in existing scripts that are fixed to use the new syntax. This
indeed looks like a better solution, and it is less of a maintenance
burden. Good idea.

>
>
> What do you think?
>
> All the best,
> simon
>

Summarizing:
Introduce the environment variable.
For fixed scripts recommend unsetting the environment variable.

That looks like a better plan. Thanks for your insights.

Best regards,
g_bor

>

[-- Attachment #2: Type: text/html, Size: 5119 bytes --]

bug#38529: Make --ad-hoc the default for guix environment proposed deprecation mechanism

[zimoun]

Hi Gábor,


On Thu, 12 Dec 2019 at 21:54, Gábor Boskovits <boskovits@gmail.com> wrote:

> zimoun <zimon.toutoune@gmail.com> ezt írta (időpont: 2019. dec. 12., Csü 17:47):

>> Maybe I miss a point. Is the aim to conserve the "--ad-hoc" option
>> with a different effect? Or why do we want to conserve this option
>> name?
>> It appears to me simpler to give another name, for example
>> "--inputs-of". And it is more meaningful.
>
> Sorry for the confusion. Ad-hoc should be retained with the same effect, so that we do not break existing scripts.
> Renamin the option would be ok. It even makes sense to me.

What I propose is:

  - keep the option "--ad-hoc" with the current behavior; so same effect
  - add a new option "--inputs-of" with the new behavior; name more meaningful
  - and two env variables; to not break existing scripts


>> First, when "--ad-hoc" is used then it reports a warning: deprecated
>> option and falls in the current behavior.
>> When "--inputs-of" is used then it falls in the new behavior.
>> Therefore, no needs of the ugly "--ignore-deprecated-ad-hoc".
>
> That could be done. The problem is caused by uses of guix environment that does not use any of these options. Those mean different things after the change.

The transition to such use-case was described below with the
introduction of 2 env variables. :-)


>>  # Alice
>>  $ guix environment foo --ad-hoc bar
>>  Warning: deprecated... explanations...
>>                instead use:
>>                 guix environment bar --inputs-of foo
>>
>>  # Bob
>>  $ guix environment bar --inputs-of foo
>>
>>
>> Second, the previous "guix environment foo" (dependencies of foo) is
>> inconsistent with the new "guix environment bar" (only the package
>> bar). Therefore, let introduce the GUIX_ENVIRONMENT_DEPRECATED
>> variable to distinguish both, as you said.
>
> Ok.

It is the easy part. ;-)


Now the hard part: avoid to break existing scripts.

>>  # Alice
>>  $ guix environment foo
>>  Warning: previous behavior requires GUIX_ENVIRONMENT_DEPRECATED=1
>>                turn off the warning: GUIX_ENVIRONMENT_NOWARNING=1
>>
>> And Alice has now a new shell with the package foo. If she wants the
>> dependencies, she has two options:
>>
>> $ GUIX_ENVIRONMENT=1 guix environment foo
>> or
>> $ guix environment --inputs-of foo
>>
>>
>>  # Bob
>>  $ guix environment bar
>>  Warning: previous behavior requires GUIX_ENVIRONMENT
>>
>> And if Bob is annoyed by the warnings each time, he globally turns off
>> with the variable GUIX_ENVIRONMENT_NOWARNING=1.
>>
>>
>> Couple of months later -- after the period adoption -- we remove the
>> variables GUIX_ENVIRONMENT_NOWARNING and GUIX_ENVIRONMENT_DEPRECATED;
>> still keeping the warning with the "--ad-hoc" option. And then, after
>> we can remove the "--ad-hoc" option if required.


> We could recommend simply to use something like:
> GUIX_ENVIRONMENT_DEPRECATED=0 guix environment ...
> Instead in existing scripts that are fixed to use the new syntax. This indeed looks like a better solution, and it is less of a maintenance burden. Good idea.

My point is: the new variable GUIX_ENVIRONMENT_DEPRECATED should only
be used by the scripts that call "guix environment pkg" without the
options "--ad-hoc" or "--inputs-of". And I think that it represents
really few scripts in real life. :-)


> Summarizing:
> Introduce the environment variable.
> For fixed scripts recommend unsetting the environment variable.

I am not to get your plan. :-)


Cheers,
simon

bug#38529: Make --ad-hoc the default for guix environment proposed deprecation mechanism

[Gábor Boskovits]

Hello,

Let me try again :)

zimoun <zimon.toutoune@gmail.com> ezt írta (időpont: 2019. dec. 13., P, 13:02):
>
> Hi Gábor,
>
>
> On Thu, 12 Dec 2019 at 21:54, Gábor Boskovits <boskovits@gmail.com> wrote:
>
> > zimoun <zimon.toutoune@gmail.com> ezt írta (időpont: 2019. dec. 12., Csü 17:47):
>
> >> Maybe I miss a point. Is the aim to conserve the "--ad-hoc" option
> >> with a different effect? Or why do we want to conserve this option
> >> name?
> >> It appears to me simpler to give another name, for example
> >> "--inputs-of". And it is more meaningful.
> >
> > Sorry for the confusion. Ad-hoc should be retained with the same effect, so that we do not break existing scripts.
> > Renamin the option would be ok. It even makes sense to me.
>
> What I propose is:
>
>   - keep the option "--ad-hoc" with the current behavior; so same effect
>   - add a new option "--inputs-of" with the new behavior; name more meaningful
>   - and two env variables; to not break existing scripts
>
>
> >> First, when "--ad-hoc" is used then it reports a warning: deprecated
> >> option and falls in the current behavior.
> >> When "--inputs-of" is used then it falls in the new behavior.
> >> Therefore, no needs of the ugly "--ignore-deprecated-ad-hoc".
> >
> > That could be done. The problem is caused by uses of guix environment that does not use any of these options. Those mean different things after the change.
>
> The transition to such use-case was described below with the
> introduction of 2 env variables. :-)
>
>
> >>  # Alice
> >>  $ guix environment foo --ad-hoc bar
> >>  Warning: deprecated... explanations...
> >>                instead use:
> >>                 guix environment bar --inputs-of foo
> >>
> >>  # Bob
> >>  $ guix environment bar --inputs-of foo
> >>
> >>
> >> Second, the previous "guix environment foo" (dependencies of foo) is
> >> inconsistent with the new "guix environment bar" (only the package
> >> bar). Therefore, let introduce the GUIX_ENVIRONMENT_DEPRECATED
> >> variable to distinguish both, as you said.
> >
> > Ok.
>
> It is the easy part. ;-)
>
>
> Now the hard part: avoid to break existing scripts.
>
> >>  # Alice
> >>  $ guix environment foo
> >>  Warning: previous behavior requires GUIX_ENVIRONMENT_DEPRECATED=1
> >>                turn off the warning: GUIX_ENVIRONMENT_NOWARNING=1
> >>
> >> And Alice has now a new shell with the package foo. If she wants the
> >> dependencies, she has two options:
> >>
> >> $ GUIX_ENVIRONMENT=1 guix environment foo
> >> or
> >> $ guix environment --inputs-of foo
> >>
> >>
> >>  # Bob
> >>  $ guix environment bar
> >>  Warning: previous behavior requires GUIX_ENVIRONMENT
> >>
> >> And if Bob is annoyed by the warnings each time, he globally turns off
> >> with the variable GUIX_ENVIRONMENT_NOWARNING=1.
> >>
> >>
> >> Couple of months later -- after the period adoption -- we remove the
> >> variables GUIX_ENVIRONMENT_NOWARNING and GUIX_ENVIRONMENT_DEPRECATED;
> >> still keeping the warning with the "--ad-hoc" option. And then, after
> >> we can remove the "--ad-hoc" option if required.
>
>
> > We could recommend simply to use something like:
> > GUIX_ENVIRONMENT_DEPRECATED=0 guix environment ...
> > Instead in existing scripts that are fixed to use the new syntax. This indeed looks like a better solution, and it is less of a maintenance burden. Good idea.
>
> My point is: the new variable GUIX_ENVIRONMENT_DEPRECATED should only
> be used by the scripts that call "guix environment pkg" without the
> options "--ad-hoc" or "--inputs-of". And I think that it represents
> really few scripts in real life. :-)
>
>
> > Summarizing:
> > Introduce the environment variable.
> > For fixed scripts recommend unsetting the environment variable.
>
> I am not to get your plan. :-)
>
>
> Cheers,
> simon

So in a more algorithmic manner:
1. if ad-hoc and inputs-of is present at the same invocation: fail
hard. (With an error like incompatible options present)
2. if only ad-hoc is present, then print a deprecation warning (yes,
we could make this suspendable with an environment variable, like you
described)
3. if only inputs-of present, then do the new behaviour.
4. if neither ad-hoc nor inputs-of present then
  a. if GUIX_ENVIRONMENT_DEPRECATED is 1: do the current behaviour,
  b. if GUIX_ENVIRONMENT_DEPRECATED is undefined, or is not 1: do the
new behaviour.

This would minimze friction, as there will be a few scripts falling under 4.
This would also allow mirgating such scripts one by one. be defining
GUIX_ENVIRONMENT_DEPRECATED to 1 in some startup file, and using
GUIX_ENVIRONMENT_DEPRECATED=0 guix environment ... in scripts that are
fixed to use the new syntax.


What do you think?

Best regards,
g_bor
-- 
OpenPGP Key Fingerprint: 7988:3B9F:7D6A:4DBF:3719:0367:2506:A96C:CF63:0B21

bug#38529: Make --ad-hoc the default for guix environment proposed deprecation mechanism

[zimoun]

Hi Gábor,

Sorry to be slow. :-)

On Fri, 13 Dec 2019 at 17:28, Gábor Boskovits <boskovits@gmail.com> wrote:


> So in a more algorithmic manner:
> 1. if ad-hoc and inputs-of is present at the same invocation: fail
> hard. (With an error like incompatible options present)
> 2. if only ad-hoc is present, then print a deprecation warning (yes,
> we could make this suspendable with an environment variable, like you
> described)
> 3. if only inputs-of present, then do the new behaviour.
> 4. if neither ad-hoc nor inputs-of present then
>   a. if GUIX_ENVIRONMENT_DEPRECATED is 1: do the current behaviour,
>   b. if GUIX_ENVIRONMENT_DEPRECATED is undefined, or is not 1: do the
> new behaviour.
>
> This would minimze friction, as there will be a few scripts falling under 4.
> This would also allow mirgating such scripts one by one. be defining
> GUIX_ENVIRONMENT_DEPRECATED to 1 in some startup file, and using
> GUIX_ENVIRONMENT_DEPRECATED=0 guix environment ... in scripts that are
> fixed to use the new syntax.
>
>
> What do you think?

I am perfectly aligned! :-)
It is exactly what I have tried to describe.
Sorry again for being slow.

Thank you.
Do you plan to implement it? Do I give a try?


All the best,
simon

bug#38529: Make --ad-hoc the default for guix environment proposed deprecation mechanism

[Gábor Boskovits]

Hello Zimoun,

zimoun <zimon.toutoune@gmail.com> ezt írta (időpont: 2019. dec. 13., P, 17:32):
>
> Hi Gábor,
>
> Sorry to be slow. :-)

I probably just did not express myself clearly enough.

>
> On Fri, 13 Dec 2019 at 17:28, Gábor Boskovits <boskovits@gmail.com> wrote:
>
>
> > So in a more algorithmic manner:
> > 1. if ad-hoc and inputs-of is present at the same invocation: fail
> > hard. (With an error like incompatible options present)
> > 2. if only ad-hoc is present, then print a deprecation warning (yes,
> > we could make this suspendable with an environment variable, like you
> > described)
> > 3. if only inputs-of present, then do the new behaviour.
> > 4. if neither ad-hoc nor inputs-of present then
> >   a. if GUIX_ENVIRONMENT_DEPRECATED is 1: do the current behaviour,
> >   b. if GUIX_ENVIRONMENT_DEPRECATED is undefined, or is not 1: do the
> > new behaviour.
> >
> > This would minimze friction, as there will be a few scripts falling under 4.
> > This would also allow mirgating such scripts one by one. be defining
> > GUIX_ENVIRONMENT_DEPRECATED to 1 in some startup file, and using
> > GUIX_ENVIRONMENT_DEPRECATED=0 guix environment ... in scripts that are
> > fixed to use the new syntax.
> >
> >
> > What do you think?
>
> I am perfectly aligned! :-)

Great!

> It is exactly what I have tried to describe.
> Sorry again for being slow.

I am sorry for the confusion, my communication tends to slugghish, an
I am also a bit bound to
do some hand-waving :) I hope to improve on that

>
> Thank you.
> Do you plan to implement it? Do I give a try?

I would like to hear something from Ludo, as he was also a participant
of the IRC discussion.

After that I would not mind if you gave it a try, if you would like.
Otherwise I will implement it.

>
>
> All the best,
> simon


Best regards,
g_bor
-- 
OpenPGP Key Fingerprint: 7988:3B9F:7D6A:4DBF:3719:0367:2506:A96C:CF63:0B21

bug#38529: Make --ad-hoc the default for guix environment proposed deprecation mechanism

[Ludovic Courtès]

Hello,

Gábor Boskovits <boskovits@gmail.com> skribis:

> So in a more algorithmic manner:
> 1. if ad-hoc and inputs-of is present at the same invocation: fail
> hard. (With an error like incompatible options present)
> 2. if only ad-hoc is present, then print a deprecation warning (yes,
> we could make this suspendable with an environment variable, like you
> described)
> 3. if only inputs-of present, then do the new behaviour.
> 4. if neither ad-hoc nor inputs-of present then
>   a. if GUIX_ENVIRONMENT_DEPRECATED is 1: do the current behaviour,
>   b. if GUIX_ENVIRONMENT_DEPRECATED is undefined, or is not 1: do the
> new behaviour.

That sounds like a good plan to me.

#4 is the trickiest, and I think it’d be good to give users a bit of
time so they can start adjusting before deprecation is in effect.

Namely, we could start by introducing ‘--inputs-of’ and emitting a
warning in case #4 to suggest the use of ‘--inputs-of’.  Apart from the
warning, case #4 would still behave the same as now.

Three (?) months later, we implement what you describe above.  Hopefully
by that time many people got used to ‘--inputs-of’.

Thoughts?

Ludo’.

bug#38529: Make --ad-hoc the default for guix environment proposed deprecation mechanism

[Konrad Hinsen]

On 16/12/2019 23:09, Ludovic Courtès wrote:
> So in a more algorithmic manner:
>> 1. if ad-hoc and inputs-of is present at the same invocation: fail
>> hard. (With an error like incompatible options present)
>> 2. if only ad-hoc is present, then print a deprecation warning (yes,
>> we could make this suspendable with an environment variable, like you
>> described)
>> 3. if only inputs-of present, then do the new behaviour.
>> 4. if neither ad-hoc nor inputs-of present then
>>    a. if GUIX_ENVIRONMENT_DEPRECATED is 1: do the current behaviour,
>>    b. if GUIX_ENVIRONMENT_DEPRECATED is undefined, or is not 1: do the
>> new behaviour.
> That sounds like a good plan to me.
>
> #4 is the trickiest, and I think it’d be good to give users a bit of
> time so they can start adjusting before deprecation is in effect.

#4 is trickiest for another reason: there is no future-proof use of 
"guix environment" that works right now and will continue to work. Nor 
is there any way to see, when looking at a command line, whether it's 
old-style or new-style, if neither --ad-hoc nor --inputs-of are present. 
This means that all existing documentation (tutorials etc.) will become 
misleading in the future. Worse, even documentation written today, in 
full awareness of a coming change, can't do better than saying "watch 
out, this will do something else in the future".

The first rule of backwards-compatibility is: never change the meaning 
of an existing valid command/API. Add new valid syntax, deprecate old 
valid syntax, but don't change the meaning of something that was and 
will be valid.

How about a more drastic measure: deprecate "guix environment" and 
introduce a new subcommand with the desired new behaviour?


Cheers,

   Konrad.

bug#38529: Make --ad-hoc the default for guix environment proposed deprecation mechanism

[Gábor Boskovits]

[-- Attachment #1: Type: text/plain, Size: 2214 bytes --]

Hello Konrad,

Konrad Hinsen <konrad.hinsen@fastmail.net> ezt írta (időpont: 2019. dec.
17., Ke 7:52):

> On 16/12/2019 23:09, Ludovic Courtès wrote:
> > So in a more algorithmic manner:
> >> 1. if ad-hoc and inputs-of is present at the same invocation: fail
> >> hard. (With an error like incompatible options present)
> >> 2. if only ad-hoc is present, then print a deprecation warning (yes,
> >> we could make this suspendable with an environment variable, like you
> >> described)
> >> 3. if only inputs-of present, then do the new behaviour.
> >> 4. if neither ad-hoc nor inputs-of present then
> >>    a. if GUIX_ENVIRONMENT_DEPRECATED is 1: do the current behaviour,
> >>    b. if GUIX_ENVIRONMENT_DEPRECATED is undefined, or is not 1: do the
> >> new behaviour.
> > That sounds like a good plan to me.
> >
> > #4 is the trickiest, and I think it’d be good to give users a bit of
> > time so they can start adjusting before deprecation is in effect.
>
> #4 is trickiest for another reason: there is no future-proof use of
> "guix environment" that works right now and will continue to work. Nor
> is there any way to see, when looking at a command line, whether it's
> old-style or new-style, if neither --ad-hoc nor --inputs-of are present.
> This means that all existing documentation (tutorials etc.) will become
> misleading in the future. Worse, even documentation written today, in
> full awareness of a coming change, can't do better than saying "watch
> out, this will do something else in the future".
>
> The first rule of backwards-compatibility is: never change the meaning
> of an existing valid command/API. Add new valid syntax, deprecate old
> valid syntax, but don't change the meaning of something that was and
> will be valid.
>
> How about a more drastic measure: deprecate "guix environment" and
> introduce a new subcommand with the desired new behaviour?
>
That is also the other option I was thinking about. Do you have any good
idea in mind as how to call it? Of course the classic guix environment2
comes to my mind, but it does not look very appealing to me.

>
>
> Cheers,
>
>    Konrad.
>
Best regard,
g_bor

>
>
>
>

[-- Attachment #2: Type: text/html, Size: 3098 bytes --]

bug#38529: Make --ad-hoc the default for guix environment proposed deprecation mechanism

[Kyle Meyer]

Gábor Boskovits <boskovits@gmail.com> writes:

> Konrad Hinsen <konrad.hinsen@fastmail.net> ezt írta (időpont: 2019. dec.
> 17., Ke 7:52):
[...]
>> How about a more drastic measure: deprecate "guix environment" and
>> introduce a new subcommand with the desired new behaviour?
>>
> That is also the other option I was thinking about. Do you have any good
> idea in mind as how to call it? Of course the classic guix environment2
> comes to my mind, but it does not look very appealing to me.

Perhaps "guix env"?

bug#38529: Make --ad-hoc the default for guix environment proposed deprecation mechanism

[Brett Gilio]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain; charset=UTF-8, Size: 596 bytes --]



Dec 17, 2019 7:34:17 AM Kyle Meyer :

> Gábor Boskovits writes:
>
>
> > Konrad Hinsen ezt írta (id?pont: 2019. dec.
> > 17., Ke 7:52):
> >
> [...]
>
> >
> > > How about a more drastic measure: deprecate "guix environment" and
> > > introduce a new subcommand with the desired new behaviour?
> > >
> > >
> > That is also the other option I was thinking about. Do you have any good
> > idea in mind as how to call it? Of course the classic guix environment2
> > comes to my mind, but it does not look very appealing to me.
> >
>
> Perhaps "guix env"?
>

+1 for guix env

bug#38529: Make --ad-hoc the default for guix environment proposed deprecation mechanism

[Bengt Richter]

Hi Gábor, Konrad, et al

On +2019-12-17 10:14:12 +0100, Gábor Boskovits wrote:
> Hello Konrad,
> 
> Konrad Hinsen <konrad.hinsen@fastmail.net> ezt írta (időpont: 2019. dec.
> 17., Ke 7:52):
> 
> > On 16/12/2019 23:09, Ludovic Courtès wrote:
> > > So in a more algorithmic manner:
> > >> 1. if ad-hoc and inputs-of is present at the same invocation: fail
> > >> hard. (With an error like incompatible options present)
> > >> 2. if only ad-hoc is present, then print a deprecation warning (yes,
> > >> we could make this suspendable with an environment variable, like you
> > >> described)
> > >> 3. if only inputs-of present, then do the new behaviour.
> > >> 4. if neither ad-hoc nor inputs-of present then
> > >>    a. if GUIX_ENVIRONMENT_DEPRECATED is 1: do the current behaviour,
> > >>    b. if GUIX_ENVIRONMENT_DEPRECATED is undefined, or is not 1: do the
> > >> new behaviour.
> > > That sounds like a good plan to me.
> > >
> > > #4 is the trickiest, and I think it’d be good to give users a bit of
> > > time so they can start adjusting before deprecation is in effect.
> >
> > #4 is trickiest for another reason: there is no future-proof use of
> > "guix environment" that works right now and will continue to work. Nor
> > is there any way to see, when looking at a command line, whether it's
> > old-style or new-style, if neither --ad-hoc nor --inputs-of are present.
> > This means that all existing documentation (tutorials etc.) will become
> > misleading in the future. Worse, even documentation written today, in
> > full awareness of a coming change, can't do better than saying "watch
> > out, this will do something else in the future".
> >
> > The first rule of backwards-compatibility is: never change the meaning
> > of an existing valid command/API. Add new valid syntax, deprecate old
> > valid syntax, but don't change the meaning of something that was and
> > will be valid.
> >

I think it is important to consider context when talking about meaning.

1. the level and the interpreter of the command:
   The first level is usually the shell (typicallly bash) from logind,
   but there is systemd and/or shepherd before that, and there is bootloader
   and UEFI and before that also accepting and/or passing commands through
   to the kernel via the kernel command line (cf. cat /proc/cmdline ).

   The general pattern I mostly see for a given interpreter is
   
   verb -adverb* (-adjective-for: object-name)* sub-command? implicit-or-object-for-verb*

   Consider whether your new name reinforces a good convention or forks it.
   Consider whether proposed usage translates easily to a natural language explanation.
   Does guix have a cli design best practices doc, BTW?
   
   right now we are talking about the use case where
   verb=guix and subcommand=environment

2. project community conventions
   Specialized areas often have their own jargon and abbreviated refrences
   so an unfortunate choice of name can cause distracting disambiguation searches.

Before settling on a new name xxx, even for a sub-command, I would say at least
first do the following at the command line:
   which xxx
   xxx --version
   xxx --help
   info xxx
   man xxx
   apropos xxx
   
   #check for same prefix, case-insensitively,
   # e.g. env might be tempting, as seen in this thread :)
--8<---------------cut here---------------start------------->8---
   echo $PATH|tr : $'\n'|while read pdir;do (find "$pdir" -maxdepth 1 -iname "env*" 2>/dev/null);done
--8<---------------cut here---------------end--------------->8---
   # -name "*xxx*" may also be a good idea, but the prefix is most important
   # env* produces
   /usr/bin/env
   /usr/bin/envsubst

   guix search xxx
   guix package -A xxx
   wikipedia search on xxx, e.g.
       lynx -dump -force_html -nolist https://en.wikipedia.org/w/index.php?search=xxx |less

   You get the idea, I'm sure ;-)
   
> > How about a more drastic measure: deprecate "guix environment" and
> > introduce a new subcommand with the desired new behaviour?
> >
SGTM, with some caveats

Good, since calling different things by the same name is always going to be problematic.
Iffy, since calling the same thing by different names may reduce future naming options,
   and may muddy the peer-name namespace, so maybe consider using sub-commands or -adverb.

> That is also the other option I was thinking about. Do you have any good
> idea in mind as how to call it? Of course the classic guix environment2
> comes to my mind, but it does not look very appealing to me.
>

Me neither :)

> >
> > Cheers,
> >
> >    Konrad.
> >
> Best regard,
> g_bor
> 

HTH in some way :)

-- 
Regards,
Bengt Richter

bug#38529: Make --ad-hoc the default for guix environment proposed deprecation mechanism

[Bengt Richter]

Forgot to add:
┌──────────────────────────────────────────────────────────────────────────┐
│          guile -c '(use-modules (ice-9 session))(apropos "env")          │
├──────────────────────────────────────────────────────────────────────────┤
│ (guile): getenv  #<procedure getenv (_)>                                 │
│ (guile): environ #<procedure environ (#:optional _)>                     │
│ (guile): setenv #<procedure setenv (name value)>                         │
│ (guile): interaction-environment #<procedure interaction-environment ()> │
│ (guile): putenv #<procedure putenv (_)>                                  │
│ (guile): unsetenv #<procedure unsetenv (name)>                           │
└──────────────────────────────────────────────────────────────────────────┘

BTW, it would be really handy to be able to type
   guile -apropos rest of line as regex
for the effect of
   ,a rest of line as regex
in the guile repl
-- 
Regards,
Bengt Richter

bug#38529: Make --ad-hoc the default for guix environment proposed deprecation mechanism

[zimoun]

Hi Konrad,

On Tue, 17 Dec 2019 at 07:52, Konrad Hinsen <konrad.hinsen@fastmail.net> wrote:
>
> On 16/12/2019 23:09, Ludovic Courtès wrote:
> > So in a more algorithmic manner:
> >> 1. if ad-hoc and inputs-of is present at the same invocation: fail
> >> hard. (With an error like incompatible options present)
> >> 2. if only ad-hoc is present, then print a deprecation warning (yes,
> >> we could make this suspendable with an environment variable, like you
> >> described)
> >> 3. if only inputs-of present, then do the new behaviour.
> >> 4. if neither ad-hoc nor inputs-of present then
> >>    a. if GUIX_ENVIRONMENT_DEPRECATED is 1: do the current behaviour,
> >>    b. if GUIX_ENVIRONMENT_DEPRECATED is undefined, or is not 1: do the
> >> new behaviour.
> > That sounds like a good plan to me.
> >
> > #4 is the trickiest, and I think it’d be good to give users a bit of
> > time so they can start adjusting before deprecation is in effect.
>
> #4 is trickiest for another reason: there is no future-proof use of
> "guix environment" that works right now and will continue to work. Nor
> is there any way to see, when looking at a command line, whether it's
> old-style or new-style, if neither --ad-hoc nor --inputs-of are present.
> This means that all existing documentation (tutorials etc.) will become
> misleading in the future. Worse, even documentation written today, in
> full awareness of a coming change, can't do better than saying "watch
> out, this will do something else in the future".

I do not understand what is the issue for the time-traveling if it is
documented.
Maybe I miss a point. It is not: "watch out, this will do something
else in the future" but "watch out, this was doing something else in
the past and the change happened the <date> in <commit>".

First, I am not convinced that there is not so much scripts that will
be broken. And second, I am not convinced neither that these very
scripts need time-traveling.


> The first rule of backwards-compatibility is: never change the meaning
> of an existing valid command/API. Add new valid syntax, deprecate old
> valid syntax, but don't change the meaning of something that was and
> will be valid.

I agree on the rule.
But it is mitigated but the number of users and the popularity of the tool. ;-)


> How about a more drastic measure: deprecate "guix environment" and
> introduce a new subcommand with the desired new behaviour?

Yes, it is probably the most adequate to do. But it is sad to loose
the good name "guix environment"... and we know that naming is hard.
;-)

What about "guix shell"?


All the best,
simon

bug#38529: Make --ad-hoc the default for guix environment proposed deprecation mechanism

[Konrad Hinsen]

Hi Simon,

> Maybe I miss a point. It is not: "watch out, this will do something
> else in the future" but "watch out, this was doing something else in
> the past and the change happened the <date> in <commit>".

Concrete example: I am writing a tutorial about using Guix for
reproducible research. It shows several uses of "guix environment", some
of them without '–add-hoc' or '–inputs-of'. I know my examples will
cease to work in a few months. What am I supposed to do about this?

> First, I am not convinced that there is not so much scripts that will
> be broken. And second, I am not convinced neither that these very
> scripts need time-traveling.

Perhaps it's just me, but I use "guix environment" quite a lot in
scripts, in order to make them more reproducible. Here's a simple
example:

   #!/usr/bin/env bash
   guix environment --container --ad-hoc gcc-toolchain <<EOF
   gcc pi.c -o pi
   ./pi
   EOF
 
>> The first rule of backwards-compatibility is: never change the meaning
>> of an existing valid command/API. Add new valid syntax, deprecate old
>> valid syntax, but don't change the meaning of something that was and
>> will be valid.
>
> I agree on the rule.
> But it is mitigated but the number of users and the popularity of the tool. ;-)

Indeed!

> Yes, it is probably the most adequate to do. But it is sad to loose
> the good name "guix environment"... and we know that naming is hard.
> ;-)

I definitely agree. As a lesson for the future, maybe we should use
not-so-nice names for new commands during a kind of beta-testing phase.

Cheers,
  Konrad

bug#38529: Make --ad-hoc the default for guix environment proposed deprecation mechanism

[zimoun]

Hi Konrad,

On Wed, 18 Dec 2019 at 10:43, Konrad Hinsen <konrad.hinsen@fastmail.net> wrote:
>
> Hi Simon,
>
> > Maybe I miss a point. It is not: "watch out, this will do something
> > else in the future" but "watch out, this was doing something else in
> > the past and the change happened the <date> in <commit>".
>
> Concrete example: I am writing a tutorial about using Guix for
> reproducible research. It shows several uses of "guix environment", some
> of them without '–add-hoc' or '–inputs-of'. I know my examples will
> cease to work in a few months. What am I supposed to do about this?

Assuming "guix environment" would stay and following the proposed
plan, you would need to add GUIX_ENVIRONMENT_DEPRECATED=1 on the top
of your script. In this would not be a problem for travelling back in
time.


> > First, I am not convinced that there is not so much scripts that will
> > be broken. And second, I am not convinced neither that these very
> > scripts need time-traveling.
>
> Perhaps it's just me, but I use "guix environment" quite a lot in
> scripts, in order to make them more reproducible. Here's a simple
> example:
>
>    #!/usr/bin/env bash
>    guix environment --container --ad-hoc gcc-toolchain <<EOF
>    gcc pi.c -o pi
>    ./pi
>    EOF

With the proposed plan, this would stay the same. Even, the --ad-hoc
option could stay forever for backward compatibility.

The only issue is for example:

    #!/usr/bin/env bash
    guix environment --container gmsh <<EOF
    mkdir build
    cd build
    cmake ..
    make
    ./bin/gmsh
    EOF

And I not convinced that this kind of scripts need to be robust for
time-travelling, I mean it is easy to correct adding the --inputs-of
option or set the GUIX_ENVIRONMENT_DEPRECATED variable.


> > Yes, it is probably the most adequate to do. But it is sad to loose
> > the good name "guix environment"... and we know that naming is hard.
> > ;-)
>
> I definitely agree. As a lesson for the future, maybe we should use
> not-so-nice names for new commands during a kind of beta-testing phase.

What do you think about "guix shell" for the new "guix environment" behaviour?

What the others think?
New name (easier) vs transitional plan (trickier)?
And new names proposal:
  -  guix env
  -  guix shell
?


All the best,
simon

Re: bug#38529: Make --ad-hoc the default for guix environment proposed deprecation mechanism

[Arne Babenhauserheide]

[-- Attachment #1: Type: text/plain, Size: 941 bytes --]


Konrad Hinsen <konrad.hinsen@fastmail.net> writes:

> Hi Simon,
>
>> Maybe I miss a point. It is not: "watch out, this will do something
>> else in the future" but "watch out, this was doing something else in
>> the past and the change happened the <date> in <commit>".
>
> Concrete example: I am writing a tutorial about using Guix for
> reproducible research. It shows several uses of "guix environment", some
> of them without '–add-hoc' or '–inputs-of'. I know my examples will
> cease to work in a few months. What am I supposed to do about this?

This is the point where we need to ask ourselves:

     Should Guix be volatile software?
     http://stevelosh.com/blog/2012/04/volatile-software/

Also: Software developers should avoid traumatic changes
      https://drewdevault.com/2019/11/26/Avoid-traumatic-changes.html

Best wishes,
Arne
--
Unpolitisch sein
heißt politisch sein
ohne es zu merken

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 1076 bytes --]

Re: bug#38529: Make --ad-hoc the default for guix environment proposed deprecation mechanism

[zimoun]

Hi Arne,

Thank you for the pointers.


On Wed, 18 Dec 2019 at 21:55, Arne Babenhauserheide <arne_bab@web.de> wrote:

> Konrad Hinsen <konrad.hinsen@fastmail.net> writes:

> >> Maybe I miss a point. It is not: "watch out, this will do something
> >> else in the future" but "watch out, this was doing something else in
> >> the past and the change happened the <date> in <commit>".
> >
> > Concrete example: I am writing a tutorial about using Guix for
> > reproducible research. It shows several uses of "guix environment", some
> > of them without '–add-hoc' or '–inputs-of'. I know my examples will
> > cease to work in a few months. What am I supposed to do about this?
>
> This is the point where we need to ask ourselves:
>
>      Should Guix be volatile software?
>      http://stevelosh.com/blog/2012/04/volatile-software/

Guix is not a volatile software and will never be. Because it is
rooted in time-travelling.
The tools "guix pull --commit=", "guix <command> --manifest=", "guix
time-machine" or the "--roll-back" avoid to break what is currently
working.
Well, the section "The situation" just cannot(*) happen with Guix.
That's why Guix is awesome! ;-)

(*) well if one correctly uses Guix which is another story ;-)
and it is not perfect yet... see all the discussion about manifest. :-)


Now, let recall the formula (already discussed in this thread :-))

Number of people         Time it takes each person
using that part of   X   to figure out what changed
the program              and how to fix it

Hum? I am not convinced that the cost would be high... Because 1. the
number of people using Guix is not so high (yet!), so 2. I am almost
sure we can name the people using "guix environment" in scripts ;-).
And 3. the time to figure out what changed is really low -- especially
with warnings and hints -- and "guix environment foo -- make" would
return an error because of dependencies missing.

Other said, I do not see myself use-cases where the scripts using
"guix environment" need to be robust for time-travelling -- the same
script used with current, past and future Guix versions -- because as
it was said elsewhere: "environment" can be seen like "temporary
profile". And temporary means... well temporary. ;-)

Then, the section "The Tradeoff" advices "from newmodule import
new_foo as foo" and IMO it is what the plan proposes with the variable
GUIX_ENVIRONMENT_DEPRECATED (tricky point #4).

Last, "volatile" vs "stable" is mitigated by "The future of 'guix
environment'" [1] which really predates the 1.0. ;-)

[1] https://lists.gnu.org/archive/html/guix-devel/2017-08/msg00300.html


As I said, I am not convinced by the argument: everything would be
broken, too much time to fix the break, etc. and this proposal would
lead to disaster for the end-user. But it is my opinion based on my
restricted personal experience.


> Also: Software developers should avoid traumatic changes
>       https://drewdevault.com/2019/11/26/Avoid-traumatic-changes.html

"Traumatic changes"? Maybe a bit extreme for the change we are talking about...



Well, at the end, what is explicitly your personal opinion?
 a. Change the behaviour of "guix environment" using the proposed plan?
or
 b. Add a new command? Which one? "guix shell", "guix env" or "guix
<you-name-it>"?


All the best,
simon

bug#38529: Make --ad-hoc the default for guix environment proposed deprecation mechanism

[Arne Babenhauserheide]

[-- Attachment #1: Type: text/plain, Size: 4202 bytes --]

Hi zimoun,

zimoun <zimon.toutoune@gmail.com> writes:

>>      Should Guix be volatile software?
>>      http://stevelosh.com/blog/2012/04/volatile-software/
>
> Guix is not a volatile software and will never be. Because it is
> rooted in time-travelling.
> The tools "guix pull --commit=", "guix <command> --manifest=", "guix
> time-machine" or the "--roll-back" avoid to break what is currently
> working.

This is taking this a bit too easy. If I can no longer pull, because
that breaks my Emacs or Gnome, then Guix is broken for me: I can no
longer update my system without first adjusting my config.

> Number of people         Time it takes each person
> using that part of   X   to figure out what changed
> the program              and how to fix it
>
> Hum? I am not convinced that the cost would be high... Because 1. the
> number of people using Guix is not so high (yet!), so 2. I am almost
> sure we can name the people using "guix environment" in scripts ;-).

I’m not so sure. Guix is already used in scientific workflows, and there
is existing third-party documentation about using `guix environment`.

And can you name the people using `guix environment` by searching
backwards in their bash history?

> And 3. the time to figure out what changed is really low -- especially
> with warnings and hints -- and "guix environment foo -- make" would
> return an error because of dependencies missing.

It took me days to figure out the exact guix environment invocation that
allows me to build the tools for a paper I’m still working on. If that
breaks, that causes massive anxiety, because I then don’t know whether
I’ll find the time to fix it before I run into deadlines.

> Other said, I do not see myself use-cases where the scripts using
> "guix environment" need to be robust for time-travelling -- the same
> script used with current, past and future Guix versions -- because as
> it was said elsewhere: "environment" can be seen like "temporary
> profile". And temporary means... well temporary. ;-)

The same script must always work with future versions. Otherwise the
software is volatile.

You don’t need to be able to go back in time, but the path forward must
be without breakage.

> Then, the section "The Tradeoff" advices "from newmodule import
> new_foo as foo" and IMO it is what the plan proposes with the variable
> GUIX_ENVIRONMENT_DEPRECATED (tricky point #4).

No, that’s the opposite: from newmodule import new_foo as foo means,
that you allow users to define an environment variable called
`GUIX_ENVIRONMENT_MODERN`.

> Last, "volatile" vs "stable" is mitigated by "The future of 'guix
> environment'" [1] which really predates the 1.0. ;-)

Yepp, but we’re after 1.0 now. This might have been a blocker for 1.0,
but it wasn’t.

>> Also: Software developers should avoid traumatic changes
>>       https://drewdevault.com/2019/11/26/Avoid-traumatic-changes.html
>
> "Traumatic changes"? Maybe a bit extreme for the change we are talking about...

I don’t think so. There’s the strong version where it’s obvious: It
leads people to leave a project instantly.

There’s the weaker version which is less obvious: That’s where people
who invested efford to follow best practices suddenly find their project
to be written in legacy style, because the best practices changed.

> Well, at the end, what is explicitly your personal opinion?
>  a. Change the behaviour of "guix environment" using the proposed plan?
> or
>  b. Add a new command? Which one? "guix shell", "guix env" or "guix
> <you-name-it>"?

I would opt for b. And then for changing guix to give the most common
commands when called without argument (as `guix`) — excluding
guix environment.

That would not avoid the slow version of traumatic changes, but if
guix environment would keep working and both guix env/guix shell/… and
guix environment used the same backend (just with different options),
then they would be minimized, because guix environment would not become
a second-class citizen.

Best wishes,
Arne
--
Unpolitisch sein
heißt politisch sein
ohne es zu merken

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 1076 bytes --]

bug#38529: Make --ad-hoc the default for guix environment proposed deprecation mechanism

[zimoun]

Hi Arne,

First, have you read the proposal?
Or are you (maybe a bit) "overreacting" about the backward compatibility?


On Thu, 19 Dec 2019 at 22:39, Arne Babenhauserheide <arne_bab@web.de> wrote:

> zimoun <zimon.toutoune@gmail.com> writes:

> > Guix is not a volatile software and will never be. Because it is
> > rooted in time-travelling.
> > The tools "guix pull --commit=", "guix <command> --manifest=", "guix
> > time-machine" or the "--roll-back" avoid to break what is currently
> > working.
>
> This is taking this a bit too easy. If I can no longer pull, because
> that breaks my Emacs or Gnome, then Guix is broken for me: I can no
> longer update my system without first adjusting my config.

So you expect that we would push a patch changing "guix environment"
and in the same time break "guix pull, isn't it?
Otherwise I do not see your argument.


> > Hum? I am not convinced that the cost would be high... Because 1. the
> > number of people using Guix is not so high (yet!), so 2. I am almost
> > sure we can name the people using "guix environment" in scripts ;-).
>
> I’m not so sure. Guix is already used in scientific workflows, and there
> is existing third-party documentation about using `guix environment`.

Please point me where.
It will save me time instead of reinventing the wheel.


> And can you name the people using `guix environment` by searching
> backwards in their bash history?

So what would break?
Your workflow: spending 5 minutes to read the warning message and then pressing:
C-a GUIX_ENVIRONMENT_DEPRACATED=1 guix environment <your-complicated-invokation>

(unfair and bitter; sorry!)


> > And 3. the time to figure out what changed is really low -- especially
> > with warnings and hints -- and "guix environment foo -- make" would
> > return an error because of dependencies missing.
>
> It took me days to figure out the exact guix environment invocation that
> allows me to build the tools for a paper I’m still working on. If that
> breaks, that causes massive anxiety, because I then don’t know whether
> I’ll find the time to fix it before I run into deadlines.

Do you mean add GUIX_ENVIRONMENT_DEPRECATED=1 at the top of your script?


> > Other said, I do not see myself use-cases where the scripts using
> > "guix environment" need to be robust for time-travelling -- the same
> > script used with current, past and future Guix versions -- because as
> > it was said elsewhere: "environment" can be seen like "temporary
> > profile". And temporary means... well temporary. ;-)
>
> The same script must always work with future versions. Otherwise the
> software is volatile.

Here is the real argument.

It is a point of view. I would like to ear the one of others.
If I understand well, Konrad agrees with you.

I am fine with: the same script must always work with future versions.

It is a strong statement and if the Guix project agrees then it must
be documented. For example in this section [1].

What do you think?


[1] http://guix.gnu.org/manual/devel/en/html_node/Managing-Software-the-Guix-Way.html#Managing-Software-the-Guix-Way


> You don’t need to be able to go back in time, but the path forward must
> be without breakage.

Talking about Reproducible Science, going back in time is the core
issue. If one is able to go back in time and to run again the (almost)
exact same version, then the future is not the issue.

Correct me if I misunderstand your point.
Today, I write a script using X tools at time T. In the future, I want
to re-run this script so all the X tools must have a path forward
without any breakage. It is your point, right? But this never happens,
there is always a breakage somewhere; and generally for good reasons.
Instead in this future, if I am able to restore the exact same X tools
as they were at time T, my script still works.

Well, this is another story.



> > Last, "volatile" vs "stable" is mitigated by "The future of 'guix
> > environment'" [1] which really predates the 1.0. ;-)
>
> Yepp, but we’re after 1.0 now. This might have been a blocker for 1.0,
> but it wasn’t.

So if the version bump, it is not an issue then, isn't it?


> >> Also: Software developers should avoid traumatic changes
> >>       https://drewdevault.com/2019/11/26/Avoid-traumatic-changes.html
> >
> > "Traumatic changes"? Maybe a bit extreme for the change we are talking about...
>
> I don’t think so. There’s the strong version where it’s obvious: It
> leads people to leave a project instantly.

Yes, me!


> There’s the weaker version which is less obvious: That’s where people
> who invested efford to follow best practices suddenly find their project
> to be written in legacy style, because the best practices changed.

Best practise depends on a lot of parameters. I did not know it was frozen.



Well, I withdraw my investment. I am not interested anymore to not
tell that I am traumatized.


Regards,
simon

bug#38529: Make --ad-hoc the default for guix environment proposed deprecation mechanism

[Arne Babenhauserheide]

[-- Attachment #1: Type: text/plain, Size: 10413 bytes --]


zimoun <zimon.toutoune@gmail.com> writes:
> First, have you read the proposal?

Yes.

> Or are you (maybe a bit) "overreacting" about the backward compatibility?

I don’t think so. I am definitely reacting strongly, but that’s because
breakages in Guix have already cost me the evenings of several weeks
this year.

But before I write anything more, I’d like to ask you to take a step
back to breathe.

We’re discussing a change in software. We disagree on the way forward,
but I’m not attacking you as a person, and I hope it does not feel that
way to you.

If it does: This is not my intention. Please take a moment to sigh
deeply, shake your head, relax, and smile — because that actually helps.
It’s what I try to do when discussions get vexing.

I am grateful that you’re taking up improvements in Guix, and there are
situations where viewpoints are different. That is OK.

>> This is taking this a bit too easy. If I can no longer pull, because
>> that breaks my Emacs or Gnome, then Guix is broken for me: I can no
>> longer update my system without first adjusting my config.
>
> So you expect that we would push a patch changing "guix environment"
> and in the same time break "guix pull, isn't it?

No, this is an example which shows that being able to roll back does not
mean that there is no problem with breaking the way forward. Using only
old versions is often not an option. Just imagine running audio software
from 5 years ago on a system that only provides pulseaudio (or whatever
will come after it). Imagine using an old KDE DCOP-based automation
workflow on a dbus-system. You need to update the libraries you use to
get it working at all.

>> > Hum? I am not convinced that the cost would be high... Because 1. the
>> > number of people using Guix is not so high (yet!), so 2. I am almost
>> > sure we can name the people using "guix environment" in scripts ;-).
>>
>> I’m not so sure. Guix is already used in scientific workflows, and there
>> is existing third-party documentation about using `guix environment`.
>
> Please point me where.
> It will save me time instead of reinventing the wheel.

It was mentioned on this list.

For the scientific workflows, see https://hpc.guix.info/

>> And can you name the people using `guix environment` by searching
>> backwards in their bash history?
>
> So what would break?
> Your workflow: spending 5 minutes to read the warning message and then pressing:
> C-a GUIX_ENVIRONMENT_DEPRACATED=1 guix environment <your-complicated-invokation>
>
> (unfair and bitter; sorry!)

I’m sorry that this makes you bitter. This is not my intention.

I’ll answer without bitterness: The original environment does not spawn
instantly. It takes many minutes until it is ready. If I then have to go
back, find the warning (it’s likely that I’d miss it, because these are
things that work, and suddenly they break, which I’m likely to only
figure out when the followup steps don’t work) and run it again, that
often means that I’m out of time to do what I actually wanted to do.

Despite that: Yes, this is a viable way. It is one of the less painful
ones. Maybe avoid calling it "DEPRECATED" and instead give it a more
descriptive name that does not imply that it will go away.

Mercurial uses HGPLAIN=1 to say "I want the version which will never
change established API". Best practice is to always use that in scripts
— and that is a stable best practice. But this is also slow to receive
new features.

If the old way to use guix environment is intended to actually be legacy
only, then it could be a way forward to also provide
GUIX_ENVIRONMENT_STABLE=1 which gives an API that is guaranteed to never
change the meaning of options again *after the change that’s been
started to brew in 2017*.

That would be a purely append-only API then, and while it would break
once, it would prevent such changes for the future.

For PR it might be possible to state that with this change, guix
environment as a tool reaches version 0.99 (to be updated to 1.0 after
sufficient testing).

>> > And 3. the time to figure out what changed is really low -- especially
>> > with warnings and hints -- and "guix environment foo -- make" would
>> > return an error because of dependencies missing.
>>
>> It took me days to figure out the exact guix environment invocation that
>> allows me to build the tools for a paper I’m still working on. If that
>> breaks, that causes massive anxiety, because I then don’t know whether
>> I’ll find the time to fix it before I run into deadlines.
>
> Do you mean add GUIX_ENVIRONMENT_DEPRECATED=1 at the top of your script?

Yes, at every script. And remember to add it to every command I still
have in history.

>> > Other said, I do not see myself use-cases where the scripts using
>> > "guix environment" need to be robust for time-travelling -- the same
>> > script used with current, past and future Guix versions -- because as
>> > it was said elsewhere: "environment" can be seen like "temporary
>> > profile". And temporary means... well temporary. ;-)
>>
>> The same script must always work with future versions. Otherwise the
>> software is volatile.
>
> Here is the real argument.
>
> It is a point of view. I would like to ear the one of others.
> If I understand well, Konrad agrees with you.
>
> I am fine with: the same script must always work with future versions.
>
> It is a strong statement and if the Guix project agrees then it must
> be documented. For example in this section [1].
>
> What do you think?

Only if this is actually the stance of the whole Guix project.

Currently this is the argument given by one person in an email
discussion. I think that it is a strong and important argument
(otherwise I would not have made it), but I’ve been wrong before.

Maybe the change to Guix environment now is for the best of the project.
I cannot actually see so clearly into the future that I could say
whether the churn due to the breaking change or the annoyance due to
suboptimal default behavior will be worse.

> [1] http://guix.gnu.org/manual/devel/en/html_node/Managing-Software-the-Guix-Way.html#Managing-Software-the-Guix-Way
>
>
>> You don’t need to be able to go back in time, but the path forward must
>> be without breakage.
>
> Talking about Reproducible Science, going back in time is the core
> issue. If one is able to go back in time and to run again the (almost)
> exact same version, then the future is not the issue.

The future is an issue, because you often have to use up-to-date
libraries. Just imagine using a rust-tool but being stuck in a 12 months
old environment that you cannot update without breakage.

> Correct me if I misunderstand your point.
> Today, I write a script using X tools at time T. In the future, I want
> to re-run this script so all the X tools must have a path forward
> without any breakage. It is your point, right? But this never happens,
> there is always a breakage somewhere; and generally for good reasons.

No, and that’s the point. That’s also the point of the article: There
are tools which almost never break. And there are tools that almost
always break.

If you use a tool out of the latter group, you’re in for a world of
pain. It’s why it took years and years for somewhat stable git-wrappers
to appear: The early wrappers that made git easier to use always broke
when git changed operation. Guix environment might actually help you
delay the update until you have time to deal with the breakage.

> Instead in this future, if I am able to restore the exact same X tools
> as they were at time T, my script still works.
>
> Well, this is another story.

This helps surviving volatility in other tools, but only if tool for
doing so isn’t volatile itself.

>> > Last, "volatile" vs "stable" is mitigated by "The future of 'guix
>> > environment'" [1] which really predates the 1.0. ;-)
>>
>> Yepp, but we’re after 1.0 now. This might have been a blocker for 1.0,
>> but it wasn’t.
>
> So if the version bump, it is not an issue then, isn't it?

It would still be an issue, but see the part about seeing into the
future above :-)

>> >> Also: Software developers should avoid traumatic changes
>> >>       https://drewdevault.com/2019/11/26/Avoid-traumatic-changes.html
>> >
>> > "Traumatic changes"? Maybe a bit extreme for the change we are talking about...
>>
>> I don’t think so. There’s the strong version where it’s obvious: It
>> leads people to leave a project instantly.
>
> Yes, me!

Have a look at your reaction here. This is just the kind of reaction
people feel when something into which they invested time suddenly stops
working as expected.

>> There’s the weaker version which is less obvious: That’s where people
>> who invested efford to follow best practices suddenly find their project
>> to be written in legacy style, because the best practices changed.
>
> Best practise depends on a lot of parameters. I did not know it was frozen.

If you manage to freeze the best practices without blocking ways into
the future, then you found part of the holy grail of software
development: You managed to find one fragment that’s so good that it
never needs to change again and everything new you do fits to it.

Typically reality isn’t quite as beautiful and change can break your
model. They say about Lisp that it’s a snowball: You can keep adding
stuff to it and it always stays a snowball. That’s close to this beauty.
But Lisp is also full of car/cdr-namings and legacy you cannot shed,
even though you might want to.

You cannot reach-and-keep perfection in a changing world, you can only
try to limit the pain for users and stay close to something which feels
right.

Volatile projects do not work to limit the pain.

Stale projects do not try to stay close to ways that feel right in a
changing reality.

A good project needs to get as close as possible to a consensus (I’m not
saying compromise here, because that’s not what I mean — the goal is
something which unites both) between not being volatile and not becoming
stale.

Best wishes,
Arne
--
Unpolitisch sein
heißt politisch sein
ohne es zu merken

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 1076 bytes --]

Re: bug#38529: Make --ad-hoc the default for guix environment proposed deprecation mechanism

[zimoun]

Hi Arne,


On Fri, 20 Dec 2019 at 02:37, Arne Babenhauserheide <arne_bab@web.de> wrote:

> > Or are you (maybe a bit) "overreacting" about the backward compatibility?
>
> I don’t think so. I am definitely reacting strongly, but that’s because
> breakages in Guix have already cost me the evenings of several weeks
> this year.
>
> But before I write anything more, I’d like to ask you to take a step
> back to breathe.
>
> We’re discussing a change in software. We disagree on the way forward,
> but I’m not attacking you as a person, and I hope it does not feel that
> way to you.
>
> If it does: This is not my intention. Please take a moment to sigh
> deeply, shake your head, relax, and smile — because that actually helps.
> It’s what I try to do when discussions get vexing.
>
> I am grateful that you’re taking up improvements in Guix, and there are
> situations where viewpoints are different. That is OK.


I am fine. :-)
Life is about managing disagreements.
And I am probably a typical grouchy French. ;-)

Well, if we go back in time, the story is:
 - the original author of "guix environment" is not happy with the
current behaviour and proposes a change (see "The future of 'guix
environment'").
 - life happens (v1.0) but not this change.
 - I am not happy with the current behaviour and other on IRC neither.
 - a plan to change is opened for discussions.

The first concern by Ludo is about the compatibility.
Then Konrad raises concrete examples.

At this point, my personal opinion is: the cost is low so the change can happen.
However, I agree with the "backward compatibility" issue and even I
propose a name for this "new" command: "guix shell".

Then you ask one question: "Should Guix be volatile software?" with
the subtitle "Software developers should avoid traumatic changes".
Nothing more.
Well, I answer you by trying to fill the gap. Note that "volatile
software" is the same argument than the Ludo's concern and the
Konrad's example. So, nothing new on the table; except you are
starting to throw "feelings" with the "traumatic change" words.

Then, your following answer is more about your feelings than concrete
examples. It is hard to know in advance how many scripts or use-cases
would be broken -- i.e., estimate the cost -- and a way is to probe;
say: "it will break X of my scripts" or "in my institute, X people use
"guix environment blabla" daily, so it is not an option", etc.
Otherwise, it is unproductive.

Well, instead of arguing about feelings because it is going nowhere or
at better a flame war about "backward compatibility", I prefer going
to spend my time elsewhere (still about Guix :-)).
I mean, I proposed, I said my opinion and I called to collect more
opinions. I feel I did my best on this front and other fronts deserve
proposals and fixes.


Kind regards,
simon

ps:
Note that I did a proposal which could be a path to reduce the burden
of "guix pull" breakage: adding tags. Feel free to comment.

https://lists.gnu.org/archive/html/guix-devel/2019-11/msg00513.html

bug#38529: Make --ad-hoc the default for guix environment proposed deprecation mechanism

[Ricardo Wurmus]

zimoun <zimon.toutoune@gmail.com> writes:

> Then you ask one question: "Should Guix be volatile software?" with
> the subtitle "Software developers should avoid traumatic changes".
> Nothing more.
> Well, I answer you by trying to fill the gap. Note that "volatile
> software" is the same argument than the Ludo's concern and the
> Konrad's example. So, nothing new on the table; except you are
> starting to throw "feelings" with the "traumatic change" words.

I’m just chiming in here to say that feelings of frustration are very
valid reasons to make or object to a change.  Guix is or can be a very
important piece of software — if it remains reliable in the toolbox of
those using it.

It is difficult striking the right balance between exciting new features
that make things possible that were previously unattainable and
dependability through stable interfaces.

The Guix command line is by far the most commonly used interface.  We
can’t just claim that the Scheme API is stable (which it actually isn’t)
and change the user-facing CLI as we please.

Personally, I think that it is fine to introduce breaking changes, but
that for changes that are likely to have a high potential for annoyance
and frustration there ought to be a documented way to work around them.
Breaking changes must be communicated through version number bumps and
accompanying announcements.

While I don’t see how we can make it happen, I do find the idea of a
stable API whose version can be selected with an environment variable
intriguing and worth thinking about.  If our Scheme API is as flexible
as we claim it shouldn’t be too hard to interpose a configuration layer
between the core facilities and the “porcelain”.

I wonder what the other maintainers think about this.

--
Ricardo

bug#38529: Make --ad-hoc the default for guix environment proposed deprecation mechanism

[Gábor Boskovits]

[-- Attachment #1: Type: text/plain, Size: 2801 bytes --]

Ricardo Wurmus <rekado@elephly.net> ezt írta (időpont: 2019. dec. 20., Pén
22:32):

>
> zimoun <zimon.toutoune@gmail.com> writes:
>
> > Then you ask one question: "Should Guix be volatile software?" with
> > the subtitle "Software developers should avoid traumatic changes".
> > Nothing more.
> > Well, I answer you by trying to fill the gap. Note that "volatile
> > software" is the same argument than the Ludo's concern and the
> > Konrad's example. So, nothing new on the table; except you are
> > starting to throw "feelings" with the "traumatic change" words.
>
> I’m just chiming in here to say that feelings of frustration are very
> valid reasons to make or object to a change.  Guix is or can be a very
> important piece of software — if it remains reliable in the toolbox of
> those using it.
>
> It is difficult striking the right balance between exciting new features
> that make things possible that were previously unattainable and
> dependability through stable interfaces.
>
> The Guix command line is by far the most commonly used interface.  We
> can’t just claim that the Scheme API is stable (which it actually isn’t)
> and change the user-facing CLI as we please.
>
> Personally, I think that it is fine to introduce breaking changes, but
> that for changes that are likely to have a high potential for annoyance
> and frustration there ought to be a documented way to work around them.
> Breaking changes must be communicated through version number bumps and
> accompanying announcements.
>
> While I don’t see how we can make it happen, I do find the idea of a
> stable API whose version can be selected with an environment variable
> intriguing and worth thinking about.  If our Scheme API is as flexible
> as we claim it shouldn’t be too hard to interpose a configuration layer
> between the core facilities and the “porcelain”.
>
This is something that needs consideration. I believe that the original
ideas presented here, and what you say about having a stable api can be
easily synchronized by naming the environment variable to something like
GUIX_CLI_API_VERSION. I would propose it to be of the form 1.0.1.0, so that
the first three numbers could be the current guix  version. Havin it this
way would allow inter releas updates bumping the last number, and the
ability to easily set a new default when the major version is bumped, which
implies a breaking change anyways. From there on the question would be what
should be the default? I would say, that is should be
<current-mayor>.0.0.0. Does that make sense?  Maybe we could come up with
something simpler, like dropping the second and third number.

>
> I wonder what the other maintainers think about this.
>
> --
> Ricardo
>
>
>
>
>

[-- Attachment #2: Type: text/html, Size: 3521 bytes --]

bug#38529: Make --ad-hoc the default for guix environment proposed deprecation mechanism

[Ludovic Courtès]

Hi!

Ricardo Wurmus <rekado@elephly.net> skribis:

> I’m just chiming in here to say that feelings of frustration are very
> valid reasons to make or object to a change.  Guix is or can be a very
> important piece of software — if it remains reliable in the toolbox of
> those using it.
>
> It is difficult striking the right balance between exciting new features
> that make things possible that were previously unattainable and
> dependability through stable interfaces.
>
> The Guix command line is by far the most commonly used interface.  We
> can’t just claim that the Scheme API is stable (which it actually isn’t)
> and change the user-facing CLI as we please.

Agreed.

> Personally, I think that it is fine to introduce breaking changes, but
> that for changes that are likely to have a high potential for annoyance
> and frustration there ought to be a documented way to work around them.
> Breaking changes must be communicated through version number bumps and
> accompanying announcements.

Yes, I think it is clear that we’d have to do this using all the tools
at our disposal, including time.

Konrad’s objection remains though: existing documents (papers, blog
posts, MOOCs, etc.) that mention ‘guix environment’ would all of a
sudden become wrong if we were to change the defaults of ‘guix
environment’.  Even if we introduce a variable to restore the old
behavior.

Perhaps that’s unavoidable in the long run, but perhaps this is not the
right time for this.

> While I don’t see how we can make it happen, I do find the idea of a
> stable API whose version can be selected with an environment variable
> intriguing and worth thinking about.  If our Scheme API is as flexible
> as we claim it shouldn’t be too hard to interpose a configuration layer
> between the core facilities and the “porcelain”.

You mean a stable Scheme API, or a stable CLI?

To me, a stable CLI is definitely the goal.  As for the Scheme API, I
would distinguish core APIs, peripheral APIs (e.g., the importers), (gnu
system …) APIs, and packages.  I’d aim for high stability for core APIs,
be laxer for peripheral APIs, even laxer for the remaining.

I’m not sure what you mean about adding a configuration layer between
the core facilities (the core Scheme APIs?) and the porcelain?

Thanks,
Ludo’.

bug#38529: Make --ad-hoc the default for guix environment proposed deprecation mechanism

[EuAndreh via Bug reports for GNU Guix]

Hello :)

Jumping in the discussion xD

Ludovic Courtès <ludo@gnu.org> writes:

> Yes, I think it is clear that we’d have to do this using all the tools
> at our disposal, including time.
>
> Konrad’s objection remains though: existing documents (papers, blog
> posts, MOOCs, etc.) that mention ‘guix environment’ would all of a
> sudden become wrong if we were to change the defaults of ‘guix
> environment’.  Even if we introduce a variable to restore the old
> behavior.
>
> Perhaps that’s unavoidable in the long run, but perhaps this is not the
> right time for this.

Wouldn't having a new name for the new behaviour avoid breakage in this
situation?

Suppose the path of adding new subcommand is chosen, and it is "guix
shell". Couldn't it adopt the new desired behaviour?

  guix shell foo --inputs-of bar     # new command
  guix environment bar --ad-hoc foo  # untouched old command

After the introduction of "guix shell", "guix environment" could become
deprecated, but no current usage of it would stop working, and no public
references to it in the internet would become misleading. "guix
environment" could say that is has been deprecated, and point to "guix
shell", but keep working the same way.

If desired, "guix environment" could be removed after X time of
deprecation has passed, but that would be optional.

What are the downsides? Am I missing something?

Thanks,
euandreh.

bug#38529: Make --ad-hoc the default for guix environment proposed deprecation mechanism

[Ludovic Courtès]

Hi,

EuAndreh <eu@euandre.org> skribis:

> Ludovic Courtès <ludo@gnu.org> writes:
>
>> Yes, I think it is clear that we’d have to do this using all the tools
>> at our disposal, including time.
>>
>> Konrad’s objection remains though: existing documents (papers, blog
>> posts, MOOCs, etc.) that mention ‘guix environment’ would all of a
>> sudden become wrong if we were to change the defaults of ‘guix
>> environment’.  Even if we introduce a variable to restore the old
>> behavior.
>>
>> Perhaps that’s unavoidable in the long run, but perhaps this is not the
>> right time for this.
>
> Wouldn't having a new name for the new behaviour avoid breakage in this
> situation?

Yes, that’s correct (that’s also one of the suggestions Konrad made).

We could take that route.  What would we call it, though?  I don’t like
“guix shell” because it doesn’t quite reflect what the command is
about.  No good idea, though.

Thanks,
Ludo’.

bug#38529: Make --ad-hoc the default for guix environment proposed deprecation mechanism

[zimoun]

Hi Ludo,

On Mon, 30 Dec 2019 at 11:35, Ludovic Courtès <ludo@gnu.org> wrote:

> > Wouldn't having a new name for the new behaviour avoid breakage in this
> > situation?
>
> Yes, that’s correct (that’s also one of the suggestions Konrad made).

Is this statement acted? Is it the consensus by all the maintainers?

And I am not clear about what will happens for "guix environment"?
Deprecate for sure.
But after X time: removed or frozen?

Removing the command "guix environment" is against the backward
compatibility argument because all the current documentation/scripts
using it will not work anymore. Other said, if the
documentation/scripts cannot be updated as it was said -- in favor for
strong backward compatibility -- then the user will be surprised that
what worked does not anymore because the command does not exist
anymore.

Therefore, if Guix goes the backward compatibility route, then the
"guix environment" should be frozen until the version 2.0 and so only
removed when the 2.0 will be released. Or I misunderstand the
arguments in favor of the backward compatibility.

As Arne described the process (bottom of [1]), "guix environment" will
become a kind-of alias of "guix shell/<name-it>". Right?



> We could take that route.  What would we call it, though?  I don’t like
> “guix shell” because it doesn’t quite reflect what the command is
> about.  No good idea, though.

Argh! Naming is hard.
Something that reflects what the command is about: "guix environment"?
(joke!! ;-))

Why do you say that "guix shell" does not reflect what the command is about?
Because the command spawns a new shell with options (expanding it,
isolating it, etc.)

Well, because we do not seem having good idea for a new name, maybe if
we argument why we collectively find that name or this name is bad or
good, one of us will find the good name. Currently, "guix shell" seems
the better option.


All the best,
simon

bug#38529: Make --ad-hoc the default for guix environment proposed deprecation mechanism

[Ludovic Courtès]

Hello!

zimoun <zimon.toutoune@gmail.com> skribis:

> On Mon, 30 Dec 2019 at 11:35, Ludovic Courtès <ludo@gnu.org> wrote:
>
>> > Wouldn't having a new name for the new behaviour avoid breakage in this
>> > situation?
>>
>> Yes, that’s correct (that’s also one of the suggestions Konrad made).
>
> Is this statement acted? Is it the consensus by all the maintainers?

All I’m saying is that what EuAndreh wrote above is correct; I’m not
stating anything as to what solution we should implement.  :-)

> And I am not clear about what will happens for "guix environment"?
> Deprecate for sure.
> But after X time: removed or frozen?

I guess that’s the whole point of deprecation.

> As Arne described the process (bottom of [1]), "guix environment" will
> become a kind-of alias of "guix shell/<name-it>". Right?

Yes.

>> We could take that route.  What would we call it, though?  I don’t like
>> “guix shell” because it doesn’t quite reflect what the command is
>> about.  No good idea, though.
>
> Argh! Naming is hard.
> Something that reflects what the command is about: "guix environment"?
> (joke!! ;-))

Yeah!

> Why do you say that "guix shell" does not reflect what the command is about?
> Because the command spawns a new shell with options (expanding it,
> isolating it, etc.)

The command does not necessarily spawn a new shell; it spawns a command
in a well-defined environment, and that command might be a shell.

Thanks,
Ludo’.

bug#38529: Make --ad-hoc the default for guix environment proposed deprecation mechanism

[zimoun]

Hey!

On Mon, 30 Dec 2019 at 16:06, Ludovic Courtès <ludo@gnu.org> wrote:
> zimoun <zimon.toutoune@gmail.com> skribis:
> > On Mon, 30 Dec 2019 at 11:35, Ludovic Courtès <ludo@gnu.org> wrote:

> > Is this statement acted? Is it the consensus by all the maintainers?
>
> All I’m saying is that what EuAndreh wrote above is correct; I’m not
> stating anything as to what solution we should implement.  :-)

Héhé, it is an answer to the questions. ;-)


> > Why do you say that "guix shell" does not reflect what the command is about?
> > Because the command spawns a new shell with options (expanding it,
> > isolating it, etc.)
>
> The command does not necessarily spawn a new shell; it spawns a command
> in a well-defined environment, and that command might be a shell.

What about "guix spawn"?

All the best,
simon

bug#38529: Make --ad-hoc the default for guix environment proposed deprecation mechanism

[Ricardo Wurmus]

zimoun <zimon.toutoune@gmail.com> writes:

>> > Why do you say that "guix shell" does not reflect what the command is about?
>> > Because the command spawns a new shell with options (expanding it,
>> > isolating it, etc.)
>>
>> The command does not necessarily spawn a new shell; it spawns a command
>> in a well-defined environment, and that command might be a shell.
>
> What about "guix spawn"?

“spawn” is a very generic verb, much like “enter” (enter what?) or
“make”.  “shell” has the awkward property of meaning different things
dependent on how you interpret it: “to shell” means to *remove* an outer
shell (like that of a nut) whereas “guix shell” as a noun would imply
*wrapping“ something in a shell.  It sends mixed signals.  We’d probably
want people to understand it as ‘spawn a command line shell’, but that’s
really not the primary purpose of ‘guix environment’.

Thinking about words some more I started to wonder: do we want verbs or
nouns?  We have some sub-commands that could be interpreted either way:

   archive
   gc
   hash

Others that are primarily understood as nouns:

   container
   environment
   graph
   package
   processes
   repl
   size
   system
   time-machine
   weather

And a majority that are primarily understood as verbs:

   build
   challenge
   copy
   deploy
   describe
   download
   edit
   import
   install
   lint
   pack
   publish
   pull
   refresh
   remove
   search
   show
   upgrade

If we were looking for verbs that express the idea of creating an
environment or to place a thing inside of an environment we could use
one of these:

   to envelop (envelop what though?  This seems to require two objects.)
   to arrange (kinda misses the point)
   to stage (in the theatric sense)
   to frame (not in the criminal sense)
   to contain (…the resulting process in a possibly leaky environment)
   to join (…all these packages to form a new whole)
   to group (…all these packages)

(As a bonus: ‘to environ’ exists, but it suffers from the same problem
as ‘to envelop’.)

Here are some nouns that might work:

   scene
   frame
   context
   union

All of them are shorter than “environment”!  :)
What do you think?

--
Ricardo

bug#38529: Make --ad-hoc the default for guix environment proposed deprecation mechanism

[zimoun]

Hi Ricardo,

Thank you for your input.

On Mon, 30 Dec 2019 at 22:10, Ricardo Wurmus <rekado@elephly.net> wrote:

> zimoun <zimon.toutoune@gmail.com> writes:

> > What about "guix spawn"?
>
> “spawn” is a very generic verb, much like “enter” (enter what?) or
> “make”.

My English is not good enough to see the drawback. :-)


Well, my personal flavor is:

>    to contain (…the resulting process in a possibly leaky environment)
>    to join (…all these packages to form a new whole)
>    context
>    union


All the best,
simon

bug#38529: Make --ad-hoc the default for guix environment proposed deprecation mechanism

[Ludovic Courtès]

Hi!

Ricardo Wurmus <rekado@elephly.net> skribis:

> zimoun <zimon.toutoune@gmail.com> writes:

[...]

>> What about "guix spawn"?
>
> “spawn” is a very generic verb, much like “enter” (enter what?) or
> “make”.  “shell” has the awkward property of meaning different things
> dependent on how you interpret it: “to shell” means to *remove* an outer
> shell (like that of a nut) whereas “guix shell” as a noun would imply
> *wrapping“ something in a shell.  It sends mixed signals.  We’d probably
> want people to understand it as ‘spawn a command line shell’, but that’s
> really not the primary purpose of ‘guix environment’.

Yeah.

> Thinking about words some more I started to wonder: do we want verbs or
> nouns?

I think verbs are preferred, but nouns are accepted.  :-)

For example, ‘time-machine’ was recently introduced, but I find it nice
that way; ’travel-in-time’ wouldn’t be better.

It’s much like Scheme APIs: we use nouns for object properties (like
‘commit-parent’) and verbs for things that are best viewed as actions
(like ‘fold’).  This is all subjective in a functional setting!

> If we were looking for verbs that express the idea of creating an
> environment or to place a thing inside of an environment we could use
> one of these:
>
>    to envelop (envelop what though?  This seems to require two objects.)
>    to arrange (kinda misses the point)
>    to stage (in the theatric sense)
>    to frame (not in the criminal sense)
>    to contain (…the resulting process in a possibly leaky environment)
>    to join (…all these packages to form a new whole)
>    to group (…all these packages)
>
> (As a bonus: ‘to environ’ exists, but it suffers from the same problem
> as ‘to envelop’.)
>
> Here are some nouns that might work:
>
>    scene
>    frame
>    context
>    union
>
> All of them are shorter than “environment”!  :)

More data points!  :-)

--8<---------------cut here---------------start------------->8---
$ wn environment -synsn

Synonyms/Hypernyms (Ordered by Estimated Frequency) of noun environment

2 senses of environment                                                 

Sense 1
environment
       => situation, state of affairs

Sense 2
environment, environs, surroundings, surround
       => geographical area, geographic area, geographical region, geographic region
$ wn environ -synsv

Synonyms/Hypernyms (Ordered by Estimated Frequency) of verb environ

1 sense of environ                                                      

Sense 1
surround, environ, ring, skirt, border
       => touch, adjoin, meet, contact
--8<---------------cut here---------------end--------------->8---

Maybe “union”, “surround”, or… “profile”?

(I’d reserve “guix spawn” or “guix run” for the tool that runs commands
in a least-authority environment, as we’ve discussed in the past.)

Ludo’.

bug#38529: Make --ad-hoc the default for guix environment proposed deprecation mechanism

[Ricardo Wurmus]

Ludovic Courtès <ludo@gnu.org> writes:

> Maybe “union”, “surround”, or… “profile”?

“profile” is a tempting choice, but it’s treacherous because we might be
blinded by the glow of the implementation of environments as volatile
profiles.  On the other hand: if we could also move some of the features
of the “package” sub-command under “profile” (e.g. those that relate to
the management of, well, profiles), that could be a winning move.

Tricky.

> (I’d reserve “guix spawn” or “guix run” for the tool that runs commands
> in a least-authority environment, as we’ve discussed in the past.)

Same.

--
Ricardo

bug#38529: Make --ad-hoc the default for guix environment proposed deprecation mechanism

[zimoun]

Hi,

Happy New Year!
(if you are using a gregorian calendar based on the January 1srt reform)


On Tue, 31 Dec 2019 at 20:09, Ricardo Wurmus <rekado@elephly.net> wrote:

> “profile” is a tempting choice, but it’s treacherous because we might be
> blinded by the glow of the implementation of environments as volatile
> profiles.  On the other hand: if we could also move some of the features
> of the “package” sub-command under “profile” (e.g. those that relate to
> the management of, well, profiles), that could be a winning move.

If the new "guix profile" does more or less the same thing than the
current "guix environment", then I find the word "profile" confusing
because the concept of "profile" is not exactly the same elsewhere.


However, if the current CLI is changed is a bit, for example splitting
the current "guix package", then why not.

I mean, let consider the new command 'profile' with subcommands:
 - guix profile new
 - guix profile list
etc. i.e., managing the profiles as it has been recently described
(sorry too lazy to correctly refer where, but e.g., this thread [1]).

*and* with the subcommand 'create' or 'temporary' or <naming-is-hard>,
i.e., "guix profile create" doing more or less what the current "guix
environment" is doing.

Wouaouw! It is far far away from the initial idea behind this thread. :-)


[1] https://lists.gnu.org/archive/html/guix-devel/2019-10/msg00565.html




All the best,
simon

Re: bug#38529: Make --ad-hoc the default for guix environment proposed deprecation mechanism

[Arne Babenhauserheide]

[-- Attachment #1: Type: text/plain, Size: 636 bytes --]

Hi zimoun,

zimoun <zimon.toutoune@gmail.com> writes:
> Konrad's example. So, nothing new on the table; except you are
> starting to throw "feelings" with the "traumatic change" words.

I do not see this as feelings, but as strategy. That’s what the article
is about: Many small breakages add up, and repeated changes to
best-practices also add up.

The volatile software article describes that software differs in how
much work it is to keep using it. The traumatic change article discusses
one aspect why people stop using projects.

Best wishes,
Arne
--
Unpolitisch sein
heißt politisch sein
ohne es zu merken

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 1076 bytes --]

Re: bug#38529: Make --ad-hoc the default for guix environment proposed deprecation mechanism

[zimoun]

Hi Arne,

First, do not take me wrong, I am not "fighting" or not going to an
"heated debate".
I am fine and I hope you are also fine.
As I said my opinion in other emails, I am not repeating here. Well, I
am not convinced it is the good one, but as I trust collective power,
I am sure Guix will find the best consensus. I am even calling since
the very beginning of this discussion to collect opinions from the
other fellow hackers.


Expressing the feelings is better than bitterness. Therefore I express
mines. :-)
I could send that privately because I am not sure it deserves to be
public. But let wash the laundry in family (translation from French
expression ;-))


On Sat, 21 Dec 2019 at 00:02, Arne Babenhauserheide <arne_bab@web.de> wrote:

> > Konrad's example. So, nothing new on the table; except you are
> > starting to throw "feelings" with the "traumatic change" words.
>
> I do not see this as feelings, but as strategy. That’s what the article
> is about: Many small breakages add up, and repeated changes to
> best-practices also add up.

Just to be on the same wavelength, traumatic means in the Collins
Dictionnary: "A 'traumatic' experience is very shocking and upsetting,
and may cause psychological damage."

https://www.collinsdictionary.com/dictionary/english/traumatic

Well, to me it could make sense in the context of the mentioned blog.
Even if I feel this very opinionated. Not to say it could hurt me; bah
I am a big boy, that's ok.

Again, to be on the same wavelength, the blog says: "The result has
been hugely divisive and intimately familiar to anyone who works with
Python, creating massive rifts in the community and wasting millions
of hours of engineer time addressing. This kind of “strong” trauma is
fairly easy to spot in advance."

Well, I understand when speaking about Python. Are we comparing the
number of Guix users with the number of Python users? Are we comparing
the number of changes between Python 2 and 3 with the change of the
default "guix environment foo"? And not all the "guix environment"
behaviour, only a specific case.

Ok, maybe we are talking about the other trauma. The blog explains:
"Since nothing has actually broken with this change, the effects are
more subtle than with strong traumatic changes." and then "The
opportunity to solve this problem by rewriting with asyncio in mind,
however, also presents me a chance to rewrite in anything else, and
reevaluate my choice of Python for the project entirely."

I am sorry, I do not understand. I am probably too dumb. On one hand,
the issue of "guix environment" is the very backward compatibility so
are we really talking about this second "trauma"? On the other hand,
because "guix environment" will be better and users probably need to
rethink how they use Guix, then they will fully drop Guix.


Maybe "feelings" (quoting, in citation quoted too) is not the right
word. My point is all is vague. Example: I have the feeling that my
students(*) do not like Scheme; do I need to switch next year to
another language? Then do I make my decision based on my feelings?
based on the feelings of the students who are retaking the year (could
be shocked)? Me, I will make my decision based on: how many students
failed? what do they understand? what could be better for all the
students? what could be a better language? what is the ratio between
the new student vs the retaking ones? how many length the Scheme
textbook is? etc. Well, analogy is just analogy.


Well, that's it. I expressed what it appears to me a trail going
nowhere. Let move forward and put energy in "backward compatibility"
discussion: does Guix want? what does it imply? which level? etc. for
example, your interesting input "GUIX_ENVIRONMENT_STABLE=1".

All the best,
simon

(*) hypothetical, I do not have real students, even if I teach a bit.
And we use Python as introduction to implemented algorithms after 1
year fighting to switch from C. Whatever! :-)

bug#38529: Make --ad-hoc the default for guix environment proposed deprecation mechanism

[Konrad Hinsen]

Hi Simon,

> Assuming "guix environment" would stay and following the proposed
> plan, you would need to add GUIX_ENVIRONMENT_DEPRECATED=1 on the top
> of your script. In this would not be a problem for travelling back in
> time.

The problem is not how I update my scripts - I can manage that, no
matter what it takes.

The problem is scripts circulating in public repositories, tutorials,
etc. New users will find them and use them for inspiration. It's very
discouraging to see examples from tutorials fail or do something weird.

The main precedent is the Python 2->3 transition. There are tons of
GitHub repositories with Python code but no indication if it's 2, 3, or
both. I even had to use one that executed with either 2 or 3, but gave
different results. It takes a lot of motivation to persist.

For guix, there's the additional issue that we use the reproducibility
of builds as an argument. Non-reproducible examples are then a bit of a
credibility problem.

Cheers,
  Konrad.

bug#38529: Make --ad-hoc the default for guix environment proposed deprecation mechanism

[zimoun]

Hi Konrad,

On Fri, 20 Dec 2019 at 12:24, Konrad Hinsen <konrad.hinsen@fastmail.net> wrote:

> The problem is scripts circulating in public repositories, tutorials,
> etc. New users will find them and use them for inspiration. It's very
> discouraging to see examples from tutorials fail or do something weird.

As I said, I am not convinced because it lacks concrete examples.
Personally, I do not know Guix ressource outside the Guix ecosystem.


> The main precedent is the Python 2->3 transition. There are tons of
> GitHub repositories with Python code but no indication if it's 2, 3, or
> both. I even had to use one that executed with either 2 or 3, but gave
> different results. It takes a lot of motivation to persist.

Except that "guix environment" will raise warnings.
Whatever.


> For guix, there's the additional issue that we use the reproducibility
> of builds as an argument. Non-reproducible examples are then a bit of a
> credibility problem.

I agree.
I do not want to fight about "backward compatibility".


As I said, talking about "guix environment", my opinion is that the
cost of the change is low.
However, we cannot know this cost, only probe and estimate: using my
probings, I estimate the cost is low.

IMHO, in this case, there is 2 ways to make a decision:
 - more probings to estimate more precisely;
or
 - say: "no backward compatibility breakage"

I am fine with both. :-)
 - I report my use-case: no cost at all
 - I propose the name "guix shell"


However, I feel I have spent enough time and energy on this topic and
I feel a blocking situation so I will move forward to another topic.
:-)

All the best,
simon

bug#38529: Make --ad-hoc the default for guix environment proposed deprecation mechanism

[Ricardo Wurmus]

zimoun <zimon.toutoune@gmail.com> writes:

>  - I propose the name "guix shell"

This is not a bad idea, especially considering that “guix environment”
was meant to get shebang support, so that you could use it as the
interpreter in a script that handles the environment configuration.

It is also shorter.

--
Ricardo

bug#38529: Make --ad-hoc the default for guix environment proposed deprecation mechanism

[Danny Milosavljevic]

[-- Attachment #1: Type: text/plain, Size: 649 bytes --]

Hi Ricardo,

On Fri, 20 Dec 2019 22:08:48 +0100
Ricardo Wurmus <rekado@elephly.net> wrote:

> zimoun <zimon.toutoune@gmail.com> writes:
> 
> >  - I propose the name "guix shell"  
> 
> This is not a bad idea, especially considering that “guix environment”
> was meant to get shebang support, so that you could use it as the
> interpreter in a script that handles the environment configuration.

Note that the Linux kernel shebang interpreter only supports ONE argument.
The good news is that whatever number of arguments you pass, it will all
be subsumed into the first argument.

#!foo bar baz

foo gets: $1="bar baz", $#=2

[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

bug#38529: Make --ad-hoc the default for guix environment proposed deprecation mechanism

[Andy Wingo]

On Fri 20 Dec 2019 22:08, Ricardo Wurmus <rekado@elephly.net> writes:

> zimoun <zimon.toutoune@gmail.com> writes:
>
>>  - I propose the name "guix shell"
>
> This is not a bad idea, especially considering that “guix environment”
> was meant to get shebang support, so that you could use it as the
> interpreter in a script that handles the environment configuration.
>
> It is also shorter.

I like this idea.  It would also allow us to deprecate "guix
environment" over a period of a year or so, and we can probably show an
equivalent "guix shell" invocation in the deprecation message.

Andy

bug#38529: Make --ad-hoc the default for guix environment proposed deprecation mechanism

[Ricardo Wurmus]

Konrad Hinsen <konrad.hinsen@fastmail.net> writes:

>> Maybe I miss a point. It is not: "watch out, this will do something
>> else in the future" but "watch out, this was doing something else in
>> the past and the change happened the <date> in <commit>".
>
> Concrete example: I am writing a tutorial about using Guix for
> reproducible research. It shows several uses of "guix environment", some
> of them without '–add-hoc' or '–inputs-of'. I know my examples will
> cease to work in a few months. What am I supposed to do about this?

I wonder if we should simply bump the version number to indicate that
this is a breaking change?

Another more difficult option would be to do what responsible API
developers on the web do: to version their API and to make the API
version selectable.  I don’t know *how* to do this elegantly, and
there’s a real maintenance cost (it seems small in this case), but
configuration files can be used for changing new defaults.

--
Ricardo

bug#38529: Make --ad-hoc the default for guix environment proposed deprecation mechanism

[Konrad Hinsen]

Hi Ricardo,

> I wonder if we should simply bump the version number to indicate that
> this is a breaking change?

That's a possibility, but who ever looks at Guix version numbers?

> Another more difficult option would be to do what responsible API
> developers on the web do: to version their API and to make the API
> version selectable.  I don’t know *how* to do this elegantly, and

That's an interesting idea which would also take care of similar
situations in the future.

One way to implement this is to have executables "guix1", "guix2"
etc. Most users would then define an alias "guix" for interactive use,
but hopefully script authors would use the versioned executables.

Cheers,
  Konrad.

bug#38529: Deprecating ‘guix environment’?

[Ludovic Courtès]

Hi Konrad,

Konrad Hinsen <konrad.hinsen@fastmail.net> skribis:

> On 16/12/2019 23:09, Ludovic Courtès wrote:
>> So in a more algorithmic manner:
>>> 1. if ad-hoc and inputs-of is present at the same invocation: fail
>>> hard. (With an error like incompatible options present)
>>> 2. if only ad-hoc is present, then print a deprecation warning (yes,
>>> we could make this suspendable with an environment variable, like you
>>> described)
>>> 3. if only inputs-of present, then do the new behaviour.
>>> 4. if neither ad-hoc nor inputs-of present then
>>>    a. if GUIX_ENVIRONMENT_DEPRECATED is 1: do the current behaviour,
>>>    b. if GUIX_ENVIRONMENT_DEPRECATED is undefined, or is not 1: do the
>>> new behaviour.
>> That sounds like a good plan to me.
>>
>> #4 is the trickiest, and I think it’d be good to give users a bit of
>> time so they can start adjusting before deprecation is in effect.
>
> #4 is trickiest for another reason: there is no future-proof use of
> "guix environment" that works right now and will continue to work. Nor
> is there any way to see, when looking at a command line, whether it's
> old-style or new-style, if neither --ad-hoc nor --inputs-of are
> present. This means that all existing documentation (tutorials etc.)
> will become misleading in the future. Worse, even documentation
> written today, in full awareness of a coming change, can't do better
> than saying "watch out, this will do something else in the future".
>
> The first rule of backwards-compatibility is: never change the meaning
> of an existing valid command/API. Add new valid syntax, deprecate old
> valid syntax, but don't change the meaning of something that was and
> will be valid.

Yeah.

Clearly there’s a tension between that and keeping Guix open to changes.

> How about a more drastic measure: deprecate "guix environment" and
> introduce a new subcommand with the desired new behaviour?

That has the advantage of avoiding the problem you mention altogether
while also allowing for further changes.

The hard question then becomes: what do we call it?  I vote against
abbreviations.  :-)

Also, what other goals would we set for that command?  How would we
frame it in the set of commands?

Thanks,
Ludo’.

bug#38529: Deprecating ‘guix environment’?

[zimoun]

On Thu, 19 Dec 2019 at 17:31, Ludovic Courtès <ludo@gnu.org> wrote:

> The hard question then becomes: what do we call it?  I vote against
> abbreviations.  :-)

"guix shell"?



Cheers,
simon

bug#38529: Deprecating ‘guix environment’?

[Konrad Hinsen]

Hi Ludo,

> Clearly there’s a tension between that and keeping Guix open to changes.

That's indeed the main problem and here as elsewhere, it is often a
topic of heated arguments.

My point of view (long form:
https://hal.archives-ouvertes.fr/hal-02117588)
is that software projects should adopt a backwards compatibility policy
early on, state it clearly in their documentation, and stick to it. That
prevents misunderstandings, bad surprises, and heated debates.

As for what that policy should be for Guix, that's a more difficult
story. For projects with versioned releases, I like the principles
of semantic versioning, but Guix is more of a rolling-release
project. (Test question: does anyone know what the current Guix version
number is? Does anyone care?) I am not aware of any good precedents
in terms of policy for such projects.

> The hard question then becomes: what do we call it?  I vote against
> abbreviations.  :-)
>
> Also, what other goals would we set for that command?  How would we
> frame it in the set of commands?

I vote for discussing the second point before the first one. Names
should reflect the functionality behind them.

How about a unified command for constructing environments and profiles
declaratively? In other words, combine "guix environment" and the
declarative parts of "guix package". We could probably get rid of
the somewhat obscure "guix environment -r" in the process.

Cheers,
  Konrad.

bug#38529: Deprecating ‘guix environment’?

[zimoun]

Hi Konrad,


On Fri, 20 Dec 2019 at 12:18, Konrad Hinsen <konrad.hinsen@fastmail.net> wrote:

> My point of view (long form:
> https://hal.archives-ouvertes.fr/hal-02117588)
> is that software projects should adopt a backwards compatibility policy
> early on, state it clearly in their documentation, and stick to it. That
> prevents misunderstandings, bad surprises, and heated debates.

Thank you for the pointer. I have not read yet.
I agree with the compatibility policy and this argument has been
raises in the "heated" debate with Arne. :-)


> As for what that policy should be for Guix, that's a more difficult
> story. For projects with versioned releases, I like the principles

The first idea which comes in mind is to introduce a pledge. Maybe in
the introduction.

"The Guix project pledges to keep backward compatibility... blabla".

However, the real question is at which level.
At the CLI level? At the exported scheme functions? All modules or
specific ones?


> of semantic versioning, but Guix is more of a rolling-release
> project. (Test question: does anyone know what the current Guix version
> number is? Does anyone care?) I am not aware of any good precedents
> in terms of policy for such projects.

I agree.

I proposed [1] to add "tags" in the meaning of "git tag". Initially,
to ease the navigation through the history when searching for
packages.
Re-hashing this "guix tag" or "guix pull --tag" proposal, one idea
could be to introduce tags, say v1.1, v1.2, v1.3 etc bumping the
version every X months, or after each core-update merge, or after
<you-name-it>, then by default "guix pull" would update to the tags.
This adds "stability" because we could tag commits that we know are
stable (no "guix pull" break, etc.)

[1] https://lists.gnu.org/archive/html/guix-devel/2019-11/msg00513.html


> > The hard question then becomes: what do we call it?  I vote against
> > abbreviations.  :-)
> >
> > Also, what other goals would we set for that command?  How would we
> > frame it in the set of commands?
>
> I vote for discussing the second point before the first one. Names
> should reflect the functionality behind them.

The starting point seems:
 - https://lists.gnu.org/archive/html/guix-devel/2017-08/msg00300.html
 - what do you feel missing about "guix environment"?

Considering my use-case, I am mostly aligned with "The future of 'guix
environment'".



All the best,
simon

bug#38529: Make --pure the default for `guix environment'?

[Leo Famulari]

On Sun, Dec 08, 2019 at 04:42:07PM +0100, Pierre Neidhardt wrote:
> --pure seems to be the more sensible behaviour.  "Impure" environments
> can have unexpected behaviours, so it makes sense to only allow them
> when the user explicitly asks for it.

I don't have an opinion about this in general except that I think that
--ad-hoc should continue to work the way it does now, without --pure
being implied.

bug#38529: Make --pure the default for `guix environment'?

[Maxim Cournoyer]

Hi Pierre,

Pierre Neidhardt <mail@ambrevar.xyz> writes:

> --pure seems to be the more sensible behaviour.  "Impure" environments
> can have unexpected behaviours, so it makes sense to only allow them
> when the user explicitly asks for it.

Unfortunately Guix packages often don't work well with --pure.  Be it
magit that depends on git, or Emacs that depend or coreutils, etc.,
there are many things that are expected to be propagated and aren't
explicitly, by omission or sometimes for closure's size sake (when the
feature is optional).  We could argue that is a good reason for the
proposed change :-).

I think environments are great mostly for hacking and trying stuff
quickly, where the guarantees of Guix do not matter as much as for
profiles (and if they did, you'd be better with guix environment
--container anyway).

So, I guess that makes me more on the side of "let's no change the
defaults for now".

Maxim

bug#38529: Make --pure the default for `guix environment'?

[Maxim Cournoyer]

tags 38529 +wontfix
thanks

Hi,

Maxim Cournoyer <maxim.cournoyer@gmail.com> writes:

> Hi Pierre,
>
> Pierre Neidhardt <mail@ambrevar.xyz> writes:
>
>> --pure seems to be the more sensible behaviour.  "Impure" environments
>> can have unexpected behaviours, so it makes sense to only allow them
>> when the user explicitly asks for it.
>
> Unfortunately Guix packages often don't work well with --pure.  Be it
> magit that depends on git, or Emacs that depend or coreutils, etc.,
> there are many things that are expected to be propagated and aren't
> explicitly, by omission or sometimes for closure's size sake (when the
> feature is optional).  We could argue that is a good reason for the
> proposed change :-).
>
> I think environments are great mostly for hacking and trying stuff
> quickly, where the guarantees of Guix do not matter as much as for
> profiles (and if they did, you'd be better with guix environment
> --container anyway).
>
> So, I guess that makes me more on the side of "let's no change the
> defaults for now".

Closing this old forgotten issue as wontfix.

Thanks,

Maxim

bug#38529: Make --pure the default for `guix environment'?

[Jesse Gibbons]

On Sun, 2019-12-08 at 16:42 +0100, Pierre Neidhardt wrote:
> --pure seems to be the more sensible behaviour.  "Impure" environments
> can have unexpected behaviours, so it makes sense to only allow them
> when the user explicitly asks for it.
> 
--pure environments sometimes miss important environment variables. Try
running any app that depends on X and doesn't fallback to a console mode in
a pure environment.
for example, "guix environment --pure --ad-hoc pavucontrol -- pavucontrol"
gives me an error: 
No protocol specified
Unable to init server: Could not connect: Connection
refused



Similarly, "guix environment --pure --ad-hoc gnubik -- gnubik" gives me an
error:
No protocol specified

(gnubik:29707): Gtk-WARNING **: 10:03:28.753: cannot open display: :1





"guix environment --pure --ad-hoc gedit -- gedit":
No protocol specified
Unable to init server: Could not connect: Connection refused

(org.gnome.gedit:31542): Gtk-WARNING **: 10:08:07.401: cannot open display:
:1



Making --pure the default for "guix environment" would make things more
complicated for users wanting to temporarily run GUI apps unless we fix this
issue first. I furthermore suspect some tests fail because of this issue.

bug#38529: Make --pure the default for `guix environment'?

[zimoun]

Hi,

On Mon, 9 Dec 2019 at 18:38, Jesse Gibbons <jgibbons2357@gmail.com> wrote:

> for example, "guix environment --pure --ad-hoc pavucontrol -- pavucontrol"
> gives me an error:
> No protocol specified
> Unable to init server: Could not connect: Connection
> refused

I do not experiment this error.


> Similarly, "guix environment --pure --ad-hoc gnubik -- gnubik" gives me an
> error:
> No protocol specified
>
> (gnubik:29707): Gtk-WARNING **: 10:03:28.753: cannot open display: :1

I do not experiment this error.


> "guix environment --pure --ad-hoc gedit -- gedit":
> No protocol specified
> Unable to init server: Could not connect: Connection refused
>
> (org.gnome.gedit:31542): Gtk-WARNING **: 10:08:07.401: cannot open display:
> :1

Instead I experiment this warning:

(org.gnome.gedit:19087): dconf-WARNING **: 20:31:27.734: failed to
commit changes to dconf: Failed to execute child process ?dbus-launch?
(No such file or directory)



All the best,
simon