Re: feature/native-comp 007a5a6 1/2: [WIP] Add and improve section headings

Re: feature/native-comp 007a5a6 1/2: [WIP] Add and improve section headings

[Stefan Kangas]

jonas@bernoul.li (Jonas Bernoulli) writes:

> branch: feature/native-comp
> commit 007a5a677573ab628426a0605eb38f8e68fe1953
> Author: Jonas Bernoulli <jonas@bernoul.li>
> Commit: Jonas Bernoulli <jonas@bernoul.li>
>
>     [WIP] Add and improve section headings
> ---
>  lisp/allout.el    | 9 +++++----
>  lisp/icomplete.el | 9 +++++----
>  lisp/net/eudc.el  | 8 ++++++--
>  3 files changed, 16 insertions(+), 10 deletions(-)

Is there any particular reason for making this change on the native-comp
branch and not on master?

Re: feature/native-comp 007a5a6 1/2: [WIP] Add and improve section headings

[Andrea Corallo via Emacs development discussions.]

Stefan Kangas <stefankangas@gmail.com> writes:

> jonas@bernoul.li (Jonas Bernoulli) writes:
>
>> branch: feature/native-comp
>> commit 007a5a677573ab628426a0605eb38f8e68fe1953
>> Author: Jonas Bernoulli <jonas@bernoul.li>
>> Commit: Jonas Bernoulli <jonas@bernoul.li>
>>
>>     [WIP] Add and improve section headings
>> ---
>>  lisp/allout.el    | 9 +++++----
>>  lisp/icomplete.el | 9 +++++----
>>  lisp/net/eudc.el  | 8 ++++++--
>>  3 files changed, 16 insertions(+), 10 deletions(-)
>
> Is there any particular reason for making this change on the native-comp
> branch and not on master?

Wow native-comp is the new master! :)

Re: feature/native-comp c36b4ee 2/2: [TODO] Remove noisy anti-noise feature

[Robert Pluim]

jonas@bernoul.li (Jonas Bernoulli) writes:

> branch: feature/native-comp
> commit c36b4eed2d76f0e804d27d35dd6281f858639f94
> Author: Jonas Bernoulli <jonas@bernoul.li>
> Commit: Jonas Bernoulli <jonas@bernoul.li>
>
>     [TODO] Remove noisy anti-noise feature
> ---
>  lisp/emacs-lisp/warnings.el | 11 -----------
>  1 file changed, 11 deletions(-)
>

Everyone is allowed to have an opinion on this feature, but just
removing it without discussion seems a little harsh.

Robert

Re: feature/native-comp c36b4ee 2/2: [TODO] Remove noisy anti-noise feature

[Jonas Bernoulli]

Robert Pluim <rpluim@gmail.com> writes:

> jonas@bernoul.li (Jonas Bernoulli) writes:
>
>> branch: feature/native-comp
>> commit c36b4eed2d76f0e804d27d35dd6281f858639f94
>> Author: Jonas Bernoulli <jonas@bernoul.li>
>> Commit: Jonas Bernoulli <jonas@bernoul.li>
>>
>>     [TODO] Remove noisy anti-noise feature
>> ---
>>  lisp/emacs-lisp/warnings.el | 11 -----------
>>  1 file changed, 11 deletions(-)
>>
>
> Everyone is allowed to have an opinion on this feature, but just
> removing it without discussion seems a little harsh.
>
> Robert

That was an accidental push.  I do think it would be best to remove
this new feature completely, but of course I would have raised the
issue before *intentionally* removing it.

This is why I think this is not a good feature:

People will make use of it.

Longer version: Of course it is nice for users who don't know any elisp
to opt-out of seeing warnings that they cannot understand, but I fear
that users who could contribute by addressing these issues or by at
least reporting them, will opt-out as well.

Even some package authors are going to enable this, possibly because
they misunderstand what it does.  I did that for example.  I thought it
would disable "that particular" warning; not all of them.  I then had to
figure out how to revert that setting, there wasn't any button for that.

Also in some contexts the buttons cause the actual warning message to be
truncated.  For example lazy native compilation shows only the first
line of each warning and if a third of that line is used up by the
buttons that are the same for each and every warning, that can turn a
sufficiently informative warning into guess work.

I am involved in an uphill battle, fixing the endless stream of compile
warnings for the packages that are installed on my system.  It would be
nice if more people did that.  Yes, these warning are a bit annoying but
the underlying issues are not going away if we make it easier to ignore
them.

I don't use package.el and don't know what happens if there are warnings
during package installation.  If the problem is that noobs get confused
because a buffer pops up with warnings that they don't understand, then
we could prevent that buffer from poping up by default.  Going further
than that seems like an own goal.

     Jonas

Silencing warnings with a button (was: feature/native-comp c36b4ee 2/2: [TODO] Remove noisy anti-noise feature)

[Stefan Monnier]

> This is why I think this is not a good feature:
>
> People will make use of it.
>
> Longer version: Of course it is nice for users who don't know any elisp
> to opt-out of seeing warnings that they cannot understand, but I fear
> that users who could contribute by addressing these issues or by at
> least reporting them, will opt-out as well.

FWIW, I tend to agree.  I understand that exposing end-users to warnings
is somewhat problematic, but I think a better solution is to try and
make sure they're presented in a way that is understandable.  E.g. make
sure that the users don't go "oh god, it's all messed up now, what do I do?",
and instead somehow explain:
- these are usually harmless (for the end users).
- these should be reported to the package maintainers (some of whom
  would never see these warnings otherwise, because they never compile
  their own package nor install it via package.el).

E.g., instead of having buttons to silence the warnings, maybe we should
have buttons to "report to maintainers".


        Stefan

Re: feature/native-comp c36b4ee 2/2: [TODO] Remove noisy anti-noise feature

[Andrea Corallo via Emacs development discussions.]

Jonas Bernoulli <jonas@bernoul.li> writes:

> Robert Pluim <rpluim@gmail.com> writes:
>
>> jonas@bernoul.li (Jonas Bernoulli) writes:
>>
>>> branch: feature/native-comp
>>> commit c36b4eed2d76f0e804d27d35dd6281f858639f94
>>> Author: Jonas Bernoulli <jonas@bernoul.li>
>>> Commit: Jonas Bernoulli <jonas@bernoul.li>
>>>
>>>     [TODO] Remove noisy anti-noise feature
>>> ---
>>>  lisp/emacs-lisp/warnings.el | 11 -----------
>>>  1 file changed, 11 deletions(-)
>>>
>>
>> Everyone is allowed to have an opinion on this feature, but just
>> removing it without discussion seems a little harsh.
>>
>> Robert
>
> That was an accidental push.  I do think it would be best to remove
> this new feature completely, but of course I would have raised the
> issue before *intentionally* removing it.

Hi Jonas,

thanks for the revert, please next time specify the reason of the revert
in the commit, (as I've recently discovered) this is the conventional
way to proceed :)

Ciao

  Andrea