bug#31978: 26.1.50; "s" in gnus *Group* error: "eieio-override-prin1: Apparently circular structure being printed"

[Tim Landscheidt]

Eric Abrahamsen <eric@ericabrahamsen.net> wrote:

>>>>>> I am using Gnus in Emacs 26.

>>>>> Would you mind cherry-picking or otherwise applying the patch in
>>>>> ffbb4e8d54, and see if that solves the problem? I think it probably
>>>>> will.

>>>> As I ran into this issue today (with Emacs 26.1): To apply
>>>> the patch, is it sufficient to redefine
>>>> eieio-persistent-validate/fix-slot-value in the running in-
>>>> stance, or do I need to do something more?

>>> That would suffice, but it wouldn't guarantee that the fix was available
>>> next time you start Emacs. For that, you'd want to make sure that the
>>> code was changed in your Emacs codebase, and then run "make" or
>>> otherwise ensure it was compiled in. Depends on how you run Emacs.

>> I redefined eieio-persistent-validate/fix-slot-value as per
>> ffbb4e8d542df44ced5afd89221b0dfb234d8525 in Emacs 26.1, and
>> it did not solve this issue, still emitting:

>> | eieio-override-prin1: Apparently circular structure being printed

>> (and barring Gnus from exiting).

> Hmm, more things have been changed in this area since I worked on it --
> is there any chance you could switch to 26.2? In particular, commit
> 5f01af6c8e almost certainly is involved.

> […]

I'm using my distribution-supplied Emacs and I don't want to
change anything major for my real-life setup (prior to my
distribution forcing me).

But the error made me think about why I had enabled the re-
gistry in the first place, and I didn't come up with a good
reason, so I disabled it and am no longer directly affected
by this bug (and have a few GBs of memory more, and Gnus
starts and quits in seconds instead of minutes :-)).