The problem might be due to this that I have several (defvar …) in the
ses-test.el, I probably added this in order to avoid some bytecompile
errors, plus somebody has added the cookie -*- lexical-binding: t; -*-
to the ses-test.el file.

SES uses buffer local variables for the cell symbols, so the interned
symbols referring to cells are per se dynamically / not lexically,
bound.

I suspect that the addition of the -*- lexical-binding: t; -*- cookies is
the root cause of the test being erroneous. If we need that cooky, then
I need another way to declare the cell symbols, maybe declaring them in
an encapsulating (let (…) …) statement and removing the (defvar …) would
solve the issue.

  V.
________________________________
De : Vincent Belaïche <vincent.b.1@hotmail.fr>
Envoyé : mardi 5 avril 2022 13:46
À : Stefan Monnier <monnier@iro.umontreal.ca>
Cc : emacs-devel <emacs-devel@gnu.org>; andrés ramírez <rrandresf@gmail.com>
Objet : RE: SES case insensitive ses-jump git-pushed

To be fully honest, if I M-x ert-run-tests-interactively a second time,
then I get this:

--------8<-----------8<-----------8<-----------8<-----------8<-------
Selector: t
Passed:  11
Failed:  2 (2 unexpected)
Skipped: 0
Total:   13/13

Started at:   2022-04-05 13:44:31+0200
Finished.
Finished at:  2022-04-05 13:44:31+0200

..........FF.

F ses-tests-renamed-cell-after-setting
    Check that setting A1 to 1 and A2 to (1+ A1), and then
    (ert-test-failed
     ((should
       (eq
(bound-and-true-p A2)
2))
      :form
      (eq *error* 2)
      :value nil))

F ses-tests-renamed-cells-row-insertion
    Check that setting A1 to 1 and A2 to (1+ A1), and then renaming A1 to ‘ses--foo’ and A2 to ‘ses--bar’ jumping
    (ert-test-failed
     ((should
       (eq ses--bar 2))
      :form
      (eq *error* 2)
      :value nil))
--------8<-----------8<-----------8<-----------8<-----------8<-------

It seems that the new tests are maybe not to blame, but that my existing
tests were wrong, as if there is some border effect remaining after the
tests have run, and spoiling a new test run.

  V.
________________________________
De : Vincent Belaïche <vincent.b.1@hotmail.fr>
Envoyé : mardi 5 avril 2022 13:43
À : Stefan Monnier <monnier@iro.umontreal.ca>
Cc : emacs-devel <emacs-devel@gnu.org>; andrés ramírez <rrandresf@gmail.com>
Objet : RE: SES case insensitive ses-jump git-pushed

Salut Stefan,

When I eval-buffer on ses-test.el, and then when I do M-x
ert-run-tests-interactively I get no error, the output in *ert* buffer
is as follows:

--------8<-----------8<-----------8<-----------8<-----------8<-------
Selector: t
Passed:  13
Failed:  0
Skipped: 0
Total:   13/13

Started at:   2022-04-05 13:41:09+0200
Finished.
Finished at:  2022-04-05 13:41:09+0200

.............
--------8<-----------8<-----------8<-----------8<-----------8<-------

  V.
________________________________
De : andrés ramírez <rrandresf@gmail.com>
Envoyé : mardi 5 avril 2022 05:33
À : Stefan Monnier <monnier@iro.umontreal.ca>
Cc : Vincent Belaïche <vincent.b.1@hotmail.fr>; emacs-devel <emacs-devel@gnu.org>
Objet : Re: SES case insensitive ses-jump git-pushed

>>>>> "Stefan" == Stefan Monnier <monnier@iro.umontreal.ca> writes:


[...]


    Stefan> I also see an error when I run `make test/lisp/ses-tests`.