Re: master 49bad2a: * lisp/cedet/semantic/analyze*.el: Use lexical-binding

Re: master 49bad2a: * lisp/cedet/semantic/analyze*.el: Use lexical-binding

[Lars Ingebrigtsen]

I'm not at all sure what's going wrong in the tests today, but perhaps
it's related to this change?

If I say "make -j16 check", I get:

SUMMARY OF TEST RESULTS
-----------------------
Files examined: 348
Ran 4741 tests, 4655 results as expected, 7 unexpected, 79 skipped
1 files contained unexpected results:
  lisp/emacs-lisp/eieio-tests/eieio-tests.log

However, if I say

"cd test; make -j16" I do not get this test failure.  Instead I get

Ran 4741 tests, 4661 results as expected, 1 unexpected, 79 skipped
1 files contained unexpected results:
  lisp/calendar/time-date-tests.log

Both these things seem to be repeatable on two different Debian bullseye
systems...  most confusing...

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no

Re: master 49bad2a: * lisp/cedet/semantic/analyze*.el: Use lexical-binding

[Lars Ingebrigtsen]

Lars Ingebrigtsen <larsi@gnus.org> writes:

> Ran 4741 tests, 4661 results as expected, 1 unexpected, 79 skipped
> 1 files contained unexpected results:
>   lisp/calendar/time-date-tests.log

This comes from:

Test test-obsolete-encode-time-value condition:
    (void-function internal--with-suppressed-warnings)

And, indeed, this gives a backtrace:

(should (equal (with-suppressed-warnings ((obsolete encode-time-value))
                 (encode-time-value 1 2 3 4 0))
               '(1 . 2)))

Eval-ing just the thing inside the `should' works, though.

So something has happened to macro expansion over the past day or so, I
think?

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no

Re: master 49bad2a: * lisp/cedet/semantic/analyze*.el: Use lexical-binding

[Stefan Monnier]

> I'm not at all sure what's going wrong in the tests today, but perhaps
> it's related to this change?

Do you say that because of the "bad" in "49bad2a"?


        Stefan

Re: master 49bad2a: * lisp/cedet/semantic/analyze*.el: Use lexical-binding

[Basil L. Contovounesios]

Lars Ingebrigtsen <larsi@gnus.org> writes:

> I'm not at all sure what's going wrong in the tests today, but perhaps
> it's related to this change?

Doesn't bug#46818 seem a likelier culprit at first glance?

[...]

> So something has happened to macro expansion over the past day or so, I
> think?

-- 
Basil

Re: master 49bad2a: * lisp/cedet/semantic/analyze*.el: Use lexical-binding

[Stefan Monnier]

> SUMMARY OF TEST RESULTS
> -----------------------
> Files examined: 348
> Ran 4741 tests, 4655 results as expected, 7 unexpected, 79 skipped
> 1 files contained unexpected results:
>   lisp/emacs-lisp/eieio-tests/eieio-tests.log

AFAICT this is now fixed,


        Stefan

Re: master 49bad2a: * lisp/cedet/semantic/analyze*.el: Use lexical-binding

[Lars Ingebrigtsen]

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

>> I'm not at all sure what's going wrong in the tests today, but perhaps
>> it's related to this change?
>
> Do you say that because of the "bad" in "49bad2a"?

Indeed.  Not many people know that this is how Git tags commits that
aren't as good as they could be.

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no

Re: master 49bad2a: * lisp/cedet/semantic/analyze*.el: Use lexical-binding

[Lars Ingebrigtsen]

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

>> SUMMARY OF TEST RESULTS
>> -----------------------
>> Files examined: 348
>> Ran 4741 tests, 4655 results as expected, 7 unexpected, 79 skipped
>> 1 files contained unexpected results:
>>   lisp/emacs-lisp/eieio-tests/eieio-tests.log
>
> AFAICT this is now fixed,

Yup; this now works.  This still fails, though:

1 unexpected results:
   FAILED  test-obsolete-encode-time-value

Confusingly enough, it only fails with "cd test; make" and not "make
check" in the directory above...

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no

Re: master 49bad2a: * lisp/cedet/semantic/analyze*.el: Use lexical-binding

[Stefan Monnier]

>>> SUMMARY OF TEST RESULTS
>>> -----------------------
>>> Files examined: 348
>>> Ran 4741 tests, 4655 results as expected, 7 unexpected, 79 skipped
>>> 1 files contained unexpected results:
>>>   lisp/emacs-lisp/eieio-tests/eieio-tests.log
>>
>> AFAICT this is now fixed,
>
> Yup; this now works.  This still fails, though:
>
> 1 unexpected results:
>    FAILED  test-obsolete-encode-time-value

I suspect this failed before as well.
Should be fixed now, tho.

I still get one error when I do (cd test; make):

    1 files contained unexpected results:
      lisp/net/ntlm-tests.log

but I hope it has nothing to do with my recent changes.

> Confusingly enough, it only fails with "cd test; make" and not "make
> check" in the directory above...

I must admit I don't understand in which way those two are different
(nor how they compare to `make test/lisp/calendar/time-date-tests`
which may also give different results).


        Stefan

Re: master 49bad2a: * lisp/cedet/semantic/analyze*.el: Use lexical-binding

[Lars Ingebrigtsen]

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

>> 1 unexpected results:
>>    FAILED  test-obsolete-encode-time-value
>
> I suspect this failed before as well.
> Should be fixed now, tho.

Yup; this works well now here too.

> I still get one error when I do (cd test; make):
>
>     1 files contained unexpected results:
>       lisp/net/ntlm-tests.log
>
> but I hope it has nothing to do with my recent changes.

I'm not getting that failure...  some of the tests depend on some things
in GNU ELPA, but they should be disabled if it can't find the resources
there.  But perhaps those checks are failing in your setup?

(Added Thomas to the CCs.)

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no