batch breaks bugfix

batch breaks bugfix

[Samuel Wales]

can you think of any differences between bugfix and main
that could result in

1) bugfix not producing certain agenda entries from the org
   outline (THE BUG)
2) with no obvious differences between the entries that show
   and the entries that do not
3) FROM A BATCH CALL.  removing --batch from the emacs call
   removes the bug.

i am merely hoping for an a priori idea about what might be
different from bugfix to main so that i don't have to do
what for unrealted reasons i cannot do at this time which is
produce an ecm.  27.1.  apologies for bad non-bug-report.

===

(info "(elisp) Startup Summary") says what --batch does
differehtly from interactive.  long ago i tried to make
testcase be like normal emacs gui use.  loads .emacs, which
batch does not.  i have no default library for batch to not
run.  i did not do the other stuff --batch does.  recently
adding (package-initialize) made no difference.

so idk what --batch is doing or not doing to screw up the
bugfix agenda but not main.

batch maint has sorting differences from main, which i will
presume are expected and fixable.

===

i wouldn't care about this issue but i like running my old
regression script to ensure agenda works ok.


-- 
The Kafka Pandemic

A blog about science, health, human rights, and misopathy:
https://thekafkapandemic.blogspot.com

Re: batch breaks bugfix

[Ihor Radchenko]

Samuel Wales <samologist@gmail.com> writes:

> can you think of any differences between bugfix and main
> that could result in
>
> 1) bugfix not producing certain agenda entries from the org
>    outline (THE BUG)
> 2) with no obvious differences between the entries that show
>    and the entries that do not
> 3) FROM A BATCH CALL.  removing --batch from the emacs call
>    removes the bug.

If you play with visibility, there might be differences. Emacs batch
mode disables some display staff and visibility relies upon display
working properly.

No other ideas.

-- 
Ihor Radchenko // yantar92,
Org mode contributor,
Learn more about Org mode at <https://orgmode.org/>.
Support Org development at <https://liberapay.com/org-mode>,
or support my work at <https://liberapay.com/yantar92>

Re: batch breaks bugfix

[Samuel Wales]

the bug occurs in recent bugfix, but not recent main, so perhaps we
could say that it is fixed in main but that the fix has not been
applied to bugfix?  [if only the fix were known.]

does org testing use --batch?

bugfix tests ok.  but idk if that tests agenda.


On 6/18/23, Ihor Radchenko <yantar92@posteo.net> wrote:
> Samuel Wales <samologist@gmail.com> writes:
>
>> can you think of any differences between bugfix and main
>> that could result in
>>
>> 1) bugfix not producing certain agenda entries from the org
>>    outline (THE BUG)
>> 2) with no obvious differences between the entries that show
>>    and the entries that do not
>> 3) FROM A BATCH CALL.  removing --batch from the emacs call
>>    removes the bug.
>
> If you play with visibility, there might be differences. Emacs batch
> mode disables some display staff and visibility relies upon display
> working properly.
>
> No other ideas.
>
> --
> Ihor Radchenko // yantar92,
> Org mode contributor,
> Learn more about Org mode at <https://orgmode.org/>.
> Support Org development at <https://liberapay.com/org-mode>,
> or support my work at <https://liberapay.com/yantar92>
>


-- 
The Kafka Pandemic

A blog about science, health, human rights, and misopathy:
https://thekafkapandemic.blogspot.com

Re: batch breaks bugfix

[Samuel Wales]

wondering if it is possible to enable display stuff without actually
dispaying anything.

On 6/18/23, Samuel Wales <samologist@gmail.com> wrote:
> the bug occurs in recent bugfix, but not recent main, so perhaps we
> could say that it is fixed in main but that the fix has not been
> applied to bugfix?  [if only the fix were known.]
>
> does org testing use --batch?
>
> bugfix tests ok.  but idk if that tests agenda.
>
>
> On 6/18/23, Ihor Radchenko <yantar92@posteo.net> wrote:
>> Samuel Wales <samologist@gmail.com> writes:
>>
>>> can you think of any differences between bugfix and main
>>> that could result in
>>>
>>> 1) bugfix not producing certain agenda entries from the org
>>>    outline (THE BUG)
>>> 2) with no obvious differences between the entries that show
>>>    and the entries that do not
>>> 3) FROM A BATCH CALL.  removing --batch from the emacs call
>>>    removes the bug.
>>
>> If you play with visibility, there might be differences. Emacs batch
>> mode disables some display staff and visibility relies upon display
>> working properly.
>>
>> No other ideas.
>>
>> --
>> Ihor Radchenko // yantar92,
>> Org mode contributor,
>> Learn more about Org mode at <https://orgmode.org/>.
>> Support Org development at <https://liberapay.com/org-mode>,
>> or support my work at <https://liberapay.com/yantar92>
>>
>
>
> --
> The Kafka Pandemic
>
> A blog about science, health, human rights, and misopathy:
> https://thekafkapandemic.blogspot.com
>


-- 
The Kafka Pandemic

A blog about science, health, human rights, and misopathy:
https://thekafkapandemic.blogspot.com

Re: batch breaks bugfix

[Ihor Radchenko]

Samuel Wales <samologist@gmail.com> writes:

> the bug occurs in recent bugfix, but not recent main, so perhaps we
> could say that it is fixed in main but that the fix has not been
> applied to bugfix?  [if only the fix were known.]

Maybe.

> does org testing use --batch?

Yup.
If you have time and interest, it would be useful if you can integrate
your regression tests with Org's test suite.

-- 
Ihor Radchenko // yantar92,
Org mode contributor,
Learn more about Org mode at <https://orgmode.org/>.
Support Org development at <https://liberapay.com/org-mode>,
or support my work at <https://liberapay.com/yantar92>

Re: batch breaks bugfix

[Samuel Wales]

i don't do anything with visibility.  make test seems ok in bugfix.
but perhaps it does not generate a comprehensive agenda, in which case
maybe --batch would cause the bug.

what i am going to try to do is just remove --batch and lower the
frame and try to get bugfix to work like main.  this is more like
normal use.  but it would be good to know what the problem is with
--batch.


On 6/19/23, Ihor Radchenko <yantar92@posteo.net> wrote:
> Samuel Wales <samologist@gmail.com> writes:
>
>> the bug occurs in recent bugfix, but not recent main, so perhaps we
>> could say that it is fixed in main but that the fix has not been
>> applied to bugfix?  [if only the fix were known.]
>
> Maybe.
>
>> does org testing use --batch?
>
> Yup.
> If you have time and interest, it would be useful if you can integrate
> your regression tests with Org's test suite.
>
> --
> Ihor Radchenko // yantar92,
> Org mode contributor,
> Learn more about Org mode at <https://orgmode.org/>.
> Support Org development at <https://liberapay.com/org-mode>,
> or support my work at <https://liberapay.com/yantar92>
>


-- 
The Kafka Pandemic

A blog about science, health, human rights, and misopathy:
https://thekafkapandemic.blogspot.com

Re: batch breaks bugfix

[Samuel Wales]

On 6/19/23, Ihor Radchenko <yantar92@posteo.net> wrote:
> Samuel Wales <samologist@gmail.com> writes:
>
>> the bug occurs in recent bugfix, but not recent main, so perhaps we
>> could say that it is fixed in main but that the fix has not been
>> applied to bugfix?  [if only the fix were known.]
>
> Maybe.
>
>> does org testing use --batch?
>
> Yup.
> If you have time and interest, it would be useful if you can integrate
> your regression tests with Org's test suite.

[so as to not ignore you --- i would like that very much but npot poss
or useful to do that atm for me for unrelated reasons or for the tests
[ignore following unless of interest]:
- not signed up
- private data included [but useful tses]
- scattered shell functions + elisp to be extracted/cleaned
- running as --batch currently errors at emacs "Unknown option
‘-maximized’" [only one hyphen] with no backtrace even with
--debug-init.  i can't find where this originates from, although it is
guessably windowy code that --batch does not run -- i want to have it
work as --batch.  i think it does not seem to be my maximizing code in
.emacs.  i wonder what org standard testcase does to get no errors
like that.
- running non-batch works but is annoying with emacs popping up for
each test and loading .emacs for a few seconds
  - [currently not at stage of reducing that to ecm -Q elisp that
might or might not resolve any batch issues.
  - for my tastes it would ahve to be made to not violate DRY so as to
forestall syncing issues and would be less like running normal emacs
anyway, thus in principle could fail to flag possible future
non-intuitive bug-causing changes in .emacs
  - still want --batch though as lower-frame does not solve annoyance]
- make info, errors, etc. not yet appended to the diffable agenda
output for convenience in showing all factors like pulling and making
and running that could affect a given output and speed.
- truly bizarre bug with certain tasks not showing up in agenda in
certain repos.  i even have 2 repos, identical, just /named/
differently, where one does this and the other does not.  this is no
longer batch vs. non-batch; it's an unknown cause.
- currently i do make oldorg for everything but my non-canonical repos
havfe a local.mk.  also tried make.  not sure if such differences
could cause any bugs subtly.
- bugfix and main sorting still not getting ts-down tsia-down right.
maint gets it right.
  - i'd have to make ecm which is not possible atm.
- having said that it was a relief to have had this script for so many
years as i was able to be confident that the agenda was picking up
every relevant task, not picking up others, sorting, formatting
correctly, org version changes, .emacs...]

>
> --
> Ihor Radchenko // yantar92,
> Org mode contributor,
> Learn more about Org mode at <https://orgmode.org/>.
> Support Org development at <https://liberapay.com/org-mode>,
> or support my work at <https://liberapay.com/yantar92>
>


-- 
The Kafka Pandemic

A blog about science, health, human rights, and misopathy:
https://thekafkapandemic.blogspot.com

Re: batch breaks bugfix

[Samuel Wales]

correction for anybody who bothered to follow all that nonsense: the
bizarre bug with not including certain entries only exists in bugfix.
i had gotten it to exhibit itself in 2 repos with different names at
one point but never mind that.

Re: batch breaks bugfix

[Max Nikulin]

On 01/07/2023 10:03, Samuel Wales wrote:
> - running non-batch works but is annoying with emacs popping up for
> each test and loading .emacs for a few seconds

For running tests that creates windows it may be convenient to use 
xephyr that is a X server running as a window in the current X session. 
Earlier there was xnest with similar features.

Re: batch breaks bugfix

[Ihor Radchenko]

Samuel Wales <samologist@gmail.com> writes:

> - bugfix and main sorting still not getting ts-down tsia-down right.
> maint gets it right.

maint is obsolete and no longer used.

> - having said that it was a relief to have had this script for so many
> years as i was able to be confident that the agenda was picking up
> every relevant task, not picking up others, sorting, formatting
> correctly, org version changes, .emacs...]

Understood. That's why Org has regression tests as well :)

-- 
Ihor Radchenko // yantar92,
Org mode contributor,
Learn more about Org mode at <https://orgmode.org/>.
Support Org development at <https://liberapay.com/org-mode>,
or support my work at <https://liberapay.com/yantar92>

Re: batch breaks bugfix

[Samuel Wales]

i still use maint as a comparison against bugfix and main to make sure
they didn't intro regressions.  i even tested 8 for a long time so i
could go back and check formatting changes etc.  if maint gets it
right tht might be a clue is all.


On 7/1/23, Ihor Radchenko <yantar92@posteo.net> wrote:
> Samuel Wales <samologist@gmail.com> writes:
>
>> - bugfix and main sorting still not getting ts-down tsia-down right.
>> maint gets it right.
>
> maint is obsolete and no longer used.
>
>> - having said that it was a relief to have had this script for so many
>> years as i was able to be confident that the agenda was picking up
>> every relevant task, not picking up others, sorting, formatting
>> correctly, org version changes, .emacs...]
>
> Understood. That's why Org has regression tests as well :)
>
> --
> Ihor Radchenko // yantar92,
> Org mode contributor,
> Learn more about Org mode at <https://orgmode.org/>.
> Support Org development at <https://liberapay.com/org-mode>,
> or support my work at <https://liberapay.com/yantar92>
>


-- 
The Kafka Pandemic

A blog about science, health, human rights, and misopathy:
https://thekafkapandemic.blogspot.com