Ihor Radchenko <[email protected]> writes:

> Morgan Smith <[email protected]> writes:
>
>> I have a very minor batch of test improvement patches.
>
> LGTM. Feel free to push.
>

Thanks for the review!

Applied, as the following commits:

3be432938 Testing: ob-haskell/ghci-info: Fix test
c11007996 Testing: Fix detection of git-annex executable
95451c771 Testing: Don't set `org-babel-temporary-directory' in tests
4d98337bc Testing: Fix various compile warnings


Org babel test failures discussion:

Foreword: Locally, I hardcoded in a rough timeout feature.  Does org
babel not have a native timeout feature for when the process just hangs?

Also I only did the roughest of investigations on some of these that I
am now recalling from memory and a build log.  I probably should have
written down my finding but honestly fixing these issues is a very low
priority for me.

>> I've got 2 failures in ob-maxima
>
> I saw some failures on ob-maxima due to changes in the latest maxima
> release. I did not diagnose them.

I think the errors didn't make it to the error buffer?  Trying to recall
from memory.

>> All the Julia tests fail
>
> Confirmed.
> ess-julia no longer has `julia' function.
> There is something called `run-ess-julia' but it does not seem like a
> drop-in replacement. We need someone familiar with julia to fix this.

I had some strange autoload issues.  My ess-julia had an autoloaded
`julia` function which is just a wrapper for `run-ess-julia' but the
autoload didn't work and I had to add "ess-julia" to my make-file flags
(BTEST_EXTRA).

With the flags in place the tests seem to run but half timeout and the
other half return something unexpected

>> One ruby test fails
>
> I saw that one.
> It is about different number of newlines.
> Also related to changes in the latest ruby.
> Most likely, the test needs to be updated, but I did not check in
> details. Maybe it is a genuine bug.
>
>> All the scheme tests fail
>
> Scheme tests pass on my side using guile.
>

For some reason, using guile on my guile based operating system is
always funky.  I'm pretty sure this is a me issue

#+begin_example
   FAILED  test-ob-scheme/cons-cell  "Error reading from stdin"
   FAILED  test-ob-scheme/list-conversion  "Error reading from stdin"
   FAILED  test-ob-scheme/prologue  "Error reading from stdin"
   FAILED  test-ob-scheme/proper-list  "Error reading from stdin"
   FAILED  test-ob-scheme/tables  "Error reading from stdin"
   FAILED  test-ob-scheme/unspecified  "Error reading from stdin"
   FAILED  test-ob-scheme/variable-assignment  "Error reading from stdin"
   FAILED  test-ob-scheme/verbatim  "Error reading from stdin"
   FAILED  test-org-attach-git/get-maybe  "Opening input file"
#+end_example


>> I have yet to try running the csharp tests using mono.  The default compiler 
>> is
>> "dotnet" which isn't available on Guix as the FSF doesn't like it or 
>> something.
>> Not sure exactly what's wrong with dotnet from a freedom perspective.  Might 
>> be
>> something about bootstrapping it.
>
> Official dotnet passes tests for me.
>

Good to know

>> These commits mostly pass CI Morgan.  Recent commits to the tests have broken
>> them with LANG set to non-english and on Emacs 28.  I might get around to
>> fixing that eventually.
>
> What exactly fails?

Test failures unrelated to this thread:

On Emacs-28 these tests fail:
FAILED  ox-html/test-link-home-and-use-abs-url/no-base-directory  
ert-with-temp-file
FAILED  ox-html/test-org-html-base-directory  file-name-parent-directory
FAILED  ox-html/test-org-html-base-directory/with-base-directory-set 
file-name-parent-directory

Because those symbols (ert-with-temp-file and file-name-parent-directory) where
introduced in Emacs 29.

With LANG=he_IL.UTF-8 I get failures in `test-org/auto-repeat-maybe' due to
hard coded abbreviated day names in timestamps.


Reply via email to