Re: [PATCH] tests: Determine current test runner before test-end is called

[vzh]

On Fri, Jun 18, 2021 at 07:51:15PM +0200, soeren@soeren-tempel.net wrote:
> From: Sören Tempel <soeren+git@soeren-tempel.net>
> 
> According to SRFI 64 test-end will remove the test-runner installed by
> the matching test-begin. As such, it seems that test-runner-current will
> return #f, if called after test-end. Since all skribilo test cases call
> test-runner-current after test-end, the following exception is currently
> thrown with Guile 3.0.6:
> 
>       Backtrace:
>       In ice-9/boot-9.scm:
>         1752:10  6 (with-exception-handler _ _ #:unwind? _ # _)
>       In unknown file:
>                  5 (apply-smob/0 #<thunk 7fe66b7d4ee0>)
>       In ice-9/boot-9.scm:
>           724:2  4 (call-with-prompt _ _ #<procedure default-prompt-handle…>)
>       In ice-9/eval.scm:
>           619:8  3 (_ #(#(#<directory (guile-user) 7fe66b7cec80>)))
>       In ice-9/boot-9.scm:
>          2835:4  2 (save-module-excursion _)
>         4380:12  1 (_)
>       In 
> /home/soeren/src/aports/testing/skribilo/src/skribilo-0.9.5/tests/ast.test:
>           116:9  0 (_)
> 
> Thereby always causing all tests, as executed by `make test`, to fail. I
> am not too familiar with Guile so it is unclear to me how the current
> code ever worked. My proposed patch stores the current failure count in
> a local variable before invoking test-end, thereby fixing the exception
> above.

Hi Sören,

It rather an issue with the Guile version you use.  In lepton-eda
we had to work around it by editing tests.  SRFI-64 was broken
in Guile somewhere between 3.0.5 and 3.0.6.  The 3.0.7 release
promises to fix the issue.  Try it first.

-- 
  Vladimir