Re: master updated (5af5ed6c62 -> f07505d1ec)

[Jim Porter]

On 9/7/2022 4:46 AM, Lars Ingebrigtsen wrote:
> In the last run on EMBA, that test didn't fail, but this one did:
>
> Test esh-proc-test/sigpipe-exits-process condition:
>      (ert-test-failed
>       ((should
>         (eq
>         (process-list)
>         nil))
>        :form
>        (eq
>         (#<process sh-stderr>)
>         nil)
>        :value nil))
>     FAILED  16/16  esh-proc-test/sigpipe-exits-process (1.013610 sec) at 
> lisp/eshell/esh-proc-tests.el:122
> Ran 16 tests, 14 results as expected, 1 unexpected, 1 skipped (2022-09-07 
> 10:27:03+0000, 2.594968 sec)
> 1 unexpected results:
>     FAILED  esh-proc-test/sigpipe-exits-process  ((should (eq
>      (process-list) nil)) :form (eq (#<process sh-stderr>) nil) :value
>      nil)

Luckily, that's an easy one to fix. See attached.

The changes for bug#21605 resulted in that test spawning an extra pipe 
process to monitor stderr, which confused the test. For simplicity, I 
just tweaked the pipeline in that test to avoid spawning the pipe process.

0001-Fix-a-race-condition-in-an-Eshell-test.patch
Description: Text document