+++ This bug was initially created as a clone of Bug #2070450 +++

Description of problem:
Scheduling at jobs in python using subprocess does not give expected result.
Somewhere something goes wrong with a `marcinDELIMITER`.


Version-Release number of selected component (if applicable):
- at 3.1.20
- Python 3.6.8


How reproducible:
Very


Steps to Reproduce:
1. open python3 console
2. paste code:
```
import subprocess
p = subprocess.Popen(
    ['/bin/at', 'now'],
    stdin=subprocess.PIPE,
)
stdout, stderr = p.communicate(
    input=b'echo something > test',
)
```

Actual results:
File created named: `testmarcinDELIMITER31aa8a18` with contents: `something`


Expected results:
File created named `test` with contents: `something` 


Additional info:
Running at from the command line works just fine e.g.: `at now <<<'echo something > test'`
Also `input=b'echo something > test\n'` works too.


So my question is this, is this a "bug" in AT, or should python subprocess code always include the newline (\n) character.

--- Additional comment from Jan Staněk on 2022-03-31 09:00:13 UTC ---

Hi, thanks for the report. You are not the only one with that problem, I found this salt issue that seems to stem from the same bug: https://github.com/saltstack/salt/issues/58510
That issue points me to an old at bug#486844. From quick glance, it seems like an at issue (missing newline before the delimiter), but I will have to investigate further.

In the mean time, the suggested workaround is to always include the trailing newline in your Python input field.

--- Additional comment from Jan Staněk on 2022-03-31 09:16:08 UTC ---

This issue is currently present in Fedora 36 Beta too.

Minimal reproducer: `$ echo -n 'touch test.txt'|at now`
Current result: File `test.txtmarcinDELIMITERxxxxxxxx` is touched/created, where `xxxxxxxx` is soime random hexadecimal number (grep basic regex: `marcinDELIMITER[[:xdigit:]]\{8\}$`).
Expected result: File `test.txt` is touched/created.

--- Additional comment from Onmeac on 2022-03-31 12:13:45 UTC ---

> https://github.com/saltstack/salt/issues/58510

Yes that was "us" too, almost a year ago, but that ticket never got anywhere and at the time we had not yet migrated to EL8.

Now we are migrating and in doing so it is becoming more noticeable (in our situation) with using salt.

Thanks for taking the time to investigate this issue.