[BUG / possible BUG] Unused variable in starttls_io()

[dcooper16]

#1610 renamed $waitsleep to $nr_waits in starttls_io(). However, $waitsleep still appears in one line:

[[ -n "$3" ]] && waitsleep=$3

The result is that $3 is ignored.

I thought the solution would be to just replace waitsleep with nr_waits on this line. However, the third parameter to starttls_io() is always 1, and setting $nr_waits to 1 would result in the loop never running and starttls_io() always returning 1. (Prior to #1610 $waitsleep was always set to 1 and so the loop never ran, but starttls_io() would always return 0.)

Before you open an issue please check which version you are running and whether it is the latest in stable / dev branch

• I am running version from the git repo :
• I am running: 3.2 and 3.3dev
• Different version, check for updates here:

Before you open an issue please consult the FAQ and check whether this is a known problem by searching the issues

• Is related to ...
• couldn't find anything

Command line to reproduce (or docker command)

N/A

Expected behavior

N/A

Your system (please complete the following information):

N/A

AI section

• I found a bug using LLM version:
• It was just me

[drwetter]

Thanks, David. That seems to be a jabber (edit1: and postgres) only issue, at least looking @ 3.3dev and 3.2 , so as a first shot I just issued a PR which should fail in t/21_baseline* .

Edit2: It did not. Probably one has to go between 2.9.5 and #1610 in order to find out what happened. Looking @ 2.9.5 at L7405 ff it looks to me that a loop was not introduced before. At least in 2.9.5 it was just one shot for testing and it seemed to work for a long time.

[drwetter]

Quick tests showed so far that in the current tests where the loop should be executed like up to 10 times it always needed one shot only.

I would suggest doing some more tests (https://list.jabber.at/) and find the right value how often the loop should be executed to retrieve a reliable verdict --> tomorrow.